(Circulation. 1999;99:1109-1125.)
© 1999 American Heart Association, Inc.
Abstracts of the 39th Annual Conference |
Poster Presentations
P1 Periodontal Disease as a Risk Factor for CVD, CHD, and STROKE: The First National Health and Nutrition Examination Survey (NHANES I) and Its Follow-Up Study
Tiejian Wu, Maurizio Trevisan, Robert Genco, Joan Dorn, Karen Falkner, Christopher Sempos, National Institutes of Health, Bethesda, MA, SUNY at Buffalo, Buffalo, NY
Oral microflora associated with periodontal disease (PD) has been proposed to be a causal factor for cardiovascular disease (CVD). Data from NHANES I and its 21-year follow-up were used to test this hypothesis. Baseline periodontal status was categorized into (1)no PD, (2)gingivitis, (3)periodontitis, and (4) edentulousness. CVD events during follow-up were ascertained by hospital records for non-fatal events and death certificates for fatal events. Relative risk (RR) and 95% confidence interval (CI) were derived from Cox regression after adjusting for demographic variables and several well-established CVD risk factors. 9,962 people were free from coronary heart disease (CHD), heart failure, and cancer at baseline. 2,844 CVD, 1,468 CHD, and 803 stroke events occurred during the follow-up. Compared to no PD, RRs (CI) of CVD were 1.05 (0.93-1.18) for gingivitis, 1.17 (1.04-1.31) for periodontitis, and 1.22 (1.10-1.34) for edentulousness. RRs (CI) at similar PD levels for CHD were 1.03 (0.87-1.21), 1.14 (0.98-1.34), and 1.13 (0.98-1.32), and for stroke were 1.03 (0.81-1.31), 1.33 (1.07-1.66), and 1.30 (1.06-1.60), respectively. Analyses stratified by age group indicated that elevated risk for CVD associated with PD is manifested mainly in those aged 25-54 years at baseline. Among this age group, RRs (CI) of CVD were 1.13 (0.96-1.33) for gingivitis, 1.40 (1.16-1.68) for periodontitis, and 1.36 (1.11-1.68) for edentulousness in comparison to no PD; RRs (CI) of CHD were 1.13 (0.80-1.29), 1.33 (1.03-1.72), and 1.25 (0.93-1.67); and RRs (CI) of stroke were 0.96 (0.64-1.46), 1.57 (1.05-2.36), and 1.46 (0.92-2.33), respectively. This study suggests that periodontal disease is a significant risk factor for CVD, CHD, and stroke especially in adults aged 25-54.
P2 Stress in the workplace and early atherosclerosis. The Los Angeles Atherosclerosis Study
Cheryl Nordstrom, Kathleen M Dwyer, Noel Bairey Merz, Anne Shircore, Ping Sun, Wei Sun, James H Dwyer, Cedars Sinai Medical Center, Los Angeles, CA, University of Southern Califonia, Los Angeles, CA, University of Southern California, Los Angeles, CA
Background. A link between psychosocial stress and atherosclerosis remains controversial. The current study assessed the relationship between workplace stress and the prevalence of atherosclerosis in the carotid arteries.
Methods. Participants were from a cohort of 573 randomly sampled utility workers asymptomatic for cardiovascular disease who were aged 40-60 years at entry (217 women and 247 men with complete data for the stress measure and covariates). Measurement of stress was introduced at the 18-month follow-up exam, and all measures were taken from that exam. Stress was measured by a 6item questionnaire concerning workplace demands and difficulty sleeping due to worries about work. The presence of raised lesions in the left and right bifurcation and distal common carotid arteries was determined from high resolution Bmode ultrasound images by agreement of two readers. The risk of any lesion was related to stress by logistic regression with adjustment for possible confounders.
Results. Carotid lesions were detected in 25.9% of men and 19.8% of women. After controlling for age, height, ethnicity, BMI, SBP, total plasma cholesterol, HDL-C, diabetes, use of cholesterol and blood pressure lowering medications, and behavioral factors (smoking, alcohol intake, exercise, job category, fat intake), the odds of carotid lesions were 4 times greater in men scoring at the 90th percentile on the stress scale than men at the 10th percentile (OR=4.3; 95% CI=1.3 to 14.9). Among women, the risk of carotid lesions was unrelated to stress (OR=1.1; 95% CI=0.3 to 3.8).
Conclusion. The finding that middle-aged men who report more workplace stress are at increased risk for carotid lesions suggests that stress plays a role in the etiology of atherosclerosis. Women in this age group may be protected from such effects.
P3 Relationship of C-reactive protein with body fat distribution and coronary risk factors in South Asians and Europeans
Nita G Forouhi, Naveed Sattar, Paul M McKeigue, Glasgow Royal Infirmary University NHS Trust, Glasgow, United Kingdom, London School of Hygiene & Tropical Medicine, London, United Kingdom
The excess risk of coronary heart disease (CHD) in South Asians
versus Europeans is not fully explained by conventional risk factors,
but the role of C-reactive protein (CRP), a marker for inflammation
that predicts CHD risk, has not been studied. We measured CRP
concentration in 113 healthy age and BMI matched S. Asians and
Europeans aged 40-55 years in west London. Measurements included
determination of % fat by DEXA scan, visceral fat area (VFA) by
abdominal CT scan, insulin sensitivity index (ISI) by short insulin
tolerance test and fat tolerance test. There were more South Asians in
the top two tertiles of CRP distribution than there were Europeans
(
2 = 8.6; p=0.013). In age, sex and smoking adjusted
regression analyses CRP concentrations were more strongly
associated with central obesity [ß=0.51, p=0.013 in Europeans and
ß=0.41, p=0.004 in S. Asians for VFA] than overall obesity
[ß=0.55, p=0.026 in Europeans and ß=0.34, p=0.096 in S. Asians for
%fat]. CRP level was significantly associated (p<0.05) with insulin
(0h and 2h) and triglyceride (0h and 8h) in both ethnic
groups, and also with HDL-cholesterol (negative) and 0h
glucose in Europeans, but not with ISI in either group. After
additionally adjusting for % fat, there remained an independent
association of CRP level with 0h glucose and insulin and
HDL-cholesterol in Europeans, and with
triglyceride level in S. Asians. We have reported new
findings of (i)higher levels of CRP in S. Asians, and (ii) association
between CRP level and central obesity and insulin level in both ethnic
groups, and have confirmed previous reports of association of CRP level
with triglyceride and HDL-cholesterol. We
conclude that CRP level is related to CHD risk in S. Asians and it
should be measured in future studies.
P4 The Effects of Vitamin C and Vitamin E on In Vivo Lipid Peroxidation: Results from A Controlled Clinical Trial
Han-Yao Huang, Lawrence J. Appel, Kevin D. Croft, Edgar R. Miller, Trevor A. Mori, Ian B. Puddey, Johns Hopkins University, Baltimore, MD, University of Western Australia, Perth, Australia
Although antioxidant vitamins may prevent lipid peroxidation, in
vivo evidence in support of this hypothesis is sparse. To
determine the effects of vitamins C and E on in vivo lipid
peroxidation, we conducted a placebo-controlled, double-masked, 2x2
factorial trial of vitamin C (500 mg/day) and vitamin E (400 IU/day) in
184 nonsmokers (mean (SD) age: 58 (14), 55% women, 50% African
American). Mean duration of supplementation was two months. Outcome
measures were changes from baseline in urinary
8-iso-prostaglandin F2
(8-iso-PGF2, a specific in vivo measure of
free radical-catalyzed peroxidation of arachidonic
acid), urinary malondialdehyde (MDA)/4-hydroxyalkenals, and serum
oxygen-radical absorbance capacity (ORAC).
At the end of supplemention, mean±SE change in urinary
8-iso-PGF2 (pg/mg creatinine) was 9.0±68.4
(p=0.89), -150.2±64.0 (p=0.02), -141.3±45.5 (p=0.003), and
-112.5±62.4 (p=0.08) in the placebo, vitamin C alone, vitamin E alone,
and combined vitamins C and E groups, respectively. Compared to
placebo, reductions in 8-iso-PGF2 in the vitamin C
alone, vitamin E alone, and combined vitamins C and vitamin E groups
approached statistical significance (p=0.09, 0.07, and 0.19,
respectively). Regression analyses suggested a subadditive
interaction (p=0.08) of the two vitamins on reduction in urinary
8-iso-PGF2. Both vitamins, alone or combined, had no
effect on urinary MDA/4-hydroxyalkenals. Vitamin C increased ORAC by
185±76 µmol/L of Trolox units (p=0.01); vitamin E had no
significant effect on ORAC.
In conclusion, supplementation with vitamin C alone or vitamin E alone
appeared to reduce in vivo lipid peroxidation as measured by
urinary 8-iso-PGF2, but combined vitamin C and vitamin E
conferred no additional benefit.
P5 The effects of chronic physical activity on endothelial function in postmenopausal women
Ronald McKechnie, Melvyn Rubenfire, Lori Mosca, The University of Michigan Medical Center, Ann Arbor, MI
National recommendations suggest moderate levels of physical activity
may be cardioprotective, however, data to substantiate this in women
are sparse. We studied the effect of chronic physical activity on
endothelial dependent function in non-smoking,
non-diabetic, postmenopausal women (n=34, mean age 65.6 ± 7.4
yrs). Endothelial function was determined by percent
and absolute change of the brachial artery diameter measured before and
after occlusion (n=33) and in response to cold pressor testing (n=32).
Physical activity was assessed by a standardized questionnaire and by
metabolic equivalents expended/week based on a daily
exercise log. Our results indicate on average, 35.3% of the women were
exercising 
3 times/week enough to work up a sweat, 4.8 flights
of stairs were climbed/day, 5.0 city blocks were walked/day and 29.4%
participated in moderately physically demanding daily activity. There
was a significant association between the number of city blocks walked
daily and the percent and absolute change in brachial reactivity in
response to cold pressor testing (Spearman R=0.35, p<0.05 and R=0.37,
p<0.05 respectively). There was also an association between moderately
physically demanding daily activity and percent change in brachial
reactivity in response to cold pressor testing (p=0.03). Other more
intense levels of physical activity were not significantly associated
with endothelial function. Multivariate
analysis adjusting for BMI and lipids mitigated the association
between city blocks walked or moderate demanding daily physical
activity and endothelial function, suggesting that
physical activity may improve endothelial function via
these mechanisms. This study supports moderate levels of physical
activity, such as regular walking, in postmenopausal women may be
cardioprotective.
P6 Risk Factors for Cardiovascular and Total Mortality Among U.S. Male Physicians with Prior Myocardial Infarction
Julia C L Wong, J Michael Gaziano, Robert J Glynn, Julie E Buring, Charles H Hennekens, Harvard Medical School, Brigham and Women's Hospital, Boston, MA
Background. Survivors of myocardial infarction(MI) are at increased risk for fatal events. The cohort of ineligibles for the Physicians' Health Study due to their prior MI afforded a unique opportunity to study the risk factors associated with cardiovascular (CV) and total mortality in a homogeneous population of U.S. male physicians.
Methods and Results. In 1982, baseline questionnaires were sent to U.S. male physicians 40 to 84 years of age to assess eligibility to participate in the Physicians' Health Study, a randomized trial of aspirin and beta carotene in the primary prevention of CV disease and cancer. A total of 6102 men who reported a prior MI and were therefore ineligible to participate in the trial provided information on CV risk factors. During a mean follow-up period of 5 years, 1103 deaths were recorded, of which 842 (76.3%) were CV. Age, current smoking, as well as histories of angina, diabetes and hypertension, were positively associated with both CV and total mortality. Moderate exercise(>once a week), moderate alcohol intake(2-7 drinks per week), and body mass index(BMI) were inversely associated. In multivariate analysis, significant predictors of CV and total mortality included age (relative risks 1.07 and 1.07 per year respectively), current smoking (1.90 and 2.17), diabetes (1.87 and 1.92), hypertension (1.44 and 1.35), angina (1.30 and 1.19), moderate exercise (0.65 and 0.64) and moderate alcohol intake (0.79 and 0.78). BMI was significantly predictive of total but not CV mortality. Conclusions. These results in a population at high absolute risk of death emphasize the clinical importance of risk factor modification for patients with prior myocardial infarction to reduce their subsequent CV and total mortality.
P7 The effect of changing treatment standards on observational epidemiology: The case of blood pressure lowering medication.
Daniel L McGee, Youlian Liao, Hongsheng Cai, Ricard S Cooper, Loyola University Medical Center, Maywood, IL
Treatment of elevated blood pressure (BP) has become increasingly prevalent within the population. For individuals, 55-74, 20% of NHANES I participants and 31% of NHANES III, took anti-hypertensives. The increasing use of BP drugs means that the observational data no longer reflect naturally occurring BP levels. If unrecognized treated BP levels are being used in analyses in place of naturally occurring levels, outcomes reflect relationships other than those thought to be under examination. We use empirical data and simulations to demonstrate the effect of this phenomena in two settings: Comparing the BP distribution in populations; and Regressing BP on Body Mass Index. A comparison of BP distributions between Black men living in the U.S. and Nigeria found that the mean systolic BP is the same in the two populations. This similarity of the average BP levels is explained by the fact that many participants from the US cohort received treatment for hypertension, which lowered the average level for the cohort. If the comparison is restricted to those not on medication and with systolic BP < 140 mmHg, the comparison shows the expected result of higher BP among the US cohort.
We demonstrate the effect of treatment on discerning relations by simulations of regressing systolic BP on body mass index (BMI). If treatment is ignored, the relationship discerned depends on the percentage receiving treatment and the treatment threshold; when only 25% of the patients are treated, the bias is small, but it increases as the percentage treated increases; reaching a 4 fold decrease in the estimated coefficient when a treatment threshold of 140 mm Hg is assumed and the entire treatable population is assumed to be treated. Deleting people on treatment from the analysis reduces the bias only partially.
P8 Angiotensin Converting Enzyme Genotype Influences Blood Pressure Response to Oral Sodium in African-Americans
John M. Flack, Beth A. Staffileno, Marwan Hamaty, Helena Kuivaniemi, Gerard Tromp, Carla Yunis, Amanda Dudley, Richard H. Grimm, Hennepin County Medical Center, Minneapolis, MN, Wake Forest University, Winston Salem, NC, Wayne State University, Detroit, MI
We examined the influence of the ACE genotype on BP response to oral sodium in 81 of 112 normotensive African-Americans in a 24-week, two-period cross-over, randomized, placebo-controlled trial consisting of two 8-week treatment periods separated by an 8-week placebo-washout phase. Baseline BP was 104/69 mm Hg and mean age was 40 years. 24-hour urinary sodium content and the urinary sodium to creatinine ratio were 27.3% (35 mmol) and 32.4% (0.23) higher, respectively, after sodium supplementation. Variability-adjusted BP change was defined as: BP difference at the end of the respective treatment periods (sodium minus placebo) divided by the intraperson SD of 3 BP measures taken over 1 month prior to dietary intervention. We previously reported the statistical superiority of variablity-adjusted compared to unadjusted BP change. After adjustment for baseline urinary sodium to creatinine ratio, SBP level, sex, age, and change in urinary sodium to creatinine ratio, variability-adjusted, but not unadjusted SBP change, varied significantly by ACE genotype. Variability-adjusted SBP change was +3.17 higher (P<.05) in the II compared to DD homozygotes and was only significant amongst II homozygotes +2.69 [95% CI 0.6, 4.8]. SBP response to dietary sodium is hetergenous in normotensive African-Americans and appears to vary according to ACE genotype.
P9 Evaluation of the Risk Stratification Guidelines Defined by the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC VI) Report
Lorraine G Ogden, Jiang He, Suma Vupputuri, Paul K Whelton, Tulane School of Public Health and Tropical Medicine, New Orleans, LA
The JNC VI recently recommended that risk strata, in addition to blood
pressure (BP) stage, should be considered in the treatment of
hypertension. We used data from the NHANES I Epidemiological Follow-Up
Study to calculate the number needed to treat to prevent one death,
according to BP stage (high-normal: 130-139/85-89mmHg; stage 1:
140-159/90-99mmHg; stages 2 and 3: 160/100mmHg) and risk group
(group A: no risk factors, no target organ disease [TOD], and no
clinical cardiovascular disease [CCD]; group B: at
least one risk factor, not including diabetes or TOD/CCD; group C:
TOD/CCD and/or diabetes). Data on BP, other risk factors including age,
gender, menopausal status, current smoking, serum
cholesterol, and diabetes, and TOD/CCD were collected on
12267 participants who were 25 to 74 years old and not taking
anti-hypertensive medication at baseline. Over an average of 19 years
follow-up, 3605 deaths were documented. Logistic regression was used to
predict the probability of death according to systolic BP and
risk group. The following table presents the number needed to treat
in each risk group and BP stage, assuming a 12 mmHg reduction in
systolic BP. The data suggest that lowering BP in persons with
risk factors or TOD/CCD is cost-effective for all BP stages. Our
findings support the JNC VI recommendation that risk strata be
considered in the treatment of hypertension. 
|
P10 Serum potassium and cardiovascular mortality
Jing Fang, Shantha Madhavan, Hillel Cohen, Michael H Alderman, Albert Einstein College of Medicine, Bronx, NY
To determine the relation of serum potassium(SK) to cardiovascular disease(CVD) mortality in general adult population, the NHANES I Epidemiological Follow-up Study(1974-1992) was examined. Of 2992 initial subjects with baseline SK, 156 without vital status in 1992, 2836 subjects with SK range 2.7-5.4 mmol/L were left for study. Mean±standard deviation(SD) for age and SK were 46.6±13.9 years and 4.07±0.34 mmol/L respectively. During 15.9 years follow-up, of 578 deaths, 272 were CVD deaths(6.03/1000 person-years). Subjects were stratified into 3 groups by mean±1 SD SK as cut points: low 2.7-3.7(N=477), middle 3.8-4.4(N=1982), high 4.5-5.4(N=377). Those with high SK were older, had higher blood urea nitrogen, serum creatinine, cholesterol, and more males, history of CVD, current smokers than those with low and middle SK. The low SK group included more diuretic users(21%) than middle and high groups(7%). In both diuretic and non-diuretic users, the age-gender-race-adjusted CVD death/1000 person-years(Table) for high SK was significantly higher than that for middle, which was similar to low SK. Cox regression analysis confirmed this, using middle SK as reference.
Thus, while low SK is not associated with increased CVD mortality, an
elevated SK(>4.4mmol/L) bears a significant, independent, positive
association with CVD mortality - particularly among diuretic
users.
|
P11 Alcohol Consumption and Blood Pressure in African-Americans, Mexican-Americans and Whites: Results from the Third National Health and Nutrition Examination Survey (NHANES III)
Suma Vupputuri, Jiang He, Lorraine G Ogden, Catherine Loria, Paul K Whelton, National Center for Health Statistics, Hyattsville, MD, Tulane School of Public Health and Tropical Medicine, New Orleans, LA
Limited information is available regarding the association between
alcohol consumption and blood pressure (BP) in minority populations in
the United States. We examined the relationship between alcohol use and
blood pressure (BP) in 4,028 African-American (AA), 4,328 Mexican-
American (MA), and 5,895 white NHANES III participants who were 18
years and not taking antihypertensive medication. BP was characterized
as the mean of six measurements obtained by trained observers using a
standard sphygmomanometer. Alcohol consumption during the preceeding 12
months was estimated using a standard questionnaire. Among study
participants, 42.1% of AA, 43.5% of MA and 47.7% of whites reported
alcohol consumption during the preceeding 12 months. The age-sex-
adjusted systolic and diastolic BPs among AA, MA,
and whites were 124.8, 123.3, and 121.4 mm Hg, and 75.2,
72.7, and 72.8 mm Hg, respectively. Among alcohol users, the
age-sex-adjusted consumption of alcohol was 1.1, 1.0 and 0.9 drinks/day
for AA, MA, and whites, respectively. After adjustment for age, sex,
education, body mass index, dietary intake of total energy,
polyunsaturated- to-saturated fat ratio, carbohydrate, sodium and
potassium, one drink of alcohol per day was significantly and
positively associated with 0.61, 0.49 and 0.89 mm Hg higher
systolic BP, and 0.60, 0.48 and 0.30 mm Hg higher
diastolic BP, among AA, MA, and whites, respectively (all
p<0.01). Our results indicate the presence of a strong and
consistent positive relationship between alcohol use and BP in
AA, MA, and whites. Further, they suggest that alcohol consumption may
be an important etiologic factor in the development of hypertension in
all three ethnic groups.
P12 Dietary Potassium Intake and Blood Pressure in African-Americans and Whites: Results from the Third National Health and Nutrition Examination Survey (NHANES III)
Suma Vupputuri, Jiang He, Lorraine G Ogden, Catherine Loria, Paul K Whelton, National Center for Health Statistics, Hyattsvilles, MD, Tulane School of Public Health and Tropical Medicine, New Orleans, LA
It has been suggested that a lower intake of dietary potassium among
African-Americans (AA) might explain their higher risk of developing
hypertension compared to whites. We addressed this question in 3,946 AA
and 9,529 white NHANES III participants who were 18 years and not
taking antihypertensive medication. Blood pressure (BP) was
characterized as the mean of 6 measurements obtained by trained
observers using a standard sphygmomanometer. Dietary potassium and
sodium were estimated by a 24-hour dietary recall. Compared to whites,
AA had a significantly higher age-sex-adjusted mean systolic BP
(123.6 vs. 121.2 mm Hg), diastolic BP (74.7 vs.
72.6 mm Hg), body mass index (27.2 vs. 26.3
kg/M2), and dietary intake of sodium (151.8 vs. 147.8
mmol/24 h), but a lower intake of potassium (61.7 vs. 73.3 mmol/24
h). Dietary potassium intake was inversely related to both
systolic and diastolic BP in AA and whites. After
adjustment for age, sex, BMI, alcohol use, and dietary intake of total
energy, sodium, carbohydrate and polyunsaturated-to-saturated fat
ratio, a 50 mmol higher dietary potassium intake was significantly
associated with a -1.45 and -1.45 mm Hg lower
systolic, and -0.84 and -0.38 mm Hg lower
diastolic BP for AA and whites, respectively. After
adjustment for dietary potassium intake in multiple linear regression
models, the difference in average systolic BP between AA and
whites was reduced from 2.46 to 2.28 mm Hg and the
corresponding difference in diastolic BP was reduced from
2.14 to 2.07 mm Hg. These data suggest that a lower dietary
potassium intake explains, in part, the differences in average BP
between AA and whites, and they emphasize the potential importance of
dietary potassium intake in the prevention of hypertension, especially
in AA.
P13 Dietary Sodium Intake and Subsequent Risk of Cardiovascular Disease in Overweight U.S. Men and Women
Jiang He, Lorraine G Ogden, Suma Vupputuri, Catherine Loria, Paul K Whelton, National Center for Health Statistics, Hyattsville, MD, Tulane School of Public Health and Tropical Medicien, New Orleans, LA, Tulane School of Public Health and Tropical Medicine, New Orleans, LA
Ecologic analyses and animal studies have suggested that a high dietary sodium intake increases the risk of stroke. However, the relationship between dietary sodium intake and risk of cardiovascular disease (CVD) is not well studied at the individual level. We analyzed data in 931 men and 1757 women who participated in the NHANES I Epidemiologic Follow-up Study and were overweight at their baseline examination. Dietary sodium and caloric intake were estimated at baseline using a 24-hour dietary recall method. Other important risk factors for CVD including age, sex, race, serum cholesterol, blood pressure, body mass index, history of diabetes, education, physical activity, cigarette smoking, and alcohol consumption were also measured at baseline and were used for adjustment in Cox proportional hazards models. Incidence and mortality data for CVD were obtained from medical records and death certificates. Over an average of 19 years of follow-up, we documented 250 stroke events (87 fatal), 647 coronary heart disease (CHD) events (214 fatal), 329 CVD deaths and 810 deaths from all causes. For an average caloric intake, a 100 mmol increase in sodium intake was associated with a 35% increase (RR, 1.35; 95% CI, 1.07 to 1.70; p=0.01) in stroke incidence, a 96% increase (RR, 1.96; 95% CI, 1.33 to 2.89; p<0.001) in stroke mortality, a 48% increase (RR, 1.48; 95% CI, 1.15 to 1.90; p=0.002) in CHD mortality, a 67% increase (RR, 1.67; 95% CI, 1.34 to 2.07; p<0.001) in CVD mortality, and a 43% increase (RR, 1.43; 95% CI, 1.24 to 1.64; p<0.001) in mortality from all causes. These findings indicate that high dietary sodium intake is a strong independent risk factor for CVD and total mortality in overweight persons.
P14 Blood Pressure Decreases in Response to a Change in Posture: Variations by Menopausal Status and Use of Hormone Replacement Therapy among Women in the Atherosclerosis Risk in Communities Study, 1987-1989
Kathryn M. Rose, Herman A. Tyroler, Sara A. Ephross, Donna K. Arnett, David Couper, Kathleen C. Light, Thomas N. Skelton, University of Minnesota, Minneapolis, MN, University of Mississippi Medical Center, Jackson, MS, University of North Carolina at Chapel Hill, Chapel Hill, NC, World Wide Epidemiology, Glaxo Wellcome, RTP, NC
A decrease in blood pressure (BP) in response to a change in posture
(supine to standing)has been associated with an increased incidence of
CHD in the Atherosclerosis Risk in Communities (ARIC)
Study. We also investigated whether a BP response to a postural
challenge varies by menopausal status and if so, whether variations are
influenced by hormone replacement therapy (HRT) use (either unopposed
estrogen or estrogen and progestin). Included were 2243
African-American and white women between the ages of 45 and 54 who
participated in the baseline examination of the ARIC Study and were
classifiable as pre- (n=1068) or post-(n=389 with and n=786 without use
of HRT) menopausal. Women in the lowest three deciles of SBP and DBP
change (cutoff values: 
-4.7 and 0.33 mm Hg, respectively)
were classified as "decreasers"; others were categorized into the
referent group. In age-adjusted analyses, postmenopausal women
were significantly more likely to experience BP decreases in response
to a postural challenge than were premenopausal women (odds ratio
(OR)=1.4, 95% confidence interval =1.1-1.8 for both SBP and DBP).
Adjustment for established risk factors for orthostatic
hypotension (ethnicity, sitting SBP, diabetes, and use of
antihypertensive medications) did not attenuate these associations.
Postmenopausal women were categorized by HRT use and again contrasted
to premenopausal women. Associations were consistent with those
noted overall; however, among the HRT group, ORs were weaker and no
longer statistically significant. In summary, a BP decrease in response
to a postural challenge, a previously demonstrated predictor of
incident CHD in the ARIC study, was more prevalent in post- than
pre-menopausal women.
P15 a prospective study of snoring and risk of hypertension and cardiovascular disease in women
Frank B Hu, Walter C Willett, JoAnn E Manson, Graham A Colditz, Alberto Ascherio, Eric B Rimm, Bernard Rosner, Frank E Speizer, Charles E Hennekens, Meir J Stampfer, Brigham and Women's Hospital, Boston, MA, Harvard School of Public Health, Boston, MA, Harvard University, Boston, MA
Regular snoring can chronically activate the sympathetic nervous system due to intermittent hypoxia from obstructed breathing, and thus may lead to systemic hypertension. In addition, hypoxemia may promote atherosclerosis. Epidemiologic studies on snoring and risk of hypertension and cardiovascular disease (CVD) have been inconsistent. We prospectively examined the association between snoring and incidence of hypertension and cardiovascular disease among 71,779 female nurses aged 40 - 65 and without previously diagnosed cardiovascular disease or cancer at baseline in 1986. Snoring was assessed through mailed questionnaires at baseline. During 8 years of follow-up, we documented 7,622 incident cases of hypertension and 1042 incident cases of major CVD events confirmed by medical records (644 coronary heart disease and 398 stroke). After adjusting for age, body mass index, alcohol use, and other covariates, snoring was associated with a significantly higher risk of hypertension (Relative Risks (RRs) were 1.29 [95% CI 1.22-1.37] for occasional snoring and 1.55 [1.42-1.69] for regular snoring). Compared with non-snorers, the age-adjusted RRs of CVD were 1.46 [1.23-1.74] for occasional snorers, and 2.02 [1.62-2.53] for regular snorers. After further adjustment for other covariates, the positive association was attenuated but remained statistically significant (RRs were 1.20 [1.01-1.43] for occasional snorers and 1.33 [1.06-1.67] for regular snorers). These data suggest that snoring is associated with significantly increased risk of hypertension and CVD in women, independent of age, smoking, body mass index, and other cardiovascular risk factors. While further study is needed to elucidate the biological mechanism underlying this association, snoring may help clinicians identify individuals at higher risk for CVD.
P16 Number of menstrual cycles is associated with coronary heart disease mortality in postmenopausal women
Miriam J Kleijn, Yvonne T Schouw, Petra H Peeters, Jan D Banga, Andre L Verbeek, Yolanda Graaf, Department of Epidemiology, Nijmegen University, Nijmegen, Netherlands, Division of Internal Medicine, Academic Hospital Utrecht, Utrecht, Netherlands, Julius Center, Utrecht University Medical School, Utrecht, Netherlands
Background and methods: It is believed that estrogen protects women from coronary heart disease. We investigated the association between pre-menopausal endogenous estrogen exposure and coronary heart disease mortality in postmenopausal women. We studied a cohort of 9450 postmenopausal women living in Nijmegen, Netherlands, aged 35-65 years at enrolment (1975) in a population based breast cancer screening project. The median follow-up was 20.5 years (maximum 21.9 years). At baseline the participants reported data on all reproductive factors. The number of menstrual cycle-years as a measure of unopposed endogenous estrogen exposure, was calculated from these data. The vital status of these women was registered until december 1998. Primary cause of death was collected at the Dutch Central Bureau of Statistics. A Cox-proportional hazards model was used to analyse the data with menstrual cycle years as independent predictor and coronary heart disease mortality (ICD-9 410-414) as endpoint.
Results: In total, 2439 women died, of whom 496 died of coronary heart disease. 1482 Women were withdrawn alive at the time of moving out of the study-area. The risk of coronary heart disease mortality was lower for women with more than 35 cycle-years compared to those with 25 cycle-years or less; age-adjusted hazard ratio 0.80 (95% confidence interval 0.68-0.96). Adjustment for other potential confounders did not alter this risk estimate.
Conclusion: The risk for coronary artery mortality in postmenopausal women is associated with the unopposed endogenous estrogen-exposure, measured as number of cycle-years, during pre-menopausal life. These results support the hypothesis that long exposure to unopposed endogenous estrogen protects against coronary heart disease.
P17 Hormone replacement therapy and the risk of stroke.
Rozenn N Lemaitre, Susan R Heckbert, Bruce M Psaty, Nicholas L Smith, Robert C Kaplan, David L Tirschwell, W T Longstreth, University of Washington, Seattle, WA
We conducted a population-based case-control study at Group Health Cooperative of Puget Sound (GHC) to assess the association of hormone replacement therapy with the risk of stroke. Cases were all women GHC enrollees with an incident fatal or non-fatal stroke between 7/89 and 12/95 (493 ischemic and 121 hemorrhagic strokes). Controls were a stratified random sample of GHC female enrollees without stroke, frequency matched to cases on age and calendar year (n=1728). Information about cardiovascular risk factors and stroke events was obtained by medical record review. Medication use was determined from GHC computerized pharmacy data. Eighteen percent of ischemic cases, 22% of hemorrhagic cases, and 24% of controls were current users of estrogen with or without progestin. In unadjusted analyses, current use of estrogen alone and current use of estrogen plus progestin were associated with decreased risk of ischemic stroke compared to never use (odds ratios: 0.56 [0.37-0.86] and 0.72 [0.53-0.98] respectively). However, these associations disappeared after adjustment for age, year, diabetes, smoking, systolic blood pressure, and cardiovascular disease (odds ratios: 0.88 [0.55-1.41] and 0.84 [0.60-1.18] respectively). There was no association of the duration of estrogen use with ischemic stroke, and the associations did not differ in diabetics, hypertensives, or current smokers. Current use of estrogen alone was not associated with hemorrhagic stroke (adjusted odds ratio: 0.92 [0.54-1.58]), while current use of estrogen plus progestin was associated with a non-significant decreased risk (adjusted odds ratio: 0.53 {0.23-1.21]). These results suggest a lack of association of hormone replacement therapy with both ischemic and hemorrhagic stroke.
P18 Parity is Associated with Carotid Artery Plaques and Reduced HDL-Cholesterol Levels: Results from the Rotterdam Study
Karin H Humphries, Iris Westendorp, John J Spinelli, Ronald G Carere, A Hofman, Jacqueline CM Witteman, Erasmus University, Rotterdam, Netherlands, St. Paul's Hospital, Vancouver, BC, Canada, University of British Columbia, Vancouver, BC, Canada
Physiologic changes in lipids and glucose metabolism during pregnancy have been postulated to increase subsequent risk of cardiovascular disease (CVD). The relationship of parity to carotid atherosclerosis was investigated in the Rotterdam Study, a prospective cohort study in subjects aged 55+ years.
Parity and CVD risk factor data were collected on 4,878 women. Carotid plaques were measured by duplex ultrasound. Logistic regression estimates for risk of presence of carotid artery plaques were controlled for age, smoking status, hypertension, and lipid levels. ANOVA was used to examine the relationship between parity and lipids, adjusted for age. Relative to nulliparous women, parous women had 34% greater risk of carotid plaques and risk increased with increasing parity. Parity was inversely associated with HDL-cholesterol levels and positively associated with total/HDL cholesterol ratio.
Conclusion: independent of known CVD risk factors, parity is associated
with an increased risk of carotid artery plaques. Furthermore,
HDL-cholesterol levels decline with increasing parity and
this effect remains after menopause.
|
P19 Erectile Dysfunction and Cardiovascular Risk Factors: Prospective Results from the Massachusetts Male Aging Study.
Henry A. Feldman, Catherine B. Johannes, Carol A. Derby, Ken P. Kleinman, John B. McKinlay, New England Research Institutes, Watertown, MA
We examined prospectively the relation between erectile dysfunction (ED) and risk of coronary heart disease (CHD) in a cohort of middle-aged men, on the premise that progressive arterial compromise contributes to both conditions. A random sample of Massachusetts men, aged 40-70 at baseline, were interviewed at home in 1986-89 (n=1709) and again in 1996-97 (n=1156, 77% of those eligible). The interviewer took body measures, medical history, 1-yr diet recall, behavioral data, psychological scales, and nonfasting blood samples. The subject completed a sexual activity questionnaire in private. For this analysis we used data from 572 non-diabetic men with complete data on ED and baseline CHD risk factors and no CHD or related medications at either time point.
Incidence of moderate or complete ED was 16%. Multiple logistic
regression showed the odds of incident ED were significantly elevated
in cigarette smokers (adjusted incidence 26%, OR=2.12, 95% CI
1.18-3.80, p=0.01), in overweight men (BMI28 kg/m2)
(adjusted incidence 23%, OR=1.92, 95% CI 1.18-3.11, p=0.008), and in
men with greater animal fat intake (OR=1.09 /1% of energy, 95% CI
1.02-1.16, p=0.01). Onset of ED was not predicted by serum lipid or
DHEAS levels, hypertension, alcohol consumption, dietary
cholesterol or fiber intake, physical inactivity, or anger
expression. All analyses were controlled for age (OR=2.55 /10
yr, 95% CI 1.87-3.48, p<0.0001) and serum testosterone (OR=0.76 /SD,
95% CI 0.59-0.97, p=0.03). ED onset was related to a composite CHD
risk score calculated from baseline risk factors (OR=1.97
/log10 CHD risk, 95% CI 0.99-3.90, p=0.05), and ED risk
exceeded CHD risk when both lay between 0-5%. These results suggest
that ED shares some behaviorally modifiable determinants with CHD and
may have value as a sentinel condition.
P20 Carotid artery elasticity varies directly with educational achievement level: The ARIC Study, 1987-1992.
Rebecca Din-Dzietham, Duanping Liao, Ana Diez-Roux, Kay Paton, Andy Brown, Mercedes Carnethon, George Howard, F J Nieto, Herman A Tyroler, Bowman Gray School of Medecine Departement of Public Health, Winston-Salem, NC, Columbia Presbyterian Medical Center, New York, NY, John Hopkins University School of Hygiene and Public Health, Baltimore, MD, Penn State University, Hershey, PA, University of Mississippi Medical Center, Jackson, MS, University of North Carolina at Chapel Hill, Chapel Hill, NC, University of North Carolina School of Public Health, Chapel Hill, NC
Education, along with other indices of SES, appears to be strongly and inversely associated with the intima-media thickness (IMT) of the common carotid artery (CCA) in the ARIC Study. We extended the ARIC study of preclinical atherosclerosis by evaluating the cross-sectional association of education with elasticity of the CCA. This study included 10,091 black and white men and women, age 45-64, who were free of clinical CHD and stroke/TIA. Arterial elasticity was assessed at baseline by the pulsatile arterial diameter change (PADC) derived from phase-locked echotracking, as the average difference of systolic and diastolic adventitial diameters, over 5 cardiac cycles. The smaller the PADC, the stiffer the artery. Educational was categorized into six levels: grade school, high school without degree, high school graduate, vocational school, some college, and graduate-professional school. Education attainment was directly associated with PADC. The least-squares mean PADC (SE) in microns, adjusted for age, height, diastolic diameter and blood pressure, pulse pressure linear and squared, ethnicity-center, gender, and smoking status in successively higher education strata were: 391(5), 393(4), 395(2), 396(4), 399(3), 408(4) (p for linear trend <0.02). To our knowledge, this is the first time such an association has been reported. If arterial dilatation/elasticity impairment precedes arterial wall thickening in the atherosclerotic process, as recent studies on endothelial dysfunction suggest, then these results indicate that low SES may lead to early pathophysiologic changes, an effect that appeared to be mediated by known risk factors such as obesity, lipids, glucose, fibrinogen, hypertension, diabetes, and carotid thickening.
P21 Life stress, social support, and psychological resources: correlation with development of cardiovascular disease in the San Antonio Heart Study
Sharon P. Gaskill, Ken Williams, James P. Burke, Ravi Duggirala, Michael P. Stern, Helen P. Hazuda, University of Texas Health Science Center at San Antonio, San Antonio, TX, University of Texas Health Science Center at San Antonio, San Antonio,, TX
We examined the association between baseline psychosocial factors and 8-year incidence of cardiovascular disease (CVD) in the San Antonio Heart Study, a population-based, longitudinal study of CVD in San Antonio, TX. CVD was defined as presence of angina by Rose questionnaire; self-reported heart attack, stroke, or heart surgery; or CVD death. Of the individuals with baseline clinic exams from 1984 to 1988 who returned to 8-year followup exams, 1645 (controls)had CVD at neither baseline nor followup, and 202 (cases) had no CVD at baseline but had developed CVD by followup. Case-control logistic regression analyses, controlling for age, gender, ethnic group, and socioeconomic status, were used to evaluate the roles of life stresses, social support, and psychological resources in the development of CVD. Odds ratios are reported for the risk differences between the 75th and 25th percentiles of the independent variables.
CVD incidence was associated with increased job hours worked (OR 1.14*), marital stress (1.37**), and overall daily stress (1.40**). Controls' spouses were significantly more caring/mature (0.74*) and appreciative/supportive (0.75**) than cases'. Self-esteem (0.73**), mastery (0.76***), and a sense of having "a great deal of" control over one's health (0.65*) were all significantly higher among controls than cases.
Compared with controls, cases had higher job, marital, and daily stress levels, and lower spouse support, self-esteem, and mastery. The persistence of these differences, even after controlling for age, gender, ethnic, and socioeconomic differences, suggests that both external stressors and support, and internal psychological resources, are significantly related to the development - and prevention - of CVD.
*p<.05, **p<.01, ***p<.001
P22 Income inequality and cardiovascular disease risk factors
Ana V Diez-Roux, Columbia College of Physicians and Surgeons, New York, NY; Bruce G Link, Mary E Northridge, Columbia University, New York, NY
Ecologic studies suggest that income inequality is associated with increased mortality, even after accounting for average income, but the mediating mechanisms remain to be determined. Using data from the CDC's Behavioral Risk Factor Surveillance System, we examined the cross-sectional associations between state income inequality (the Robin Hood (RH) index from the 1990 Census) and four cardiovascular risk factors (body mass index (BMI), history of hypertension, sedentarism, and smoking) in 44 U.S. states. The hypotheses tested included: income inequality is related to levels of risk factors, the effects of income inequality on risk factor levels are greater in persons of lower income than in those of higher income, and associations of income inequality with risk factors persist after adjustment for individual level income. Multilevel models (including state level and individual level variables) were used. For hypertension, sedentarism, and BMI, state inequality was associated with increased risk factor levels among persons of low income (annual income <$25,000), with associations persisting after adjustment for individual-level income (adjusted odds ratios (OR) and 95% Confidence Limits (CL) per 10 unit increase in RH index in women: hypertension OR 1.6 CL 1.2-2.2; sedentarism OR 2.1 CL 1.3-3.4; mean difference in BMI (kg/m2)0.95 CL 0.2-1.7). Conversely, inequality was positively associated with smoking at higher income levels (>$25,000)(in women OR 1.5 CL 1.0-2.2), but associations were weaker or non-existent at lower income levels. Similar patterns were observed in both genders but associations were weaker and not statistically significant in men. Inequality per se appears to be related to risk factor levels. For three of the four risk factors, low income persons may be more vulnerable to the effects of inequality than higher income persons.
P23 Results of a Clinical Trial of Exercise on Long-Term Survival in Myocardial Infarction Patients
Joan M. Dorn, John Naughton, Dai Imamura, Maurizio Trevisan, On behalf of the NEHDP staff, State University of New York at Buffalo, Buffalo, NY
This study examined whether participation in a supervised exercise program improved the 19-year survival in 30-64 year old male myocardial infarction (MI) patients. The men (n=651) were participants in the National Exercise and Heart Disease Project, a three-year multi-center randomized clinical trial conducted in the U.S.(1976-79). The treatment group (n=315)exercised for eight weeks in a laboratory. Thereafter, they jogged, cycled or swam in a gym/pool setting guided by an individualized target heart rate. They were encouraged to attend 3 days/week,one hour/session. The control group (n=319) was to maintain normal routines, but not participate in any regular exercise program. Participants were followed until their death or 12/31/95. Vital status was determined for 634 men (97.4%). 162(51.4%) exercise group members and 150(47.0%) controls died; 104(33.0%) and 109(34.2%) from CVD respectively. Cox's proportional hazards analysis revealed the following all-cause mortality risk estimates [95% confidence intervals (95%CI)]in the exercise group compared to controls after an average follow-up of three 0.69(0.39-1.25), five 0.84(0.55-1.28), ten 0.95(0.71-1.29), 15 1.02(0.79-1.32) and 19 years 1.09(0.87-1.36). CVD mortality risk estimates (95% CI)for the same follow-up periods were 0.73(0.37-1.43), 0.98(0.60-1.61), 1.21(0.79-1.60), 1.14(0.84-1.54), and 1.16(0.88-1.52). These findings indicate that exercise program participation resulted in non-significantly reduced mortality risks early in the follow-up. Benefits diminished as time since participation increased, suggesting protective mechanisms associated with exercise may be of a short-term nature. The public health implications are that in MI survivors, as in healthy individuals, the effects of exercise cannot be stored and exercise must be maintained to offer survival benefits.
P24 Relationships of Physical Fitness and Body Weight with CVD Risk Factors in Young Adults: CARDIA.
Cora E. Lewis, O. Dale Williams, Steve Sidney, Albert Oberman, Kaiser, Oakland, CA, University of Alabama, Birmingham, AL, University of Alabama at Birmingham, Birmingham, AL
Is physical fitness more important than body weight in decreasing CVD
risk? We examined the relation of fitness and body weight with CVD risk
factors in blacks and whites. At baseline (1985-6), men (N=1914) and
women (N=2271) aged 18-30 years had treadmills and measures of body
mass index (BMI, kg/m2), blood pressure (BP), fasting
lipids and insulin levels; measures were repeated at Year 7 with the
same protocols. At baseline, insulin and LDL-C levels were higher with
lower maximum treadmill time (TRT) and higher BMI in both men and
women; HDL-C levels were lower with higher BMI but were not
independently related to TRT. The table shows mean 7-year change in
risk factors by TRT change: decrease (>60 sec), increase ( 60 sec),
or stable (±60 sec); and weight change: decrease (>2.5 kg), increase
( 2.5 kg), or stable (±2.5 kg) in women. Results were similar in
men. Relationships with SBP were inconsistent. Change in weight
but not in fitness was associated with significant change in lipids and
insulin. To the extent that TRT is a precise measure, fitness is not
more important than weight control in relation to these risk factors in
CARDIA.
|
P25 Is there an association of fitness level with fibrinogen level in children independent of obesity? The Columbia University BioMarkers Study.
Carmen R Isasi, Sarah C Couch, Thomas Starc, Lars Berglund, Richard J Deckelbaum, Steven Shea, Columbia University, New York, NY, University of Cincinnati, Cincinnati, OH
In recent years plasma fibrinogen has emerged as a new risk factor for cardiovascular disease in adults. Relatively little is known about correlates of increased fibrinogen earlier in life. We evaluated the association between fitness and fibrinogen level in 193 children 4 to 25 years old (mean age=11.1 years, SD=3.8); 68% were Hispanic and 46% male. Fitness level was measured using a submaximal, non-effort-dependent, treadmill test (PWC170 protocol). Mean(S.D)values for plasma fibrinogen, body mass index (BMI), fitness level and resting heart rate were 278.2(48.6)mg/dl, 22.2(5.3)kg/m2, 22.5(8.9)% grade and 81.7(13.7)bpm, respectively. Boys had lower levels of fibrinogen and higher levels of fitness than girls but did not differ in BMI. As previously reported, the Pearson correlation coefficient for fibrinogen with BMI was positive(r=.17, p<0.05). Fitness level was inversely associated with fibrinogen (r=-.24, p<0.001). Fibrinogen levels showed a graded inverse relationship with tertiles of fitness (p for linear trend < 0.001). In multivariate analysis, after adjusment for age, sex, race/ethnicity, BMI and presence of the A allele in the 455 position of the ß-fibrinogen promoter gene, fitness level remained inversely associated with fibrinogen level (ß=-1.3, 95% CI=-2.3,-0.35). Consistent with these findings, resting heart rate was correlated with fibrinogen (r=.18, p<0.05) and fibrinogen level increased with tertiles of resting heart rate (p for linear trend = 0.002). In multivariate analysis, after adjustment for the same covariates, resting heart rate was significantly associated with fibrinogen level (ß=0.82, 95% CI=0.17,1.5). These findings indicate that there is an association between fitness and fibrinogen level independent of obesity in children.
P26 Diet-Related Decrease in Waist to Hip Ratio Predicts Fall in Plasma Fibrinogen
Francisco Lopez-Jimenez, Y. Wady Aude, Gervasio A Lamas, Gerardo Rojas, Eric H Lieberman, Arthur S Agatston, Mount Sinai Medical center, Miami Beach, FL
Background: High fibrinogen is an independent risk factor for cardiovascular disease and has been associated with obesity. However, modest weight loss has failed to decrease fibrinogen levels.
Methods: To assess the effect of waist to hip ratio (WHR) change after diet on plasma fibrinogen, we studied 42 overweight patients who participated in a randomized trial comparing a relatively high-fat, low carbohydrate diet with the National Cholesterol Education Program Step II Diet over a 3-month period. Fasting fibrinogen level was measured using the Von Clauss method (clotting rate assay). Patients were asked to avoid changes in their level of physical activity. Subjects with recent acute medical conditions were not included.
Results: The mean age of participants was 45.9±10 years and 64% were female. Baseline weight was 98.4±19 Kg, body mass index (BMI) 35.3±5, WHR 0.89±0.1, and fasting fibrinogen 361±59 mg/dl. Hypertension was present in 36%, smoking in 10% and diabetes in 7%. Independent baseline correlates for fibrinogen were BMI, p=0.002; WHR, p=0.04; and weight, p=0.04. WHR improvement after diet was positively associated to a decrease in fibrinogen level, r=0.34, p=0.03. However, neither change in BMI nor weight loss correlated with a change in fibrinogen level, p=0.5 for each. A linear regression analysis including BMI change, WHR change and weight loss, confirmed that only WHR change was significantly associated with a decrease in fibrinogen level, p=0.009.
Conclusions: Improvement in the distribution of body fat, rather than total weight loss is associated with a significant decrease in fibrinogen.
P27 Obesity and Its Relationship to Physical Activity and Inactivity During Adolescence
Sue Y. S. Kimm, Nancy W. Glynn, Eva Obarzanek, on behalf of the NGHS Collaborative Research Group, National Heart, Lung and Blood Institute, Bethesda, MD, University of Pittsburgh, Pittsburgh, PA
Obesity is seen more in sedentary than in active people and TV viewing has been implicated in childhood obesity. This report examines physical activity, inactivity, and caloric intake in relation to obesity during adolescence. Three hundred ten black and 378 white girls from the NHLBI Growth and Health Study (NGHS) with Caltrac (CLT) readings, weekly hours of TV viewing (TV), and 3-day activity and food diaries (which provided activity score (AD), minutes of sedentary activity (SED) and caloric intake) at ages 11-12 (Y3) and 13-14 (Y5) were studied. The correlation between the measures of activity (AD and CLT) and inactivity (SED and TV) were examined with Spearman coefficients. CLT was significantly correlated with AD at Y3 (r=0.20, p<0.01) and Y5 (r=0.24, p<0.01) and negatively correlated with TV (r=-0.10, p<0.01 for both). CLT was weakly (r=0.08, p=0.03) related to SED in Y5. AD was not related to TV nor SED. TV and SED were not significantly correlated. CLT and AD were both negatively correlated with adiposity (sum of triceps, subscapular and suprailiac skinfolds, SSF) at Y3 (r=-0.10, p<0.01 for both) but not at Y5. TV, but not SED, was correlated (r=0.19, p<0.01 for both) with SSF. Regression models showed that at ages 11-12, TV was positively (p<0.01) and AD inversely (p<0.01) related to SSF. CLT was inversely but weakly (p=0.07) associated with SSF. At ages 13-14, TV was positively and calories negatively (p<0.01 for both) associated with SSF. Caltrac is viewed as the gold standard for activity assessment: among the measures examined with CLT, AD had the highest correlation, followed by TV. However, TV rather than AD, was consistently associated with adiposity. Whether this finding is due to greater accuracy of TV as a surrogate of inactivity or as a marker of environmental factors yet unidentified, needs further examination.
P28 Body Mass Index, Leptin, and Tumor Necrosis Factor Receptor and Risk of Insulin Resistance Syndrome among Middle Aged and Older US Men
Nain-Feng Chu, Jie Yu, Walter C Willett, Eric B Rimm, Harvard School of Public Health, Boston, MA
Obesity plays a central role in the development of insulin
resistance, diabetes and CHD. In epidemiologic studies of obesity,
simple measures such as the body mass index (BMI) are used to identify
at risk individuals. Among older populations, BMI may not be a good
indicator of CHD because it cannot accurately represent
age-related changes in body composition. Adipocyte tissue produces a
variety of molecules such as leptin and tumor necrosis factor-alpha
(TNF-) which may regulate adipocyte mass and induce insulin
resistance. Circulating soluble TNF-R75 (sTNF-R75), a receptor for
TNF, is stable, reproducible and reflects TNF- activity.
Among a random sample of 468 male health professionals, 47-83 years of
age, we calculated a composite score for insulin resistance syndrome
(IRS) based on quintile ranks of participant's TG, HDL-C, and SBP. In
multivariate linear regression models we examined the
association between measures of adiposity and IRS controlling for age,
smoking, and physical activity. Among men less than 65 years of age,
BMI (p<0.001) and leptin (p<0.001) were strongly associated with IRS.
After simultaneously controlling for both, the association
for leptin was no longer significant, but BMI continued to be
predictive of risk. sTNF-R75 was not associated with IRS in this group.
However, among older men only sTNF-R75, and not BMI nor leptin, was
significantly associated with IRS, but this association was limited to
men with a BMI > 25 kg/m2 (p<0.05, partial
r2 =0.062). Our data suggest that among older men neither
BMI nor plasma leptin levels are sufficient to identify those at risk
of CHD. Further, our data suggest the possibility that the TNF receptor
is a better marker of IRS and provides further evidence that TNF-
plays a role in the development of IRS.
P29 Childhood onset of overweight and subsequent CVD risk factor levels near age 40 years: The Fels Longitudinal Study
Roger M. Siervogel, Wayne A. Wisemandle, Shumei S. Guo, Wm. Cameron Chumlea, Bradford Towne, Wright State University School of Medicine, Yellow Springs, OH
Our aim was to examine the consequences of childhood overweight,
defined as body mass index (BMI, kg/m2) >85th age/sex
NHANES II percentile, or adult onset of overweight, defined as BMI
>25, on the subsequent levels of other cardiovascular
disease (CVD) risk factors in middle age. We used data from over 15,000
serial examinations of selected participants in the Fels Longitudinal
Study to address this question. Methods: Adult variables
included total body fat (TBF, kg), fat-free mass, and percent body fat
from hydrostatic weighing; triglycerides, total
cholesterol (C), low-density lipoprotein-C, and
high-density lipoprotein-C (HDL-C, mg/dl); and systolic
(SBP, mmHg) and diastolic blood pressure. Lifetime
data were used to identify three groups: those never overweight (OW) as
defined above, those with onset of OW in childhood (<18 yrs), and
those with adult onset of OW (18 yrs). Additional inclusion criteria
were at least one adult visit after age 24 years and, in the OW groups,
maintenance of OW in at least 75% of all visits subsequent to
onset. Participants not meeting these criteria were not included in
these analyses. Age of onset was the age at which OW was first
measured. Analyses of data (e.g., ANOVA) from the examination
nearest to age 40 years were performed. Results: As
exemplified in the table for males, mean levels of most variables
in both sexes, showed a pattern of increasing CVD risk corresponding to
duration of OW, i.e., childhood onset OW worse than adult onset OW
worse than never OW. Conclusion: The earlier the onset of OW
the more serious the consequences in terms of increased levels of risk
factors for obesity and cardiovascular disease.
|
P30 Association of Plasma Leptin Level and Obesity on Insulin Resistance Syndrome among School Children in Taiwan — The Taipei Children Heart Study
Nain-Feng Chu, Dan-Jiang Wang, Shih-Ming Shieh, Eric B. Rimm, Dept. of Epidemiology, Harvard School Public Health, Boston, MA, Harvard School Public Health, Boston, MA, Tri-Service General Hospital, Taipei, Taiwan ROC
Leptin, an adipose tissue-derived product of the obesity (ob) gene, may regulate appetite and could be an important determinant of insulin resistance, diabetes, and CHD. The purpose of this study was to evaluate the association of plasma leptin with obesity and insulin resistance syndrome (IRS) among school children in Taiwan.
We randomly selected 1264 children (617 boys and 647 girls) aged 13.3 years (12-16 y) from Taipei. Obesity measurements included body mass index (BMI) and waist-to-hip circumference ratio (WHR). We calculated an IRS summary score by adding the quartile ranks from the distribution of systolic BP, TG, HDL-C and insulin levels of each subject. A higher score corresponds to higher levels of BP, TG, and insulin and lower levels of HDL-C.
Boys had a higher BMI and WHR, BP, and IRS score and lower leptin, insulin, TG, and HDL-C than girls. BMI, WHR and plasma leptin levels were significantly associated with the IRS summary score (and each of it s components) even after adjusting for age, cigarette smoking, alcohol drinking, and puberty development. Among boys and girls, after adjusting for WHR, plasma leptin levels were still significantly associated (P<0.001) with the IRS score. The association between plasma leptin level and the IRS score was somewhat attenuated in boys (P<0.002) and girls (P<0.06) after adjusting for BMI. The final model that included the standard covariates, BMI, leptin, but not WHR, was most the predictive of the IRS summary score.
Insulin resistance syndrome in childhood characterized by hyperlipidemia, hypertension, and hyperinsulinemia may be an early marker of CHD risk. Our results suggest that BMI and leptin, in combination, are the best predictive markers of insulin resistance syndrome among school children in Taiwan.
P31 Genetic and Environmental Influences on Waist-Hip Ratio and Waist Circumference in an Older Swedish Twin Population
Tracy L. Nelson, George P. Vogler, Nancy L. Pedersen, Toni P. Miles, Karolinska Institute, Stockholm, Sweden, Pennsylvania State University, University Park, PA, University of North Carolina, Chapel Hill, NC, University of Texas Health Sciences Center-San Antonio, San Antonio, TX
The objective of this study was to investigate the genetic and environmental influences on waist-hip ratio (WHR) and waist circumference (WC) measurements in males and females. Measurements were taken from 1989-91 as part of the Swedish Adoption/Twin Study of Aging (SATSA). The SATSA sample contains both twins reared together as well as twins reared apart. We had 322 subjects (50 monozygotic (MZ) and 82 dizygotic (DZ) male pairs; 67 MZ and 123 DZ female pairs); age range 45-85 years (average age 65 years). Measurements included WHR, WC and body mass index (BMI).
Using maximum likelihood quantitative genetic analyses, we found in males, additive genetic effects accounted for 28% of the variance in WHR and 46% of the variance in WC. In females, additive genetic effects were found to account for 48% of the variance in WHR and 66% of the variance in WC. The remaining variance in males was attributed to unique environmental effects (WHR-72%; WC-54%) and in females the remaining variance was attributed to unique environmental effects (WHR-46%; WC-34%) and age (WHR-6%). When BMI was added into these models it accounted for a portion of the genetic and environmental variance in WHR and over half the genetic and environmental variance in WC.
In conclusion, there are both genetic and environmental influences on WHR and WC, independent of BMI in both males and females, and the differences between the sexes are significantly different.
P32 Body Mass Index (BMI) in Middle Age and Health Care Costs in Older Age: The Chicago Heart Association Detection Project in Industry (CHA) Study
Kiang Liu, Martha L Daviglus, Dan Garside, Philip Greenland, Lynn Lowe, Alan Dyer, Jeremiah Stamler, Northwestern University, Chicago, IL
This study examines relationships between BMI in middle-age and subsequent health care costs at age 65 years or older, utilizing data from the CHA study. The sample consists of 7030 men and 6739 women in the CHA study ages 40-64 at baseline (1967-73) who in 1984-94 were or became eligible for Medicare. U.S. Health Care Finance Administration (HCFA) inpatient and outpatient charge data were used to estimate and compare average annual costs for cardiovascular disease (CVD) and total health care for various baseline BMI levels. Average annual costs for both CVD and total health care significantly increase with higher baseline BMI for both men and women, with adjustment for baseline age, number of cigarettes, serum cholesterol, SBP, ethnicity, education, diabetes, year of Medicare eligibility and ECG abnormalities.
These findings Suggest that prevention of obesity early in life is
important not only for reducing risks of
hypercholesterolemia, hypertension, diabetes,
and CVD but also for reducing health care costs among survivors into
older age.
|
P33 Estrogen Status Versus Androgenicity in Cardiovascular Disease Risk Factors in Mid-Aged Women
Sybil L Crawford, Catherine B Johannes, Sonja M McKinlay, New England Research Institutes, Watertown, MA
Past research has found an association in women between cardiovascular disease (CVD) risk factors and sex hormone binding globulin (SHBG), an indirect indicator of androgenicity. Other studies have examined changes in CVD risk factor levels during menopause, when estrogen levels decline. This study compares directly the associations of SHBG and estrogen – estrone (E1) and estradiol (E2) – with CVD risk factors in mid-aged women. Data were from a longitudinal population-based study of 372 Caucasian women initially aged 51-61. Observations concurrent with hormone replacement therapy were omitted from analyses. CVD risk factors included high-density lipoprotein cholesterol, total cholesterol, triglycerides, Apoprotein-A-I, Apoprotein B, systolic and diastolic blood pressure, and body mass index. A log transformation was applied to SHBG, E1, E2, triglycerides, and body mass index due to skewness. Correlations of the hormones with the CVD risk factors were computed before and after adjustment for age, smoking, exercise, and ethanol consumption, accounting for within-woman correlation. All three hormones were associated with a better risk profile. Relationships with all CVD risk factors were stronger for SHBG than for E1 or E2. Correlations ranged (in absolute value) from 0.11 to 0.42 for log SHBG, versus 0.02 to 0.20 for log E1 and 0.004 to 0.08 for log E2. Adjustment for age and lifestyle factors had little impact on the correlations. Results were similar using menopause status defined based on menstrual bleeding patterns instead of E1 or E2, and for the six-year change in hormones versus corresponding change in CVD risk factors. These findings suggest that SHBG, a marker of androgenicity, is a better predictor of CVD risk factors in mid-aged Caucasian women than are measures of menopausal status.
P34 High "Normal" albuminuria is associated with increased cardiovascular risk
Hans L Hillege, Wilbert M Janssen, Annette A Bak, Harry G Crijns, Wiek H van Gilst, Dick de Zeeuw, Paul E de Jong, University Hospital Groningen, Groningen, Netherlands, University of Groningen, Groningen, Netherlands, University of Utrecht, Utrecht, Netherlands
Background: Microalbuminuria(MA) has been recognised to be an early marker of cardiovascular disease. The commonly used cut-off values for albumin concentration for untimed samples are between 20 and 200 mg/L (30-300 mg/24 hr). We tested whether lower values of albuminuria are already associated with known cardiovascular riskfactors.
Methods: In the PREVEND study (Prevention of REnal and cardioVascular ENd stage Disease) we screened 80.000 subjects aged 30-75 in the city of Groningen. Participants collected morning urine samples and filled in a short questionnaire regarding presence of diabetes mellitus (DM), use of antihypertensive medication (AHM), lipid lowering drugs, family history of CVD, smoking (SM), history of myocardial infarction and cerebrovascular accident (CVA).
Results: The table shows the distribution of MA in the first 35,652 subjects in four "normoalbuminuric" and one conventional defined microalbuminuric (>20 mg/L) group and the corresponding age and sex adjusted odds ratios of cardiovascular risk factors (prevalences in parentheses).
Conclusion: Microalbuminuria is associated with
smoking, diabetes and the use of antihypertensive drugs in ranges 10-20
mg/L. This suggests that the criteria for MA should be
re-evaluated.
|
P35 Urinary Albumin Excretion and Carotid Intima-Media Thickness in the Elderly: The Cardiovascular Health Study
John H. Summerson, Joseph C. Konen, Ronny A. Bell, Gregory L. Burke, John Port, Trevor J. Orchard, Carolinas Medical Center, Charlotte, NC, Johns Hopkins University, Baltimore, MD, University of Pittsburgh, Pittsburgh, NC, Wake Forest University School of Medicine, Winston-Salem, NC
Urinary albumin excretion below that detectable by
routine dipstick (microalbuminuria, MA) is associated with
cardiovascular disease (CVD) morbidity and mortality in
patients with diabetes mellitus and in the general population. MA has
also been found to be related to an earlier stage of the
atherosclerotic process manifested as carotid artery intima-media
thickness (IMT) in middle-aged diabetic and nondiabetic individuals.
However, associations between MA and carotid artery IMT in the elderly,
or between MA and internal carotid artery (ICA) IMT have not been
established. Three hundred seventy-six subjects aged 65 years
(Mean age= 72±4.5 yrs; 61% female; 94% white; 20% diabetes) without
macroalbuminuria from the Forsyth County, NC clinic site of
the Cardiovascular Health Study participated in this
cross-sectional investigation. A urinary albumin excretion
ratio (UAER; g albumin / g creatinine) was
determined for each participant from an overnight urine collection one
year after B-mode ultrasound measurement of common carotid artery (CCA)
and ICA IMT. Variables regressed in stepwise,
multivariate analysis against carotid artery
IMT included sex, race, age, diabetes, smoking, waist to hip ratio,
fasting insulin, lipids and lipoproteins, and hypertension. In the
final model, log-transformed UAER was positively independently
associated with both CCA (p=.029) and ICA (p=.032) IMT. CCA and ICA IMT
increased by 0.04 and 0.11 mm, respectively, per unit change in
log-UAER. Addition of prevalent CVD to the model did not significantly
change these associations. An interaction term of (log-UAER x diabetes
status) was not significant for either CCA or ICA IMT. Urinary
albumin excretion at levels below that indicative of renal
dysfunction is associated with carotid artery IMT and, thus,
atherosclerosis in the elderly.
P36 Diabetes Mellitus, Subclinical Cardiovascular Disease and Risk of Incident Cardiovascular Disease and All-Cause Mortality
Lewis H Kuller, Priscilla Velentgas, Joshua Barzilay, Norman J Beauchamp, Daniel H O'Leary, Peter J Savage, DECA/NHLBI, Rockville, MD, Johns Hopkins Radiology, Baltimore, MD, Kaiser Permanente of Georgia, Tucker, GA, Tufts-New England Medical Center, Boston, MA, University of Pittsburgh, GSPH, Pittsburgh, PA, University of Washington, Seattle, WA
We tested the hypothesis that the presence of subclinical cardiovascular disease (CVD) was the primary determinant of subsequent risk of clinical CVD and total mortality among diabetics in the Cardiovascular Health Study (CHS) without a prior history of clinical CVD. CHS participants, all aged 65 and over, were followed for 6.5 years (median). Diabetic participants at baseline were classified as prevalent (n=722), based on self-report or medication use, or new (n=621), based on fasting and 2-hour glucose challenge. Subclinical CVD was established from ankle/brachial blood pressure, carotid artery stenosis and wall thickness, EKG and echocardiographic abnormalities and the Rose Angina and Claudication Questionnaire. Total mortality varied from 9.8/1000 for normoglycemic individuals with no subclinical CVD to 17.8/1000 for diabetics with no subclinical CVD to 40/1000 for diabetics with subclinical CVD. Diabetics with no subclinical CVD at baseline were at much lower risk of clinical CVD than diabetics with subclinical disease. New diabetics were at much lower risk of CVD than prevalent diabetics. In multivariate Cox regression analyses that included traditional CVD risk factors, subclinical disease was a strong independent predictor of CVD mortality (RR 2.51), incident MI (RR 1.93), and incident CHD (RR=1.99) among diabetics (all statistically significant). Of other risk factors examined, only diastolic blood pressure was consistently related to risk of CVD among diabetics. Fasting glucose levels were weakly related to total mortality and cardiovascular risk, with relative risks of 1.04 to 1.09 for a 20 mg increase in blood glucose. These results show that the primary determinant of cardiovascular risk among older diabetics without a history of CVD is the presence of subclinical CVD.
P37 Diabetes and Coronary Heart Disease Mortality: Findings from The Physicians' Health Study Enrollment Cohort
Paulo A Lotufo, J.Michael Gaziano, Charles H Hennekens, Umed A Ajani, Julie E Buring, JoAnn E Manson, Brigham and Women's Hospital, Boston, MA
Diabetes Mellitus(DM)is associated with an increased risk of
coronary heart disease(CHD) mortality. Recently, epidemiologic
studies have suggested that CHD death rates among persons with diabetes
and without prior CHD are similar to those of nondiabetics with prior
myocardial infarction (MI). In a cohort of 91,285 US male physicians
aged 40-84-years at entry, we examined the comparative rates of CHD
mortality in four populations: (1)men free of both DM (either type)and
CHD(MI or angina)at baseline, (2)diagnosed DM and no history of CHD,
(3)diagnosed CHD and no history of DM, and (4)diagnosed with both DM
and CHD. During 5 years of follow-up, we documented 1,244 deaths from
CHD(ICD-9:410-4;798). Using Cox proportional hazard models (shown in
the table), we observed major differences in risk between these groups.
Compared to men with neither DM nor CHD at baseline, men with DM alone
had a nearly three-fold elevation, men with CHD and no DM a 5-fold
elevation, and men with both DM and CHD a more than 10-fold elevation
in CHD mortality. These prospective data confirm that DM is associated
with a substantial increase in CHD mortality, although the magnitude of
excess risk is smaller than that conferred by prior MI. The presence of
both DM and prior CHD, however, identifies a particularly high risk
group.
|
P38 Gender difference in the predictive role of hyperglycemia for CAD morbidity and mortality in patients with type 1 diabetes
Peter Stella, Trevor J. Orchard, Dorothy J. Becker, University of Pittsburgh, Pittsburgh, PA
Although poor glycemic control is an independent risk factor for all-cause mortality in type 1 diabetes, the predictive role of hyperglycemia for CAD mortality is not clear. To examine the role of glycemic control in CAD morbidity and mortality in type 1 diabetes, eight-year incidence data were analysed from the Pittsburgh Epidemiology of Diabetes Complications Study, a 10-year prospective study of 658 subjects with type 1 diabetes diagnosed in childhood (mean age at baseline:27.6 yrs, mean duration:19.4 yrs). CAD [clinic physician diagnosed angina or confirmed myocardial infarction (MI)]developed in 77 patients, while 25 patients died from CAD (23 MI and 2 chronic heart failure with previous MI) by hospital records and death certificates. Univariate analyses of baseline HbA1 are shown in the table below.
Cox Proportional Hazard modeling showed HbA1 to predict only female CAD
mortality (p=0.018). This gender-specific role of hyperglycemia may
explain the lack of the usual male excess of CAD mortality in diabetes
and the maintenance of such in CAD morbidity.
|
P39 Diabetes, impaired fasting glucose, and the prevalence of microvascular disease in the elderly
Susan R. Heckbert, Peter J. Savage, Joshua I. Barzilay, Vera Bittner, Nicholas L. Smith, Bruce M. Psaty, Adrian Dobs, Lewis H. Kuller, Trevor J. Orchard, Emory University School of Medicine, Altanta, GA, Johns Hopkins University, Baltimore, MD, National Heart, Lung, and Blood Institute, Bethesda, MD, University of Alabama, Birmingham, AB, University of Pittsburgh, Pittsburgh, PA, University of Washington, Seattle, WA
Objective. Much remains unknown about the prevalence of microvascular disease in elderly people with unrecognized or untreated Type 2 diabetes (DM) or impaired fasting glucose (IFG; 110-125 mg/dL). In the Cardiovascular Health Study, we examined the prevalence of renal and retinal microvascular disease and alterations in cardiac autonomic tone according to baseline diabetes status by ADA fasting criteria.
Methods. Urinary albumin excretion was measured at year 7 and retinal photography at year 9 of followup. Cardiac parasympathetic function was assessed by a measure of heart rate (HR) variability from baseline 24 hr ECG monitoring; low HR variability was defined as <=5th %ile.
Results. At baseline there were 4085 (69%) participants with No DM, 774 (13%) with IFG, 442 (8%) with unrecognized or untreated DM (New DM), and 501 (9%) with Treated DM. The mean age was 73 yrs. The Table gives age-adjusted prevalences. Findings were similar using WHO criteria, with diabetes status updated over time, with adjustment for antihypertensive use and BP, and in the original and minority cohorts.
Conclusion. Microvascular disease manifestations were
substantially more common in elderly people with IFG than in those with
No DM, and were common in both New DM and Treated DM.
|
P40 QT interval, heart rate, and carotid atherosclerosis in non-diabetic subjects. The Insulin Resistance Atherosclerosis Study (IRAS)
Andreas Festa, Pentti Rautaharju, Ralph D'Agostino Jr, Daniel H O'Leary, Marian Rewers, Leena Mykkanen, Steven M Haffner, Dept of Preventive Medicine, Univ of Colorado Medical School, Denver, CO, Dept of Public Health, Wake Forest Univ School of Medicine, Winston Salem, NC, Dept of Radiology, Tufts Univ School of Medicine, Boston, MA, EPICARE Center, Winston Salem, NC, UTHSCSA, San Antonio, TX
Both prolonged QT interval and elevated heart rate have been identified as unfavorable prognostic factors of cardiovascular morbidity and mortality, and atherosclerotic disease may be an underlying unifying patho-mechanism. The aim of this study was to investigate the association of QT interval duration and heart rate to atherosclerosis, as determined by ultrasonographic measurement of carotid intima-media thickness (IMT) in a non-diabetic, tri-ethnic population (n=939).
IMT of the common carotid artery (CCA) correlated positively with heart rate-corrected QT interval duration (r=0.15 for QT60 and QT index, r=0.13 for QTc; p=0.0001, respectively), whereas no relationship between IMT of the internal carotid artery (ICA) and QT interval was found (r=0.01, p=NS). In a multiple regression analysis adjusting for demographic variables and prevalent coronary heart disease, the association of CCA IMT to heart rate-corrected QT interval remained highly significant (p=0.0001). Adjustment for cardiovascular risk factors weakened the relationship (p=0.002, p=0.006, and p=0.002 for the association of CCA IMT to QT60, QTc, and QT index, respectively). In contrast to QT interval, there was no significant relation of heart rate to carotid IMT.
In summary, we found a significant relation of QT interval duration to carotid atherosclerosis in non-diabetic subjects that was independent of prevalent coronary artery disease and was mediated in part by cardiovascular risk factors. QT interval may therefore serve as a marker for clinically undetected atherosclerotic disease. These findings yield potential clinical implications for cardiovascular risk factor management in healthy, non-diabetic subjects with prolonged QT interval.
P41 Insulin Sensitivity and Systemic Atherosclerosis
George Howard, Richard Hamman, Marian Rewers, William Hiatt, Steve Haffner, University of Colorado Health Science Center, Denver, CO, University of Texas Health Sciences Center at San Antonio, San Antonio, TX, Wake Forest University School of Medicine, Winston Salem, NC
Atherosclerosis is a systemic disease expressed in several arterial beds. IRAS indexed atherosclerosis in 1475 participants in four arterial beds: (1 & 2) in the common and internal carotid arteries by intimal-medial thickness; (3) in the coronary arteries by prevalent coronary heart disease; and (4) in the peripheral arteries by the ankle-brachial index. Pairwise comparison of these indices showed a significant (p<0.01) association of moderate size (i.e., r2=0.1), with the significance of the association reflecting the systemic nature of atherosclerosis and the moderate association its focal nature (and measurement imprecision). Maximum likelihood factor analysis suggested at least one (p<0.0001), but not more than one (p=0.09), factor underlying these four indices. The first principal component score was used as an index of systemic atherosclerosis (SAI). The relationship between insulin sensitivity (SI from the FSIGT) and SAI was assessed after adjustment for: (1) age, ethnicity, gender and clinical center, (2) diabetes status and model 1 factors, (3) syndrome X risk factors (hypertension, BMI, HDL, and triglycerides) and model 2 factors. For those participants with SI>0, there was a significant inverse association between insulin sensitivity and atherosclerosis (p < 0.0001), that remained after adjustment for diabetes (p=0.005). However, in participants with SI = 0, the average SAI was lower than predicted from the trend in those with SI>0 (p < 0.007). About one-half of the effect of insulin sensitivity on SAI was mediated by syndrome X factors, and the association of SI and SAI was of borderline significance (p = 0.08). The impact of insulin sensitivity on SAI appeared largest in whites, intermediate in Hispanics, and unimportant in blacks. These data suggest an association of insulin sensitivity with systemic atherosclerosis, which may differ by ethnicity.
P42 Familial clustering of features of multiple metabolic syndrome with special reference to plasminogen activator inhibitor-1: The NHLBI Family Heart Study
Yuling Hong,, Mike A. Province,, Stephen S. Rich,, Paul N. Hopkins,, Roger R. Williams,, Donna K. Arnett,, James Pankow,, D.C. Rao,, University of Minnesota, Minneapolis, MN, University of North Carolina, Chapel Hill, NC, University of Utah, Salt Lake City, UT, Wake Forest University, Winston-Salem, NC, Washington University, Saint Louis, MO
A bivariate familial correlation model was used to evaluate whether plasminogen activator inhibitor-1 (PAI-1) and other components of the multiple metabolic syndrome (MMS) share any familial influences and to assess the overall heritability of PAI-1 and other components of MMS in 2673 members of 541 randomly selected Caucasian families participating in the Phase II examination of the NHLBI Family Heart Study. The maximal (univariate) heritability estimates for PAI-1 34%, fasting insulin 36%, body mass index (BMI) 55%, triglycerides 42%, HDL cholesterol 60%, systolic blood pressure (SBP) 34%, and diastolic blood pressure (DBP) 33%, respectively. More interestingly, 70%, 65%, 56%, 44%, 21%, 21% of these heritable genetic and familial influences on PAI-1 were also common to insulin, triglycerides, BMI, HDL cholesterol, SBP, and DBP, respectively. Also, 69%, 50%, and 53% of the genetic and familial influences on insulin were common to BMI, triglycerides, and HDL cholesterol, while 18% and 29% of the genetic and familial influences on BMI were shared by triglycerides, and HDL cholesterol, respectively. However, genetic influences on insulin and BMI appeared to be independent of those on blood pressure levels although the phenotypic correlations between blood pressure and both insulin and BMI were significant. In conclusion, genetic/familial influences on PAI-1, fasting insulin, BMI, triglycerides, HDL cholesterol, and blood pressure levels are moderate to large. There is clear evidence of genetic/familial clustering of PAI-1 and other components of MMS investigated in the present study. A set of genes common to PAI-1, insulin, BMI, and other components of MMS may be involved in the clustering.
P43 Alcohol consumption and plasminogen activator inhibitor type 1 in the NHLBI Family Heart Study
Luc Djousse, R. Curtis Ellison, Yuqing Zhang, James Pankow, Donna K. Arnett, Yuling Hong, Michael A. Province, Boston University School of Medicine, Boston, MA, University of Minnesota, Minneapolis, MA, University of North Carolina, Chapel Hill, NC, Washington University, St. Louis, MO
Elevated plasma concentration of plasminogen activator inhibitor type 1 (PAI-1) is associated with cardiovascular diseases. Although studies suggest a positive relation of alcohol to PAI-1 in general, the effects of light to moderate alcohol consumption on PAI-1 are not well established.
We used data collected on a random sample of 1862 Caucasian subjects in
the NHLBI Family Heart Study to assess the association between
different levels of habitual alcohol consumption and PAI-1 among women
and men. We created the following groups of alcohol consumption:
Non-drinkers, Ex-drinkers, 1.4, 1.5-4.9, 5.0-14.9, and 15
g/d for women, and non-drinkers, ex-drinkers, 1.4, 1.5-4.9,
5.0-14.9, 15-29.9, and 30 g/d for men. We fitted a regression
model, adjusting for anthropometric, metabolic, and
lifestyle factors. Subjects in the highest alcohol category were
leaner, had higher HDL, smoked more cigarettes, and consumed less
dietary fiber compared to non-drinkers. In the categories of
non-drinkers, ex-drinkers, 1.4 g/d through the highest category,
multivariate adjusted geometric mean PAI-1 were 10.80,
9.39, 10.17, 10.91, 11.13, and 16.28 ng/ml, respectively, for women and
18.54, 15.80, 15.33, 16.44, 16.94, 23.81, and 29.07 ng/ml,
respectively, for men. Comparing the lowest trough the highest alcohol
category to non-drinkers, the multivariate adjusted
regression coefficients (95% confidence interval) of ln(PAI-1) were
-0.11 (-0.28, 0.06), -0.03 (-0.21, 0.15), -0.02 (-0.24, 0.21), 0.01
(-0.18, 0.21), and 0.54 (0.29, 0.79), respectively, for women and -0.12
(-0.33, 0.07), -0.19 (-0.38, -0.01), -0.03 (-0.25, 0.19), -0.07 (-0.24,
0.11), 0.28 (0.03, 0.52), and 0.46 (0.23, 0.69), respectively, for
men.
These results suggest a J-shape relation, with only consumption above 15 g/d associated with increased PAI-1 in both genders.
P44 Plasminogen Activator Inhibitor-1 is Not Associated with the Amount of Fibrous Plaque in Arteries, but Contributes to Increased Risk of Heart Attack
Helena Kuivaniemi, Sungpil Yoon, Deepak Thatai, James Marsh, Gray T Malcom, Jack P Strong, C Alex McMahan, Gerard Tromp, Louisiana State University Medical Center, New Orleans, LA, University of Texas Health Science Center, San Antonio, TX, Wayne State University School of Medicine, Detroit, MI, Wayne State University School of Medicine, Det, MI, Wayne State University School of Medicine, Detroit, MI
Myocardial infarction is a major clinical manifestation of athero-sclerosis, a condition characterized by inflammatory-fibroprolifera-
tive response to insult to the vascular endothelium. We have investi-gated the role of plasminogen activator inhibitor-1 in early stages of atherosclerosis by using the PDAY collection (Pathobiolological Determinants of Atherosclerosis in Youth) and found no genetic association with the promoter polymorphism located in the gene for plasminogen activator inhibitor-1 and the amount of fibrous plaque present in the arteries of deceased young individuals. In contrast, when analyzing 280 patients referred to the Detroit Medical Center for coronary arteriograms, the 4G allele of plasminogen activator inhibitor-1 was found to contribute to an increase in risk, as reported previously in other studies, in patients with clinically manifested coronary heart disease. These findings suggest that plasminogen activator inhibitor-1 is not involved in the initial development of fibrous plaque, and does not contribute to the early stages of atherosclerosis, whereas its role in increasing the risk for heart attack is probably related to the formation of thrombus and a complex atherosclerotic lesion, which aggravates the disease and leads to serious clinical manifestations.
Supported by American Heart Association, Mid-America Research Consortium.
P45 Carotid Arterial Elasticity and retinal arteriolar narrowing
Duanping Liao, Rebecca Din-Dzietham, Jingping Mo, Ronald Klein, Daniel Jones, Lloyd E Chambless, Larry Hubbard, Richey A Sharrett, National Institute of Health/NHLBI, Bethesda, MD, Penn State University College of Medicine, Hershey, PA, University of Mississippi, Jackson, MS, University of North Carolina at Chapel Hill, Chapel Hill, NC, University of Wisconsin-Madison, Madison, WI, University of Wiscosin-Madison, Madison, WI
Decreased elasticity in large and medium sized arteries is an important marker of arterial sclerosis. Generalized retinal arteriolar narrowing represents the degree of microvascular damage due to atherosclerosis and elevated blood pressure. We examined the relationship between arterial elasticity and generalized retinal arteriolar narrowing in a bi-racial, population-based sample of 8031 men and women aged 45-64 at baseline from the Atherosclerosis Risk in Communities (ARIC) study. Arterial elasticity was measured as adjusted arterial diameter change in the left common carotid artery (AADC in micron, simultaneously adjusted for diastolic BP, pulse pressure, pulse pressure squared, diastolic arterial diameter, and height). A larger AADC reflects a greater elasticity. Generalized retinal arteriolar narrowing was measured as the ratio of the diameters of retinal arteries and veins (A/V ratio), which were assessed from digitized retinal photograph. The age, ethnicity-center, sex, smoking, hypertension, and diabetes adjusted means (SE) of A/V ratio from the lowest to the highest quartiles of AADC (from the lowest to highest arterial elasticity) were 0.833 (0.002), 0.836 (0.002), 0.837 (0.002), and 0.841 (0.002) respectively (p < 0.01). The similarly adjusted mean AADC (SE) in the lowest vs. highest quartiles of A/V were 394 (4) and 409 (4) micron respectively (p =0.007). The associations were similar between smokers and non-smokers, European and African Americans, and hypertensives and normotensives. Excluding individuals with type 2 diabetes did not alter the association. Data from this population-based study suggest a link between greater stiffness (reduced elasticity) in carotid artery and increased narrowing of arterioles, and indicate the burden of arterial sclerosis in different arterial beds.
P46 Intima-Media Thickness of the Carotid Bifurcation and Internal Carotid Artery and Risk of Myocardial Infarction. The Rotterdam Study.
Antonio Iglesias del Sol, Michiel L. Bots, Jacqueline C.M. Witteman, Albert Hofman, Diederick E. Grobbee, Erasmus University Medical School, Rotterdam, Netherlands, Julius Center of Patient Oriented Research, Utrecht, Netherlands
Background Common carotid intima-media thickness (IMT) has been found to be associated with the risk of myocardial infarction. In this study we examined whether also IMT of the carotid bifurcation and internal carotid artery are predictive of future myocardial infarction.
Methods and Results We used a nested case-control approach among 7983 subjects of 55 years and older participating the Rotterdam Study. At baseline (1990 through 1993), ultrasound images of the carotid bifurcation and the internal carotid artery were made. New cases of myocardial infarction were determined using hospital discharge records. For this analysis, we selected the first 118 myocardial infarctions and 4 random controls per case who remained free from myocardial infarction during follow-up. Results were adjusted for age and sex. Using logistic regression, the odds ratio (OR) for myocardial infarction associated with bifurcation IMT was 4.1 (95% confidence interval, 1.8-9.2) for the highest compared to the lowest quartile. For the highest compared to the lowest quartile of internal carotid artery, the OR for was 8.4 (2.1-33.4). The ORs for myocardial infarction per standard deviation increase of bifurcation IMT (0.55 mm) and internal carotid artery IMT (0.64 mm) were 1.4 (1.1-1.6) and 1.2 (0.9-1.6), respectively.
Conclusions The present study provides evidence that increased carotid bifurcation and internal carotid artery IMT are strong predictors of future coronary heart disease.
P47 Sex specific differences in risk factors for aortic atherosclerosis. The Rotterdam Study
A. Elisabeth Hak, Huibert A.P. Pols, Albert Hofman, Jacqueline C.M. Witteman, Erasmus University Medical School, Rotterdam, Netherlands
Background. Manifestations of atherosclerosis are usually considered to be more prevalent in men. However, the localization of the atherosclerotic process has been indicated to differ by gender and risk profile. We evaluated the prevalence of aortic atherosclerosis and the relation to classical cardiovascular risk factors in men and women.
Methods. The population consisted of 7983 men and women aged 55 years and over. Measurements included blood pressure, serum total cholesterol, HDL cholesterol, insulin, body mass index and waist-to-hip ratio. Smoking history was assessed during an interview. All participants were examined radiographically for calcified deposits in the abdominal aorta, which has been shown to represent intimal atherosclerosis. Calcifications were scored according to the length of the calcified area and classified as mild and advanced.
Results. The prevalence of advanced aortic atherosclerosis showed a sharp increase with age. The prevalence was higher in men than in women up to the age of 75 years. After this age the sex ratio reversed with the prevalence in men and women being 46% and 55%, respectively. Risk factors significantly associated with atherosclerosis in men were systolic blood pressure, total cholesterol and smoking. In women, HDL cholesterol, insulin and diabetes mellitus were additionally associated with aortic atherosclerosis. Associations with waist-to-hip ratio and body mass index in both sexes were largely mediated by other risk factors.
Conclusion. In the elderly, aortic atherosclerosis is more prevalent in women than in men. Risk factors known to cluster in the metabolic syndrome that has been associated with insulin resistance exert a stronger effect on female than on male risk of aortic atherosclerosis.
P48 Pulse pressure and risk of myocardial infarction in an elderly population
Nicole M. Popele, Diederick E. Grobbee, Michiel L. Bots, Albert Hofman, Jacqueline C.M. Witteman, Erasmus University Medical School, Rotterdam, Netherlands, Julius Center for Patient Oriented Research, Utrecht, Netherlands
Objective: Isolated systolic hypertension has been shown to be associated with an increased risk of cardiovascular disease. Pulse pressure is increasingly recognised as an important indicator of vascular ageing. We have studied the associations of pulse pressure and the presence of isolated systolic hypertension with incident myocardial infarction in an elderly population.
Design and Methods: The relationships were studied within the Rotterdam study, a prospective population based cohort study of subjects aged 55 years and over. Pulse pressure was defined as the difference between systolic and diastolic blood pressure. Isolated systolic hypertension was defined as a systolic blood pressure above 160 mmHg and diastolic blood pressure lower than 90 mmHg. Subjects using blood pressure lowering medication or with a history of myocardial infarction or heart failure at baseline were excluded. A total of 3552 subjects were included in the analysis. Mean duration of follow up was 4.2 year. In this period a fatal or non-fatal myocardial infarction occurred in 74 subjects.
Results: The relative risk of myocardial infarction in subjects with isolated systolic hypertension was 2.39 (95% confidence interval 1.44-3.95), compared to normotensive subjects. An increase of 10 mmHg in pulse pressure was associated with a relative risk of myocardial infarction of 1.24 (1.11-1.40). Both analyses were adjusted for age, sex, and other cardiovascular risk factors. After additional adjustment for mean arterial pressure, an increase of 10 mmHg in pulse pressure was associated with an relative risk of 1.14 (1.00-1.31).
Conclusion: The results of this prospective study show that, in addition to isolated systolic hypertension, also pulse pressure is a pivotal predictor of myocardial infarction in an elderly population.
P49 Assessment of Coronary Artery Bypass Graft Patency with Electron Beam Computed Tomography for Secondary Prevention
Heiko Pump, Svenja Schimpf, Cornelia Sehnert, Stephan Moehlenkamp, Raimund Erbel, Rainer Seibel, Department of Cardiology, University of Essen, Essen, Germany, Department of Cardiology,University of Essen, Essen, Germany, MRI, University of Witten, Muelheim, Germany, MRI,University of Witten, Muelheim, Germany
Purpose: The aim of this study was to assess bypass graft patency noninvasively with electron beam computed tomography (EBCT) and to evaluate the efficiency of the EBCT examination, compared with coronary angiography (CA).
Materials and Methods: We examined 56 patients (52 m, 4 f, mean age: 64.9 years) with 153 coronary artery bypass grafts. CA was performed prior to EBCT examination. All patients were scanned in an C150-EBCT. The patency was examined with one or two bolus- injections of contrast medium via cubital vein. Results were compared to conventional angiography in a blinded fashion.
Results: Eighty-two percent (126/153 ) of all grafts could be assessed with EBCT and 116/126 (92%) grafts were correctly estimated by EBCT. Overall sensitivity: 71%, overall specificity: 99%, negative predictive value: 91% and positive predictive value: 96%.
Conclusion: EBCT with intravenous injection of contrast agent permits the non-invasive examination of bypass graft patency and could play an important role in the management of this patient group.
P50 Electron-beam computed tomography in the Pittsburgh cohort of the Cardiovascular Health Study
Anne B Newman, Kim Sutton-Tyrrell, Barbara L Naydeck, Daniel Edmundowicz, Lewis H Kuller, University of Pittsburgh, Pittsburgh, PA
The Cardiovascular Health Study has shown noninvasive measures of subclinical atherosclerosis to be powerful predictors of cardiovascular events in older adults. These measures did not include a direct measure of coronary atherosclerosis. We tested the hypothesis that older adults with no peripheral atherosclerosis would have lower coronary artery calcification (CAC) by electron beam CT (EBCT) than those with clinical or subclinical cardiovascular disease (CVD). We also evaluated whether cardiovascular risk factors were associated with high CAC among those with no evidence of subclinical CVD. EBCT was done at study years 10-11 for 132 participants of mean age 78.2 (range 70-93); 54% women, 11% African-American. CVD was present in 28 (21.2%). CAC scores ranged from 16-5459 (median=383) in those with CVD and from 0-1715 in those without CVD (median=183; p=.01). Of those without CVD (n=104), 17 had a CAC of 0. For these 104, baseline risk factors were assessed using the third quartile to define a high score (CAC>Q3=565). Preliminary analysis of risk factors showed significant bivariate associations between CAC>Q3 with male gender, lower HDL and normotensive status. Of note, changes in white matter grade (WMG) by cerebral MRI positively correlated with the CAC, r=.34, p<.002. Age-sex adjusted WMG (coded 0-9) remained independently associated with CAC>Q3 (OR=2.27, 95%CI=1.15-4.48). Healthy older adults, even those with no CVD by other peripheral measures, have a wide range of CAC. In the healthy subgroup (no CVD), those with higher scores had higher WMG by cerebral MRI.
P51 Racial Differences In Risk Of Pulmonary Embolism Or Deep Venous Thrombosis
Arthur L. Klatsky, Mary Anne Armstrong, Jacqueline Poggi, Kaiser Permanente Med Ctr, San Francosco, CA, Kaiser Permanente Med Ctr, Oakland, CA
Sparse U.S. population study data suggest a slightly higher prevalence of pulmonary embolism (PE) and deep venous thrombosis (DVT)in Blacks than in Whites, but geographical comparisons of "non-white/white" data show that this is not evident in the Pacific region (CA, OR, WA). We prospectively studied PE/ DVT hospitalizations in 128,934 persons (72,019 White, 34,661 Black, 5777 Hispanic, 13,593 Asian) in relation to traits determined at prior health examinations. There were 303 persons hospitalized for either PE or DVT; 206 persons were hospitalized first for PE. Cox proportional hazards models with 9 covariates were used. In multivariate models the following RR's (95% confidence intervals) were found for PE: Black/White = 1.5(1.1-2.1, p = 0.01), Hispanic/White = 0.8(0.3-2.1). There were no PE cases in Asians (p vs. expected rate < 0.0001). For DVT the RR's (CI's) were: Black/ White = 0.5(0.3-0.8, p= 0.01), Hispanic/ White = 0.5(0.2-1.7), Asian/ White = 0.3(0.3-1.0, p= 0.05). For PE/DVT combined, the RR's (CI's) were: Black/White = 1.1(0.4-1.4), Hispanic/White = 0.7 (0.3-1.5), Asian/White = 0.2(0.1-0.5, p = 0.002). These relationships were similar in men and women. Covariates significantly related to risk were age, male sex, body mass index, and a composite coronary risk/symptom variable; covariates not related were education, marital status, smoking, and alcohol. Exploration of symptoms, diagnostic tests, and possible delay in diagnosis is underway in the hope of finding partial explanations for the Black/White disparity (higher PE risk with lower DVT risk of Blacks); these findings will be presented. The very low PE/DVT risk of Asians shown in these data may explain U.S. geographic variations in "white/nonwhite" PE/DVT risk. Asians may have genetic (or other) protective traits against risk of PE/DVT.
P52 A Prospective Study of Predictors of Subarachnoid & Intracerebral Hemorrhage in a Large Population
Arthur L. Klatsky, Mary Anne Armstrong, Stephen Sidney, Gary D. Friedman, Kaiser Permanente Med Ctr, Oakland, CA
Reported data show that high blood pressure (BP), Black race, and smoking are predictors of both subarachnoid hemorrhage (SAH) and intracerebral hemorrhage (ICH), but the role of other risk traits is less clear. We prospectively studied SAH and ICH in 128,934 persons (mean baseline age = 41 years, 56% women, 56% White, 27% Black, 11% Asian) in relation to traits determined at prior health examinations. There were 133 SAH subjects (mean age at diagnosis = 55 years, 66% women, 49% Black, 11% Asian) and 299 ICH subjects (mean age at diagnosis = 69 years, 52% women, 36% Black, 13% Asian). Cox proportional hazards models with 12 covariates were used. Relative risks estimated from separate Cox models for SAH (p value) and ICH (p value) follow: age per 10 yrs (SAH) = 1.3 (<0.001) and (ICH) = 2.4 (<0.001), Black vs. White (SAH) = 2.0 (<0.001) and (ICH) = 1.6 (<0.01), Asian vs. White (SAH) = 1.8 (<0.05) and (ICH) = 1.7 (<0.01), smoke 1+ ppd (SAH) = 3.4 (<0.001) and (ICH) = 1.6 (<0.01), systolic BP per 10 mmHg (both SAH and ICH) = 1.2 (<0.001), diastolic BP per 6 mmHg (SAH) = 1.2 (<0.01) and (ICH) = 1.3 (<0.001), cholesterol per 10mg/dl (SAH) = 0.98 (ns) and (ICH) = 0.97 (<0.05). Education, marital status, body mass index (BMI), sex and alcohol were not significantly associated with risk of either SAH or ICH. These relationships were similar in men and women. These data confirm positive relationships of age, higher BP (systolic and diastolic), Black race and smoking to both SAH and ICH risk, with smoking more related to SAH and age more to ICH. Noteworthy are increased risk of Asian Americans for both SAH and ICH, a slightly inverse risk of ICH in relation to cholesterol level, & absence of relationships to glucose, alcohol and BMI.
P53 CVD Risk Factors and Lower Extremity Arterial Disease: The Chicago Heart Association Detection Project in Industry (CHA).
Miriam B Rodin, Martha L Daviglus, Kiang Liu, Daniel B Garside, Philip Greenland, Jeremiah Stamler, Northwestern University Medical School, Chicago, IL
Data from the Chicago Heart Association Study were analyzed for
associations of CVD risk factors measured in middle-age and lower
extremity arterial disease (LEAD) occurring in older age.
Participants were 18,074 (9,828 men and 8,436 women) ages 40-64 at
baseline examination (1967-73) who survived to Medicare eligibility
(65 years) during 1984-94. Follow-up averaged 24 years. LEAD cases
were identified from death certificates (ICD-8 443.9) and from Medicare
records Parts A and B (ICD-9 443.9). LEAD occurred in 258 men and
183 women (17 and 12 cases, respectively, from death certificates).
Based on multivariate Cox regression analyses,
age, systolic blood pressure (SBP), serum
cholesterol (Chol), cigarette smoking (Smok), former Smk,
and diabetes were significantly associated with LEAD in both men and
women (Table). For non-diabetic women, baseline weight and post-load
plasma glucose, and DBP in men, were also significantly associated with
LEAD. In conclusion, these data confirm the findings of shorter
duration studies that major CVD risk factors are predictive of LEAD as
well as CHD in later life.
|
P54 Analysis of Candidate Genes for Intracranial Aneurysms
Gerard Tromp, Sompong Vongpunsawad, Helena Kuivaniemi, Antti Ronkainen, Juha Hernesniemi, University of Helsinki, Helsinki, Finland, University of Kuopio, Kuopio, Finland, Wayne State University School of Medicine, Detroit, MI
The prevalence of intracranial aneurysms is about 1-6%. In the U.S.A. alone, 28,000 individuals die annually due to ruptured intracranial aneurysms and about the same number of new cases with permanent neurological disability are identified as a consequence of ruptured intracranial aneurysms. The goal of our studies is to identify susceptibility loci located on the human chromosomes that contribute to the development of intracranial aneurysms. Towards this goal, we identified 85 families with two or more members with intracranial aneurysms from the catchment area of Kuopio University Hospital in Finland. Blood samples were collected from the family members and DNA isolated for affected sib-pair linkage analysis. The first screen of the human genome carried out with markers spaced with 10 cM density and 48 sib-pairs, identified 7 areas exceeding the threshold lod score of 0.8. These regions of human chromosomes included 19q13.2 with a lod score of 2.63. Search of genomic databases revealed a host of possible, biologically relevant candidate genes in the location. We analyzed the sequences for the gene encoding urokinase-type plasminogen activator receptor to determine if mutations were present in patients with intracranial aneurysms. Detailed DNA sequencing analysis revealed two sequence changes: 1) TCA (leucine) to CCA (proline) in exon 7; and 2) T to C change in intervening sequence 3. In further analyses, neither sequence change co-segregated with the disease in the families. Future analyses will include other candidate genes from the same region.
P55 Associations between Angiotensinogen Gene Polymorphism and Left Ventricular Mass and Geometric Pattern
Weihong Tang, Donna Arnett, Richard Devereux, D.C. Rao, Albert Oberman, Paul Hopkins, Dalane Kitzman, Cornell University Medical College, New York, NY, The University of Alabama at Birmingham, Birmingham, AL, University of Minnesota, Minneapolis, MN, University of Utah, Salt Lake City, UT, Wake Forest University, Winston-Salem, NC, Washington University, St. Louis, MO
While the angiotensinogen (AGT) M235T polymorphism is associated with hypertension and plasma angiotensinogen, its association with left ventricular (LV) mass and LV geometric patterning is not well defined. In this study, hypertensive sibships and randomly selected subjects of the HyperGEN Echocardiography Study [503 whites (mean age = 61.7 years) and 532 African-Americans (mean age = 51.5 years)] were evaluated. Left ventricular structure [LV mass and relative wall thickness (RWT)] was obtained using 2D guided M-mode echocardiography. The AGT M235T variant was analyzed using standard polymerase chain reaction (PCR) techniques. The frequency of T235 allele was significantly higher in African-Americans (85.2%) than in whites (43.5%) (p<0.05). Both groups were in Hardy-Weinberg equilibrium. In ANCOVA models adjusted for weight, height, systolic blood pressure, age, sex, and hypertensive medication status, LV mass was significantly higher in those with MM genotype (170.20 g) compared with those with TT or TM genotypes (163.69g ) in whites (p<0.05), but not in African-Americans. The adjusted mean LV mass for TT/TM and MM genotypes in African-Americans were 172.95 g and 175.21 g, respectively. In African-Americans, the adjusted mean value of RWT was significantly higher in the TT/TM genotypes (0.36) than in MM homozygotes (0.33) (p=0.05). In whites, the adjusted mean RWT for TT/TM and MM genotypes was 0.35 and 0.34, respectively. The polymorphism was not significantly associated with either systolic functional parameters or diastolic functional parameters. However, there was a significant interaction between systolic blood pressure and the genotype with regard to the value of atrial filling peak velocity measured at mitral valve annulus. In conclusion, we observed significant associations between genetic variation of angiotensinogen gene and LV mass and RWT. This result suggests that this polymorphism may play a minor functional role in left ventricular hypertrophy and LV remodeling.
P56 Risk Factors for Congestive Heart Failure in U.S. General Population: NHANES I Epidemiologic Follow-Up Study (NHEFS)
Jiang He, Lorraine G Ogden, Suma Vupputuri, Catherine Loria, Paul K Whelton, National Center for Health Statistics, Hyattsville, MD, Tulane School of Public Health and Tropical Medicine, New Orleans, LA
Despite significant progress in the prevention and treatment of cardiovascular disease in the past two decades, the incidence of congestive heart failure (CHF) has been increasing steadily in the U.S. population. We studied risk factors for the incidence of CHF in 14188 NHEFS participants who were aged 25 to 74 years and did not have a history of CHF at their baseline examination. Age, sex, race, serum cholesterol, blood pressure, body mass index, history of diabetes, education, physical activity, cigarette smoking, and alcohol consumption were obtained at the baseline examination using standard methods. Incidence data for CHF and coronary heart disease (CHD) were obtained from medical records and death certificates. Cox proportional hazards models were used to examine the relationship between risk factors and incidence of CHF adjusting for age, gender, race, and time-dependent history of CHD. The 1,382 CHF cases were documented over an average of 19 years of follow-up. Incidence of CHF was positively associated with history of diabetes (RR, 1.94; 95% CI, 1.64 to 2.30; p<0.001), hypertension (RR, 1.47; 95% CI, 1.32 to 1.64; p<0.001), cigarette smoking (RR, 1.51; 95% CI, 1.33 to 1.70; p<0.001), and overweight (RR, 1.37; 95% CI, 1.23 to 1.53; p<0.001), while inversely associated with moderate or high physical activity (RR, 0.76; 95% CI, 0.68 to 0.84; p<0.001) and 12 years or more schooling (RR, 0.74; 95% CI, 0.66 to 0.83; p<0.001). Serum cholesterol and alcohol consumption were not significantly associated with incidence of CHF. Our study indicates that diabetes, hypertension, cigarette smoking, overweight, physical inactivity, and low education are independent risk factors for CHF. Modification of these risk factors should play an important role in the prevention of CHF in the general population.
P57 Epidemiology of Congestive Heart Failure in Three Ethnic Groups
Jasenka Demirovic, Ronald J Prineas, Mark F Rudolph, Univ of Miami School of Med, Miami, FL, Univ of Minnesotta, Minneapolis, MN, University of Miami School of Medicine, Miami, FL
Congestive heart failure (CHF) is a major medical and public health
problem but little is known about the
epidemiology of this condition among free
living minority populations. To determine the prevalence of CHF and
associated risk factors in a multi-ethnic community, we examined a
population sample of 2,759 elderly (65) African American (AA),
Hispanic-Cuban (HC) and white non-Hispanic (WNH) men and women of Miami
Dade County, Florida. There were 153 (5.6%) cases of self-reported CHF
in the sample. The unadjusted prevalence rate of CHF was virtually
identical among elderly AA and WNH (4.8%) and was significantly
(p=0.04) lower than the rate among HC (6.8%). Sex and ethnic-specific
analyses showed that the highest prevalence rate of CHF was
among HC women (8.2%). A multiple logistic analysis was used
to calculate odds ratio (OR) and 95% confidence intervals (CI) of CHF
in relation to age, sex, ethnicity, medical history and smoking and
alcohol drinking habits. Age (65-74 vs 85+ years, OR 1.7, CI 1.2-2.4,
p=0.005), HC ethnic group (OR 1.7, CI 1.2-2.6, p=0.008), history of
hypertension (OR 1.5, CI 1.0-2.1, p=0.02), history of myocardial
infarction (OR 2.3, CI 1.5-3.5, p=0.0001) and history of diabetes
mellitus (OR 1.9, CI 1.3-2.8, p=001) were directly, siginificantly and
independently associated with the prevalence of self-reported CHF.
These results confirm the findings of prior studies in elderly white US
residents which indicate that heart attack, hypertension and diabetes
mellitus are major risk factors for CHF. The results also show that
elderly AA and WNH have similar CHF prevalence rates. The higher CHF
prevalence in elderly HC women requires further investigation.
P58 Electrocardiographic Indicators of Autonomic Dysfunction and Risk of Primary Cardiac Arrest Among Patients Without Clinically Recognized Heart Disease
Eric A. Whitsel, Rachel M. Pearce, Trivellore E. Ragunathan, Danyu Lin, Pentti M. Rautaharju, Sheila A. Weinmann, Gail D. Anderson, Patrick G. Arbogast, David S. Siscovick, Cardiovascular Health Research Unit, Seattle, WA, Department of Medicine, Seattle, WA, EPICARE Center, Winston-Salem, NC, University of Michigan, Ann Arbor, MI, University of Washington, Seattle, WA, Veteran Affairs Puget Sound Health Care System, Seattle, WA
Whether ECG indicators of autonomic dysfunction (tachycardia, low RR interval variation and QT prolongation) are related to risk of primary cardiac arrest (PCA) among patients without clinically recognized heart disease remains unknown. We conducted a population-based, case-control study of out-of-hospital PCA among enrollees in Group Health Cooperative of Puget Sound (GHC). Cases (n=498) were GHC enrollees aged 18 to 79 years with incident PCA between 1980 and 1994 and an available ECG but no history of heart disease before their index date. Controls (n=529) were a demographically similar stratified random sample of GHC enrollees. Resting 12-lead ECGs obtained from ambulatory care medical records were classified by the Novacode. ECG variables analyzed included heart rate (HR, median [b/m]), RR interval variation (RRV, range [ms]) and Rautaharju's QT index (QTI [%]). Clinical characteristics of enrollees were determined from abstracted medical records and use of medication from computerized outpatient pharmacy files. Mean (interquintile range, IQR) HR, RRV and QTI were 72 (23), 77 (90) and 104 (12), respectively. Risk of PCA (odds ratio, [95% CI]) was estimated separately for an IQR increase in HR, an IQR decrease in RRV and an IQR increase in QTI using conditional logistic regression. After adjustment for clinical and pharmacologic risk factors, odds ratios were 1.36 [1.07-1.74], 1.12 [0.89-1.41] and 1.55 [1.25-1.93], respectively. After further adjustment for other ECG features including ventricular conduction, ectopy, hypertrophy, ischemia and infarction, odds ratios were 1.25 [0.96-1.62], 1.06 [0.83-1.35] and 1.36 [1.07-1.72]. These findings suggest that of several 12-lead ECG indicators of autonomic dysfunction, QT prolongation is related to risk of PCA.
P59 Use of inhaled beta-agonists and the risk of primary cardiac arrest
Rozenn N Lemaitre, David S Siscovick, Bruce M Psaty, Gail D Anderson, Rachel M Pearce, Trivellore E Raghunathan, Danyu Lin, Sheila A Weinmann, Eric A Whitsel, University of Michigan, Ann Arbor, MI, University of Washington, Seattle, WA, Veteran Affairs Puget Sound Health Care System, Seattle, WA
Drugs that increase sympathetic function might influence the risk of primary cardiac arrest (PCA). We used data from a population-based case-control study conducted at Group Health Cooperative of Puget Sound (GHC) to assess the association of inhaled beta-agonists with the risk of PCA. Cases were all GHC enrollees, aged 40 to 79, with incident PCA between 1/80 and 12/94 (n=2125). Controls were a stratified random sample of GHC enrollees without PCA, frequency-matched to cases on age, gender, calendar year, and use of digoxin or nitroglycerin (n=4027). Information about clinical characteristics and other PCA risk factors was obtained by medical record review. Medication use was determined from GHC computerized pharmacy data. Mean age of the study subjects was 67 years. The proportion of cases and controls who received inhaled beta-agonists within 3 months prior to their index date were 7.2% and 3.4% respectively. We classified users according to the number of beta-agonist cans received within 3 months prior to index date. Never users (defined as no use in past 2 years) were the reference group. The risk of PCA was estimated from conditional logistic regression and adjusted for matching factors and other PCA risk factors including pulmonary disease. Compared to never use, the odds ratio of PCA associated with one can inhaled beta-agonist in the 3 month interval was 1.31 (95% CI: 0.83-2.07) and the odds ratio associated with two or more cans was 2.27 (95% CI: 1.58-3.26, p for test for trend < 0.001); the odds ratio associated with one or more cans prior to the 3 month interval was 0.81 (95% CI: 0.59-1.11). Adjustment for theophylline use changed the odds ratios only slightly. These results suggest that use of inhaled beta-agonists is associated with an increased risk of PCA in a dose-response fashion.
P60 Acute Coronary Syndrome in Young Adults: Gender Differences in Utilization of Interventional Therapy.
Kwame O Akosah, Beth Gower, Brenda Rooney, Linda Groom, Kris Paul, Rajah Sundaram, Gundersen Lutheran Medical Center, La Crosse, WI
Young women are generally considered to have a low risk of developing early coronary artery disease. However, the extent to which this is true in a high risk population remains unclear. Moreover, it is unclear how this misconception contributes to utilization of resources in young women admitted with acute coronary syndrome. The objective of this study was to determine the influence of gender on the utilization of invasive cardiac interventions in young adults. There were 2042 admissions at this community hospital for acute syndrome over a two year period; 449 (22%)were aged 50 or younger (mean 43±3). Among the 449 patients, 132 (30%) were women. There were no gender differences in the rates of previous MI (9.9% vs 12.9%), previous bypass surgery [CABG] (4.6% vs 3.5%), previous PTCA (7.6% vs 5.1%) or previous angiography (25.8% vs 22.4%) in females and males respectively. However, more young males underwent angiography during hospitalization (72.4% vs 62.8%, p<0.05) and PTCA (26.2% vs 14.4%, p<0.007) and a trend towards more CABG (22.4% vs 15.9%, p=NS). The prevalence of hypertension (34.1% vs 36.3%) and prevalence of family history of CAD and hyperlipidemia were higher in males versus females (29.4% vs 17.4%, p<0.009) and (39.4% vs 27.3%, p<0.014) respectively, but obesity was more common among females (28.8% vs 19.6%, p<0.032). Both current smoking and history of smoking were equally high in both gender (51% vs 60%) and (72% vs 81%, p=NS) respectively. This data shows that there is gender bias in the utilization of cardiac interventions among young people with premature coronary artery disease.
P61 Trends in classification of strokes and in validated discharge rates by sex, southeastern New England, 1980-1991.
Carol A. Derby, Kate L. Lapane, Henry A. Feldman, Richard A. Carleton, Brown University, Providence, RI, Memorial Hospital of Rhode Island, Pawtucket, RI, New England Research Institutes, Watertown, MA
Hospital discharge rates for stroke are influenced by many
factors including incidence, trends in criteria for hospitalization,
and trends in diagnostic and classification criteria. Data
on validated strokes in a well defined population are rare. This study
examined sex-specific trends in confirmed stroke, the distribution of
stroke types and secular trends in risk factors in the Pawtucket Heart
Health Program study communities for 1980-1991. Analyses were
based on 3,975 discharges with ICD-9 diagnosis code 431-432, 434, 435,
or 436-437. Medical records were reviewed according to a validation
algorithm for 96%, and 57% were confirmed. Temporal trends were
evaluated with repeated measures linear regression. For non-validated
discharges there was no consistent trend in non-fatal stroke,
while fatal stroke discharges declined in men and women and across each
decade of age(p = 0.0001). In contrast, decreasing trends were
observed for both non-fatal and fatal confirmed strokes (p
0.0004). The declines appear to have been greater for men than women (p
for sex by time interaction 0.04). The proportion of strokes
with a CT scan increased from 39% to 86%, the percent classified as
thromboembolic increased from 21% to 56%, and the percent of
ill-defined strokes declined from 48% to 10%. The percent of stroke
diagnoses validated increased less, from 52% to 63%. The prevalence
of hypertension was constant across the period (35% men, 23% women),
and rates of control plateaued after 1985. Temporal trends in detection
and classification of strokes may explain the lack of a trend in
non-fatal icd-9 stroke discharge rates. Secular trends in hypertension
control do not appear to explain the declines in confirmed strokes.
Steeper declines in men suggest gender differences in benefits from
prevention and treatment.
P62 Recent changes in mortality from cardiovascular disease in China: 1987-1996
Xin-Hua Zhang, Ting-Rui Guan, Li-Sheng Liu, Jia-Wen Mao, Anthony Rodgers, Stephen MacMahon, Center for Health Statistics Information, Beijing, People's Rep of China, Chinese Academy of Medical Sciences, Beijing, People's Rep of China, The University of Auckland, Auckland, New Zealand
Data on cardiovascular mortality (CVM) are available from about 100 million men and women (0-85y+) in selected rural and urban areas of China. Data for the period 1987-1996 (age-adjusted to the world population) were included in this analysis. Throughout this period, cardiovascular disease was the leading cause of death in China, although CVM rates declined from between 203 and 273 per 100,000 in 1987 to between 158 and 260 per 100,000 in 1996. The decline was greater in rural areas than in urban areas for both men (15% vs 5%) and women (22% vs 14%). Over this period, the male to female sex ratio of CVM increased from 1.25 to 1.38, but still remains much lower than that in most Western populations. Overall CVM rates were similar to those in many Western populations, but the composition of CVM was markedly different. In all years and subgroups, stroke accounted for more than half of all CVM. From 1987 to 1996, stroke deaths declined by 14% in urban and rural women and by 7% in rural men and 4% in urban men. In contrast, deaths from ischaemic heart disease increased by 20%-24% in rural and urban men and by 13%-16% in rural and urban women; by 1996, ischaemic heart disease accounted for 15% of CVM in rural areas and 26% in urban areas. In 1987, pulmonary heart disease accounted for 20% of CVM in rural areas and 10% in urban areas, but deaths from this cause declined by about 60% in all subgroups during the following decade. Over this period, deaths from rheumatic heart disease also decreased by about 40%, but deaths from other heart disease increased by between 15% and 27%. In conclusion, these results indicate that stroke remains the leading cause of CVM in China, although death from ischaemic heart disease is steadily increasing in all major population subgroups.
P63 Differences in Case Fatality of Myocardial Infarction by Socioeconomic Status in Finland; The Finmonica MI Register Study
Veikko V. Salomaa, Heikki Miettinen, Matti Niemela, Jyvaskyla Central Hospital, Jyvaskyla, Finland, Loimaa Regional Hospital, Loimaa, Finland, National Public Health Inst., Helsinki, Finland
Finland is a Scandinavian country with a government subsidized health
care and a comprehensive social security system. We have examined the
association of socioeconomic status (SES) with the 28-day case fatality
(CF) of myocardial infarction (MI) using the FINMONICA MI Register data
for years 1983-1992. Patients with their first ever MI (6485 male and
1942 female) aged 35-64 years were included in the study. Information
on indicators of SES was obtained by record linkage of the MI
register data with the files of Statistics Finland. The
age-standardized 28-day CF was significantly higher among men with low
taxable income than among those with high income (Table). The
differences between the income classes did not diminish over time. The
28-day CF was further divided to out-of-hospital and in-hospital CF,
but similar differences were observed in both. Findings among women
were consistent with those among men. Thrombolytic
treatments and revascularizations were
significantly more frequent among patients with high income than among
those with low income. Time interval from the onset of symptoms to
medical presence was shorter in men with high income than in men with
low income. In conclusion, there are considerable socioeconomic
differences in the CF and treatment of MI in Finland. These differences
showed no signs of reduction during the period 1983-1992.
|
P64 Trends in attack rates and aspects of medical care for unstable angina
Wayne D Rosamond, Beth D Weatherley, Aaron R Folsom, Jianwen Cai, Verna Lamar, Lawton S Cooper, NHLBI, Bethesda, MD, University of Minnesota, Minneapolis, MN, University of North Carolina, Chapel Hill, NC
Data from the National Hospital Discharge Survey indicate that hospitalizations for unstable angina increased to 570,000 in 1991 and then declined to 285,000 in 1995. Although clinical practice guidelines for the diagnosis and management of unstable angina are available, little is known about their application in community hospitals. Since 1987, the Atherosclerosis Risk in Communities Study has conducted community surveillance of coronary heart disease, including hospital discharges of unstable angina (ICD9-CM 411), among persons 35-74 years of age in Forsyth County, NC; Washington County, MD; suburban Minneapolis, MN; and Jackson, MS. Between 1987 and 1994, 6,376 hospitalizations with a primary diagnosis of unstable angina occurred. The mean age was 61 years, 37% were female, and 15% were African American. The attack rate of unstable angina among men increased 54% from 2.70 per 1,000 in 1987 to 4.16 per 1000 in 1992, then declined to 2.83 per 1000 in 1994. The rate in women changed little from a rate of 1.77 per 1000 in 1987. The use of aspirin in the treatment of unstable angina doubled from 40% of cases in 1987 to 80% in 1994. For all years combined, cardiac catheterization was used in 51% of cases. Its highest level of use (62%) was in 1992. After adjusting for age, race, center, and year of discharge the female to male odds ratios (95% confidence interval) of receiving aspirin and cardiac catheterization were 0.53 (0.47, 0.60) and 0.68 (0.61, 0.77), respectively. The adjusted black to white odds ratios were 0.55 (0.43, 0.67) and 0.56 (0.47, 0.67), respectively. These data suggest that trends in hospitalizations for unstable angina in these communities were similar to those seen nationally. The use of aspirin and cardiac catheterization has evolved, although significant differences in their use among race and gender groups remain.
P65 Time of Onset, Awakening with Symptoms, and Delay in Seeking Care for Acute Stroke
Wayne D Rosamond, Dexter L Morris, Kelly R Evenson, Jeffrey A Vallee, Kristie W Schmidt, Suzanne Cook, GlaxoWellcome, Inc., RTP, NC, University of North Carolina, Chapel Hill, NC
A substantial proportion of stroke patients awake with symptoms of stroke. Currently most treatment guidelines for acute stroke measure symptom onset as the last time the patient was known to be symptom free. This influences the estimate of pre-hospital delay (time from stroke symptom onset to Emergency Department (ED) arrival) for patients who awake with symptoms and therefore may limit the amount that delay can be reduced among these patients. We conducted a prospective registry of patients with stroke-like symptoms presenting to ED's in three states. Among 389 patients enrolled to date, the median time from symptom onset to ED arrival was 4.0 hours with 33% reporting that they awoke with symptoms. Patients who awoke with symptoms had significantly longer median pre-hospital delay times than those who did not wake with symptoms (5.8 hours vs. 2.6 hours, respectively). Onset of symptoms while sleeping was more likely among older (>80 years of age) patients. Median pre-hospital delay times varied markedly by time of symptom onset as shown below. Much of this variation is likely related to the proportion of patients who are asleep when their event occurs.
P66 Geographic Proximity of Patient's Residence to Invasive Cardiac Services: Impact on Utilization
Patrice M. Gregory, Edmond S. Malka, Alan C. Wilson, Jasmine K. Arora, John B. Kostis, George G. Rhoads, UMDNJ-Robert Wood Johnson Medical School, New Brunswick, NJ
Objective: To study the effect of geographic distance from patient's residence to invasive cardiac services on utilization of these services.
Study Design and Subjects: A historical cohort study of 35,191 New Jersey residents aged 65 and older hospitalized with primary diagnosis of acute myocardial infarction(AMI) between 1992-1996.
Methods: Hospital discharge records were linked on a statewide basis for follow-up of re-admission within 90 days of initial hospitalization for AMI. Use of coronary angioplasty(PTCA) and coronary artery bypass graft surgery(CABG) were outcome measures. Distance from patient's residence to nearest hospital with invasive cardiac services(PTCA or CABG services) was calculated based on straight-line distance from centers of zip codes and categorized as: 0 to <2, 2 to <5, 5 to <10, 10 to <15, 15 to <20, 20 to <25, >=25 miles. Adjusted odds of PTCA or CABG use at each distance category were compared to odds at 25 miles or more.
Results: A sharp linear decline in adjusted odds ratios for PTCA use was found with increasing distance from patient's residence to nearest hospital providing invasive cardiac services. Adjusted odds of PTCA use were 3.1, 2.7, 2.2, 1.9, 1.7 and 1.1 times higher for each distance category, respectively, in comparison to that of 25 miles or more. Every distance category under 20 miles showed statistically significant increases in PTCA use compared to 25 miles or more. Use of CABG was also higher (2.0, 1.7, 1.7, 1.2, 1.5, 0.9 times higher, respectively) for patients residing closer to invasive cardiac services. Results were similar for male, female, White and Black subgroups.
Conclusion: Closer geographic proximity to invasive cardiac services was strongly associated with increased utilization. The patterns may indicate under or over-utilization of these cardiac services.
P67 The association of physical activity and markers of inflammation in a healthy elderly population.
Dominic F Geffken, Mary Cushman, Greg Burke, Joseph F Polak, Pam A Sakkinen, Russell P Tracy, Harvard Medical School, Boston, MA, University of Vermont College of Medicine, Burlington, VT, Wake Forest University School of Medicine, Winston-Salem, NC
Background: Higher levels of physical activity are associated with lower risk of cardiovascular disease (CVD). There is growing evidence that development of the atherosclerotic plaque is associated with inflammation. We investigated the cross-sectional association of physical activity and inflammation markers in a healthy elderly population.
Methods: Data from the Cardiovascular Health Study, a cohort of 5,888 men and women over age 65, were analyzed. Concentrations of the markers of inflammation, C-reactive protein (CRP), fibrinogen, Factor VIII:C (FVIII:C), white blood count (WBC), and albumin were compared cross-sectionally by quartiles of self-reported physical activity.
Results: Compared to the lowest quartile, those in the
highest quartile of physical activity had 21%, 8%, 4%, 3% lower
concentrations of CRP, WBC, FVIII:C, and fibrinogen respectively,
adjusting for gender, presence of CVD, age, race, smoking, body mass
index, diabetes, and hypertension. Participants 72 years of age
showed associations similar to participants, 65-71 years of age. No
signifiacnt interaction was observed between physical activity and
gender. There was no cross-sectional association of physical activity
and albumin.
Conclusions: Higher levels of physical activity were associated with lower concentrations of inflammation markers in this elderly cohort. These data suggest a hypothesis that reduced inflammation is a beneficial effect of exercise. Prospective studies and basic research are required for verification.
P68 The association between C-reactive protein, diabetes and other CVD risk factors among elderly Japanese-American men; The Honolulu Heart Program
Beatriz L Rodriguez, Jess David Curb, Pamela Sakkinen, Diane E Bild, Cecil M Burchfiel, Robert Abbott, Irwin Schatz, Katsuhiko Yano, Russell Tracy, NIH, NHLBI, Jackson, MS, NIH, NHLBI, Bethesda, MD, University of Hawaii, Honolulu, HI, University of Vermont, Colchester, VT, University of Virginia, Charlottesville, VA
Elderly Japanese-American men have very high rates of diabetes (36%). Diabetes tends to influence the relationships between C-reactive protein, a marker of inflammation, and other CVD risk factors. These relationships were examined in a random sample of 400 Japanese American men ages 71-93 years old who participated in the Fourth Examination of the Honolulu Heart Program (1991-93). Because of this interaction effect, analyses were stratified by diabetic status. Prevalent cases of CVD were excluded from the analyses. Using stepwise regression analyses among the non diabetic subjects, the variables HDL cholesterol, FEV1 and SBP were inversely, significantly and independently associated with CRP. Waist-hip ratio showed a direct association (p<0.05). Alcohol showed an inverse borderline association (p=.07). Age, BMI, physical activity index, triglyceride, glucose, insulin, ankle brachial index, and pack-years of smoking did not enter the model (p<0.10). Since fibrinogen and WBC count are highly correlated to CRP (p<0.0001) and are also indicators of inflammation they were not used in these models. Among the diabetic subjects, only FEV1 was inversely related to CRP. These analyses suggest that CRP levels are closely related to other risk factors such as HDL, SBP, FEV1 and waist-hip ratio in non-diabetics. However, diabetes appears to overide the association of CRP levels with most of these factors in this elderly population of Japanese-American men.
P69 Effect of Postmenopausal Hormones on Inflammation Proteins: The PEPI Study
Mary Cushman, Claudine Legault, Elizabeth Barrett-Connor, Marcia L Stefanick, Craig Kessler, Howard L Judd, Pamela A Sakkinen, Russell P Tracy, Georgetwon University, Washington, DC, Stanford University, Stanford, CA, University of California, Los Angeles, Los Angeles, CA, University of California, San Diego, La Jolla, CA, University of Vermont, Burlington, VT, University of Vermont, Colchester, VT, Wake Forest University, Winston-Salem, NC
Background: Postmenopausal hormone use in women with coronary disease was recently associated with increased proximate risk of coronary events. Observational studies in healthy women suggest the opposite effect. Since higher levels of inflammation markers predict vascular disease, effect of hormones on these markers is of interest. Methods: Four inflammation markers, C-reactive protein (CRP), soluble E-selectin, von Willebrand factor (vWF), and coagulation factor VIIIc were measured at baseline, 12 and 36 months in 365 women in the Postmenopausal Estrogen/Progestin Intervention Trial, a randomized placebo-controlled trial of the effects of four hormone preparations (estrogen alone or combined estrogen + progestin as either micronized progestin or continuous or cyclical medroxyprogesterone acetate) on cardiac risk factors. Results: Compared to placebo, all four hormone therapies resulted in increased CRP and decreased E-selectin (p = 0.0001), with no effects on vWF or factor VIIIc. For example, for estrogen alone, over 36 months CRP rose from 1.03 to 1.93 mg/L compared to 1.21 to 1.26 mg/L with placebo (p = 0.0001). E-selectin declined from 35.8 to 29.7 ng/ml with estrogen compared to 34.5 to 35.4 ng/ml with plabebo (p = 0.0001). Effects were similar among the 4 active treatments. At 36 months the percent changes from baseline due to any active treatment were: 85% increase in CRP, 18% decrease in E-selectin. Conclusions: Hormone therapy had varied effects on inflammation. Any cardioprotection from hormones may not be mediated by CRP or the physiological factors it reflects, but results suggest it may be related to lowered soluble E-selectin. Effects on CRP may relate to possible adverse early effects of therapy. To understand clinical sequelae of the findings, studies of these changes in relation to subsequent clinical events are needed.
P70 Longitudinal stability of coagulation, fibrinolysis, and inflammation factors in stored plasma samples
Michael R Lewis, Russell P Tracy, University of Vermont, Colchester, VT
Background: Molecular epidemiologic studies of risk factors for
atherosclerosis and thrombosis often rely on
analysis of plasma samples which may have been stored at
-70°C for several years. Limited data exist, however, concerning the
stability of most assayed factors in frozen samples. Methods and
Results: In our laboratory we have assayed replicate samples of
frozen plasma pools (controls) over time as a quality control
procedure. We reviewed data collected on 14 different control pools
over periods ranging from 7-59 months for the following factors:
C-reactive protein, D-dimer, factor VIIc, fibrinogen,
plasmin-2-antiplasmin complex, PAI-1 antigen, protein C
antigen, protein S antigen, and t-PA antigen (over 6,900 individual
data points). We calculated monthly means for each control pool, for
each factor. We used linear regression analysis to screen for
changes with respect to time. For example, we analyzed 5
different control pools for fibrinogen. There was no significant
association of fibrinogen values with time in any of the 5
analyses. We interpreted these results to mean no
consistent change for fibrinogen in samples frozen at -70°C
over the interval studied (up to 25 months). Significant associations
between monthly means and time were found only sporadically in the
other 17 individual regression analyses we performed. Nonlinear
regression (quadratic, cubic, logarithmic) yielded no additional
information. Taken together, we found no consistent evidence of
significant sample degradation over time for the factors studied.
Conclusion: Our finding of longitudinal stability in the
biochemical properties of frozen plasma bolsters the presumption of
assay validity on which molecular epidemiologic studies are based.
P71 C-reactive protein is related to carotid intima-media thickness in healthy middle-aged women
A. Elisabeth Hak, Michiel L. Bots, Coen D.A. Stehouwer, Kees H. Polderman, Albert Hofman, Jacqueline C.M. Witteman, Erasmus University Medical School, Rotterdam, Netherlands, Free University Hospital, Amsterdam, Netherlands, Julius Center of Patient Oriented Research, Utrecht, Netherlands
Background. Recent data suggest that inflammation is involved in the process of atherosclerosis. C-reactive protein, a major acute phase protein, has been shown to be related to the risk of cardiac events. In this study the association between levels of C-reactive protein and intima-media thickness of the common carotid arteries was examined.
Methods. The population comprised 186 healthy women, aged 43 to 55 years, selected from the general population. All participants were free from cardiovascular disease. Common carotid intima-media thickness was assessed by ultrasonography.
Results.Mean common carotid intima-media thickness was 0.61 mm (standard deviation 0.09 mm). C-reactive protein was significantly asssociated with common carotid intima-media thickness (ßcoefficient = 0.021, 95% confidence interval 0.003 to 0.039, adjusted for age). After stratification by smoking status associations between C-reactive protein and common carotid intima-media thickness appeared to be present in former and current smokers only (ßcoefficient = 0.040, 0.013 to 0.067 in ever smokers, ßcoefficient = 0.004, -0.020 to 0.028 in never smokers).
Conclusion. The results of this population based study in healthy middle-aged women show that C-reactive protein is associated with common carotid intima-media thickness in ever smokers. These data suggest that C-reactive protein may mark permanent underlying vascular damage due in part to smoking. This may explain why the association between inflammation and intima-media thickness is more pronounced not only in current, but also in past smokers.
P72 C-reactive protein, cardiovascular risk factors and mortality in a prospective study in the elderly
Timo E Strandberg, Reijo S Tilvis, Department of Medicine, University of Helsinki, Helsinki, Finland
A marker of inflammation, C-reactive protein (CRP) has been implicated to associate with cardiovascular disease (CVD) events. However, there are few studies in the elderly, and moreover, smoking may be an confounder in the association. In the population-based Helsinki Aging Study, CRP was measured with an enzyme-immunoassay method (sensitivity 0.3 mg/L) from the baseline serum samples of 476 elderly individuals (75-85 years, 73 % women, 8 % smokers) along with CVD risk factors. Mortality follow-up lasted 5 years. The range of CRP concentrations was 0.18-170.0 mg/L (median 1.60 mg/L). At baseline, CRP concentration was inversely correlated with serum albumin (r= -0.302, p<0.001), serum creatinine (r= -0.164, p<0.001), serum cholesterol (r= -0.143, p=0.003), and HDL cholesterol (r= - 0.273, p<0.001). The correlation with serum triglycerides was nonsignificant, whereas weak positive correlations were observed with BMI (r= 0.092, p=0.050) and plasma insulin (r= 0.135, p=0.007). During the 5-year follow-up 192 individuals died, 48 % due to CVD. When adjusted for age and sex, baseline CRP concentration (per 10 mg/L) significantly predicted total mortality (risk ratio 1.28, 95% confidence interval 1.19-1.60), and also tended to predict CVD mortality only (RR 1.37, 95 % CI 0.94-1.99).
The results show that even low concentrations of C-reactive protein are correlated with important cardiovascular risk factors and significantly predict 5-year mortality in this elderly population with low prevalence of smoking.
P73 C-Reactive Protein and Incident Myocardial Infarction in Japanese American Men.
Pamela Sakkinen, Robert D Abbott, J David Curb, Beatriz L Rodriguez, Katsuhiko Yano, Russell P Tracy, Kuakini Medical Center, Honolulu, HI, University of Hawaii, Honolulu, HI, University of Vermont, Colchester, VT, University of Virginia, Charlottesville, VA
C-reactive protein (CRP) has been shown to predict incident
cardiovascular disease (CVD) in several populations.
However, there is little information in racial groups other than
Caucasian. We used a nested case-control design to determine the
relation between CRP and risk of incident myocardial infarction (MI) in
48-70 year old Japanese American men from the Honolulu Heart Program.
The average follow-up was 21 years, yielding 369 cases compared to 1348
controls free of coronary heart disease or stroke during this
same time period. Traditional CVD risk factors, viz. total
cholesterol, serum glucose, hypertension, body mass index,
and cigarettes smoked/day, were more common in cases than controls (all
P0.001); values for the latter three were higher with increasing CRP
levels (all P0.001, comparing values in 3rd tertile of CRP to 1st).
The age-adjusted percentage of men with MI increased with increasing
CRP tertile in both younger men (48-55 years, P trend 0.001) and
older (56-70 years, P trend 0.05). However, after adjusting for age
and CVD risk factors, the Odds Ratio for MI comparing the 3rd CRP
tertile to the 1st was significant in the younger men (1.7, CI =
1.1-2.7) but not in the older (1.2, 0.8-1.9). When stratified by
smoking status, the adjusted 3rd vs 1st Odds Ratio was significant in
the current smokers (1.8, 1.1-3.1) but not in former or never smokers;
when stratified on diabetes status, the 3rd vs 1st Odds Ratio was only
significant in non-diabetics (1.6, 1.1-2.2). This latter finding was
due to a relatively large incidence of MI in those with diabetes in the
first CRP tertile. We conclude that, over an extended follow-up,
baseline CRP levels were related to incident MI in middle-aged Japanese
Americans but not in the elderly. This moderately strong association
was seen primarily in smokers, and in those without diabetes.
P74 C-reactive protein, Serum Amyuloid A Protein, Cardiovascular Disease and Death: The Framingham Heart Study Experience
Peter W.F. Wilson, Ralph B. D'Agostino, Carl Franzblau, Halit Silbershatz, Philip A. Wolf, Boston University, Boston, MA, Framingham Heart Study/NHLBI, Framingham, MA
Elevated C-reactive protein (CRP) and serum amyloid A protein (SAA) may indicate vascular inflammation and aid in the prediction of cardiovascular disease (CVD). Associations between CVD and levels of CRP and SAA were examined for 630 men and 901 women without CVD at the time of their 1980-1982 clinic examination. An ultrasensitive CRP assay was used (Hemagen Diagnostics), employing serum frozen at –20 degrees centigrade. Analyses were performed separately for men and women, using sex-specific quartiles of CRP (approximately Q1=0-1.1, Q2=1.1-3.2, Q3=3.3-7.7, Q4=7.7-50 ug/mL) and odds ratios (OR) were derived with the lowest quartile as the referent group. The mean age of the participants was 68 years at baseline and they were followed 12 years for incident CVD and death. Significant associations with adverse outcomes were generally associated only with the top quartile levels of CRP. After adjustment for age, smoking, total cholesterol, HDL cholesterol, systolic pressure and diabetes, the top quartile of CRP was associated with an increased odds for all-cause mortality (men, OR=1.79, P=0.001; women, OR=1.62, P=0.009) and cardiovascular mortality (men, OR=1.96, P=0.02; women OR=2.14, P=0.02). After multivariate adjustment the top quartile of CRP was consistently associated with a significantly increased odds for coronary heart disease (CHD), CHD death, non-CVD death, and cancer death in men, but not in women. Quartiles of SAA were generally not associated with an increased risk for adverse outcomes. Conclusion: The top quartile of CRP was highly associated with increased all-cause and CVD mortality in both sexes after adjustment for common cardiovascular risk factors, and this marker of inflammation was consistently associated with several adverse outcomes in elderly men.
P75 Risk factors for the unstable atherosclerotic plaque
Richard GJ Gibbs, Nessa Carey, , Adam Mitchell, Roger M Greenhalgh, Alun H Davies, Imperial College School of Medicine, London, United Kingdom, Imperial College School of Medicine, United Kingdom, Imperial College School of Medicine, London, United Kingdom
Aim: To determine possible morphological (degree of internal carotid artery stenosis and plaque echolucency) and biological factors (the presence of the microorganism C.pneumoniae) which could discriminate between embolising and non embolising plaques in a cohort of patients undergoing carotid endarterectomy.
Methods: 91 consecutive symptomatic patients were recruited. All patients underwent transcranial Doppler insonation of their ipsilateral middle cerebral artery for 30 minutes 24 hours before surgery and embolic signals were recorded onto a DAT recorder. The degree of ipsilateral carotid artery stenosis and Gray Weale classification of the plaque were assessed using Duplex. The presence of chlamydial DNA was detected using a nested polymerase chain reaction (PCR).52 patients had preoperative CT scans available for analysis.
Results 91 patients were entered.12 of the patients were embolising
preoperatively. 27% of the plaques were positive for
C.pneumoniae. Conclusion: Younger male patients and those
with echolucent plaques are significantly more likely to embolise
preoperatively. The presence of C. pneumoniae does not
effect observed embolisation. The rationale for antibiotic therapy for
C.pneumoniae in atherosclerotic plaques needs to be
defined.
|
P76 The relation between tooth loss and incidence of ischemic stroke
Kaumudi J Joshipura, Alberto Ascherio, Eric B Rimm, Chester W Douglass, Walter C Willett, Harvard School of Dental Medicine, Boston, MA, Harvard School of Public Health, Boston, MA
We evaluated the association between baseline tooth loss and the incidence of ischemic stroke in an eight-year follow-up of 42,151 health professional men who were free of cardiovascular disease (CVD) at baseline. Compared to men with 25 or more teeth at baseline, those with 10 or fewer teeth had a relative risk (RR) of 1.75 (95% Confidence Interval 1.03, 2.99), those with 11-15 teeth had a RR of 1.97 (95% CI 1.07, 3.64), and those with 17-24 teeth had a RR of 1.48 (95% CI 1.02, 2.13). These relative risks were adjusted for age, smoking, obesity, alcohol, exercise, aspirin, family history of CVD, profession, hypertension and hypercholesterolemia. When modeled as a continuous variable, each tooth present was associated with a decremental risk of 3%; RR of 0.97 (0.96, 0.99). The risk ratios were attenuated slightly after adding quintiles of total fruit and vegetable intake to the model; the RR was 1.69 for 0-10 teeth, 1.97 for 11-15 teeth and 1.46 for 16-24 teeth compared to 25 or more teeth. The decision to extract teeth is generally preceded by chronic infection in the form of periodontal disease or pulpal infection from dental caries; extraction decisions are also governed by economic considerations and preferences of the dentists and patients. Periodontal disease has recently been associated with CVD. When the analyses was repeated separately for men with and without a history of periodontal disease, tooth loss was associated with an increased risk of ischemic stroke in both groups, suggesting that factors other than antecedent periodontal disease may also be playing a role. The association between tooth loss and ischemic stroke is stronger than our previously reported association between tooth loss and coronary heart disease from the same population. Our results suggest that people who have lost teeth show higher incidence of ischemic stroke.
P77 Impaired Lower Lung Function and Coronary Heart Disease (CHD), The Atherosclerosis Risk in Communities(ARIC) Study
Verna L Lamar, Duanping Liao, Millicent Higgins, Wayne Rosamond, Gerardo Heiss, , PSU Hershey Medical Center, Hershey, PA, University of Michigan, Ann Arbor, MI, University of North Carolina, Chapel Hill, NC
The putative relationship between lung function(LF) and CHD remains
unclear. The aim of this study was to examine the association between
LF measured by forced expiratory volume at one
second(FEV1)and incident CHD after nine years of follow-up
of 14519 ARIC participants without CHD at baseline. An inverse
association between quartiles of LF and incident CHD (n=697 events)was
observed when adjusting for smoking and other CHD risk factors in
multivariable analyses (p for trend=0.02). However,
treating smoking as a covariate may not completely adjust for its role
as a confounder. After stratification by smoking, former and never
smokers in the lowest quartile of LF had the greatest risk of CHD.
After multivariable adjustment (see below)the magnitude of the
association was not as strong as reported by other cohort studies. We
conclude that the relationship between LF and CHD for the most part can
be explained by smoking. However, a nonsignificant excess risk of CHD
persisted among never smokers in the lowest quartile of lung
function.
|
P78 egg consumption and risk of cardiovascular disease in women
Frank B Hu, Meir J Stampfer, JoAnn E Manson, Graham A Colditz, Bernard Rosner, Frank E Speizer, Charles H Hennekens, Walter C Willett, Brigham and Women's Hospital, Boston, MA, Harvard School of Public Health, Boston, MA
Background. Restriction of egg consumption has been widely recommended to lower blood cholesterol and prevent coronary heart disease (CHD), but epidemiologic data on the relation of egg intake to the risk of CHD and stroke are sparse.
Methods. We examined the relation between egg consumption and incidence of CHD and stroke in a cohort of 80,082 women 34 through 59 years of age and without previously diagnosed CHD, stroke, cancer, hypercholesterolemia, or diabetes in 1980. Egg consumption was assessed at baseline and updated periodically during follow-up using validated dietary questionnaires.
Results. During 14 years of follow-up, we documented 939 cases of major coronary disease (658 nonfatal myocardial infarctions and 281 fatal CHD) and 563 cases of stroke (296 cases of ischemic stroke, 176 cases of hemorrhagic stroke, and 91 cases of unknown type). After adjustment for age, smoking, and other coronary risk factors, we found no evidence of an overall positive association between egg consumption and risk of CHD. The relative risks across categories of intake (<1/week, 1/week, 2-4/week, 5-6/week, 1+/day) were 1.0, 0.82, 0.99, 0.95, 0.82 (95% confidence interval 0.60-1.13, P for trend=0.95). Egg consumption was not significantly associated with either ischemic stroke (RR for 1+ eggs /day was 0.81, 0.46-1.42, P for trend = 0.81) or hemorrhagic stroke (RR for 1+ eggs /day was 1.07, 0.56-2.03, P for trend = 0.81).
Conclusions. Although we cannot exclude a small overall increase in risk or a larger elevated risk in some subgroups, our findings suggest that egg consumption within the range of our population (up to 1 egg/day) is unlikely to substantially increase risk of coronary disease or stroke in women.
P79 Independent associations of dietary fiber and saturated fat with fasting insulin in young black and white adults: The CARDIA Study.
Mark A Pereira, David S Ludwig, Candyce H Kroenke, Martha L Slattery, Linda Van Horn, David R Jacobs, Children's Hospital, Boston, MA, Northwestern University Medical School, Chicago, IL, University of Minnesota, Minneapolis, MN, University of Utah Medical School, Salt Lake City, UT
Recent epidemiologic evidence suggests that dietary fiber may play an
important role in the insulin resistance syndrome. We examined
associations of dietary fiber (g/1,000 kcal/day) and saturated fat
(%kcal/day), as measured by The CARDIA Diet History, with fasting
insulin concentration (µU/mL) measured in serum from exam year 7 of
young black (n=1,591) and white (n=1,885) adults of the CARDIA Study.
The table shows geometric means of insulin by mutually modeled
quintiles of dietary fiber and saturated fat with adjustment for age,
sex, energy intake, body mass index, waist-hip ratio, physical
activity, education, alcohol consumption, cigarette smoking, and
vitamin supplementation. There was an inverse dose-response
relationship between dietary fiber and fasting insulin in blacks and
whites. For saturated fat there was a positive dose-response
relationship in whites only. The coefficient of fasting insulin
regressed on dietary fiber was attenuated after inclusion of whole
grain intake in the model (regression coefficient reduced by 13% in
blacks and 42% in whites), while little attenuation was observed from
adjustment for refined grain, fruit, or vegetable consumption. We
conclude that the relation between dietary fiber and fasting insulin is
independent of other lifestyle factors, body habitus, and saturated
fat. Among food sources of dietary fiber, whole grains may be a
particularly important source of dietary fiber and related nutrients
which may offer protection from insulin resistance.
|
P80 Plasma homocysteine increases with menopause
A. Elisabeth Hak, Kees H. Polderman, Coen D.A. Stehouwer, Iris C.D. Westendorp, Albert Hofman, Jacqueline C.M. Witteman, Erasmus University Medical School, Rotterdam, Netherlands, Erasmus University Medical Shool, Rotterdam, Netherlands, Free University Hospital, Amsterdam, Netherlands
Background. Plasma homocysteine is possibly influenced by menopause, but data are inconsistent. Changes in homocysteine with menopausal status are difficult to study, due to the high correlation of menopausal status with age. Therefore we measured plasma homocysteine in a meticulously selected population in which the contrast between estrogen status between pre- and postmenopausal women of the same age was maximized.
Design. Plasma homocysteine was compared in 93 premenopausal and 93 postmenopausal women who were matched by age (range 43-55 years).
Setting. Women were selected from respondents to a mailed questionnaire about the menopause, which was sent to all women in the Dutch town of Zoetermeer (n=12,675, respons 61%).
Subjects.Postmenopausal women who were at least three years after menopause or whose menses had stopped naturally before age 48 were age-matched with premenopausal women with regular menses and without menopausal complaints.
Results.The mean age was 50.6 years in premenopausal and 51.1 years in postmenopausal women. Plasma homocysteine was related to menopausal status; the geometric mean was 10.70 µmol/l in premenopausal and 11.49 µmol/l in postmenopausal women (increase of 7%, 95% confidence interval 0.3 to 14%, p=0.036). Adjustment for creatinine and smoking did not influence this difference. The difference was present within three years after onset of menopause and did not show a trend with the number of postmenopausal years.
Conclusion.The result of this population based study indicate that plasma homocysteine is affected by menopause. The increase in homocysteine might add to the growing incidence of cardiovascular disease after menopause.
P81 A prospective study of plasma total homocysteine, MTHFR genotype, and risk of myocardial infarction and stroke in the elderly: the Cardiovascular Health Study.
Stephen M Schwartz, David S Siscovick, Manuel R Malinow, Mary Cushman, David L Hess, Jennifer Doherty, Rachel Pearce, David L Hess, Steven J Kittner, Walter H Ettinger, Lewis H Kuller, Russell P Tracy, Bowman Gray School of Medicine, Winston-Salem, NC, Oregon Regional Primate Research Center, Beaverton, OR, University of Maryland, Baltimore, MD, University of Pittsburgh, Pittsburgh, PA, University of Vermont, Burlington, VT, University of Washington, Seattle, WA
Few data exist on the association between plasma total homocysteine
(tHcy) and cardiovascular disease (CVD) in the elderly.
We used a case-cohort design to examine tHcy and its genetic and
nutritional determinants in relation to incident CVD within a
prospective study of 65+ year-old men and women. We included 290
myocardial infarction/fatal coronary heart disease (MI/CHD)
cases and 187 ischemic stroke cases identified during 5.5 years
of follow-up, and 818 participants in the subcohort. Adjusted for age,
gender, and CVD risk factors, the relative risks (RR) of MI/CHD
associated with increasing tHcy (<10, 10-12.9, 13-15.9, 16.0
µmol/L) were 1.0, 1.3 (95% CI 0.9, 1.8), 1.5 (95% CI 0.9, 2.3) and
2.2 (95% CI 1.5, 3.4) (trend p <0.001). The corresponding
results for stroke were: 1.0, 1.2 (95% CI 0.9, 1.8), 1.8 (95% CI 1.1,
2.9), 1.4 (95% CI 0.9, 1.8) (trend p = 0.085). The RRs
were strongest among persons with clinical CVD at baseline. Age and
gender-adjusted RRs for MTHFR TT677 were 1.5 (95% CI 0.9,
2.3) for MI/CHD and 1.7 (95% CI 1.1, 2.8) for stroke. Participants
with plasma folate, B6, and B12 levels all in the lowest three deciles
were at increased risk of MI/CHD (RR=1.8, 95% CI 0.8, 3.6) and stroke
(RR=1.9, 95% CI 0.8, 4.5), but a trend with increasingly poor plasma B
vitamin status was evident only for MI/CHD (trend p =
0.008). These data suggest that the risk of MI/CHD, and to a lesser
extent stroke, is associated with elevated tHcy among the elderly.
Clinical trials of tHcy-lowering therapies should include the elderly,
since the number of CVD cases potentially prevented in this population
may be particularly large.
P82 Homocysteine, Arteriosclerosis, & the "Reverse Causality" Hypothesis: Ignorance of Renal Function is Not Bliss
Andrew G Bostom, Linda Bausserman, Paul F Jacques, Gintaras Liaugaudas, Jacob Selhub, Irwin H Rosenberg, Jean Mayer USDA-HNRCA, Boston, MA, Memorial Hospital of Rhode Island, Pawtucket, RI, The Miriam Hospital, Providence, RI
Absent any tenable biological explanation, the hypothesis has been put
forth that: a) arteriosclerosis itself somehow
independently raises total homocysteine (tHcy) levels, and therefore,
b)observed associations between tHcy levels and
arteriosclerotic outcomes are spurious, and due to
"reverse causality". Based upon the proven link between subclinical
arteriosclerosis and
nephrosclerosis from autopsy studies, and the
exquisite, continuous association between glomerular
filtration rate (GFR)& tHcy levels, we provide objective data that
offers a reasoned alternative to the purely speculative reverse
causality hypothesis. We determined levels of cystatin C, a sensitive
indicator of subtle declines in GFR, along with tHcy, folate, B12,
pyridoxal 5'-phosphate,albumin, and creatinine
levels, in 91 consecutive patients with stable coronary artery
disease (CAD){mean age 62±8 years; 80.2% men}, whose
creatinine was 1.4 mg/dL. General linear modeling
with analysis of covariance revealed that cystatin C
was strongly and independently predictive (partial R= 0.351,;
p=0.001)of fasting tHcy levels, after adjustment for age, sex, vitamin
status, albumin, & creatinine. Given the known
powerful association between GFR & tHcy levels, we suggest that
subclinical decrements in renal function/GFR (discernible by cystatin C
levels) secondary to nephrosclerosis, & resultant
increases in tHcy, may antedate, & hence contribute to, the development
of clinical arteriosclerosis, including CAD.
P83 Methylene Tetrahydrofolate Reductase (MTHFR) and Nitric Oxide Synthase (ecNOS) Genes and Risks of Peripheral Arterial Disease and Coronary Heart Disease : Edinburgh Artery Study
Gerald R Fowkes, Amanda J Lee, Cathryn M Hau, Alexander Cooke, Michael Connor, Gordon DO Lowe, University of Edinburgh, Edinburgh, United Kingdom, University of Glasgow, Glasgow, United Kingdom
Hyperhomocysteinaemia and reduced nitric oxide synthesis may each result in endothelial dysfunction predisposing to atherogenesis. Genetic variants of MTHFR and ecNOS influence homocysteine metabolism and nitric oxide synthesis respectively and might increase the risk of atherosclerosis. The aim of our study was to identify, in a general population sample, the risks of peripheral arterial disease (PAD)and of coronary heart disease (CHD) related to MTHFR (175;198) and ecNOS (4;5) polymorphisms. In the Edinburgh Artery Study, 940 men and women aged 60 to 79 years had DNA extracted from a venous blood sample. Based on clinical examinations, three groups were identified: PAD (n=80), CHD (n=137) and healthy controls (n=300). Distributions of ecNOS and MTHFR genotypes did not differ significantly between groups with and without disease. However the ecNOS-4 allele (freq 0.13) was related to CHD, OR=1.45 (95% CI [0.91, 2.32], p=0.1) especially in non- smokers, OR=2.47 (95% CI [1.42,4.34], p=0.02). No association was found with PAD. The MTHFR-175 allele (freq 0.31) was not related to CHD, but with a reduced risk of PAD, OR=0.54 (95% CI [0.32, 0.90], p=0.02). Neither the ecNOS-4 allele or MTHFR-175 allele was related to the ankle brachial pressure index in the whole population. The ecNOS-4 allele was associated with a slightly increased risk of CHD, especially in non-smokers, but otherwise the MTHFR and ecNOS genotypes appeared to have little influence on risks of PAD and CHD.
P84 Relationship between occasional fish consumption and acute myocardial infarction in South Asian Indians
Linda E. Kelemen, Janice Pogue, Prem Pais, Sonia S. Anand, Salim Yusuf, McMaster University, Hamilton, ON, Canada, St. John's Medical College Hospital, Bangalore, India
South Asian (SA) Indians have a high rate of coronary heart disease (CHD). The reasons for this increased risk are unclear. We conducted an exploratory analysis to determine the relationship between vegetarianism, fish intake and first acute myocardial infarction (AMI) among SA who participated in a case control study in Bangalore, India. Three hundred AMI patients aged 30 to 60 years were prospectively recruited as cases and were age and sex matched with 300 hospital-based controls. Subjects completed a questionnaire assessing dietary and lifestyle habits and medical history. Fasting blood was analyzed for glucose(G) and lipids(L) according to a standardized protocol. Weight, height, waist(W) and hip(H) circumference were also measured. More controls than cases were vegetarian (55 vs 45%,P=0.067 for trend). Following adjustment for dietary (meat, alcohol, tea, coffee, dairy and type of oil intake) and non-dietary (W-H ratio, smoking, hypertension, family history of CHD, G and L) factors, multivariate analysis showed an inverse association for fish intake and risk of AMI (OR,0.45; 95%CI,0.23-0.90). Compared with subjects who consumed no fish, 1 fish meal per week was significantly associated with a 60% decrease for the risk of AMI (OR,0.40; 95%CI,0.16-0.96). When non-vegetarian non-fish eaters were compared to vegetarian non-fish eaters, the former was observed to have a 2-fold increase for risk of AMI (OR,2.01; 95%CI,1.14-3.56). These findings suggest that a vegetarian diet which includes fish may be protective for AMI in SA Indians even after adjusting for known CHD risk factors. To our knowledge, no study has previously reported this relationship in SA. Future studies are needed to further elucidate the complex relationship between meat and fish consumption and risk for AMI in this population.
P85 The Age-Associated Distribution of Serum Alpha-Tocopherol in an Adult Population: The Minnesota Heart Survey
Myron Gross, Kathryn H Schmitz, David R Jacobs, Russell V Luepker, Donna K Arnett, James Wessman, University of Minnesota, Minneapolis, MN
Serum alpha-tocopherol (vitE) may be protective
against recurrent myocardial infarction. Few studies have identified
its distribution and determinants in the general population. In
particular, the relationship between age and serum vitE is unclear. The
1995-96 Minnesota Heart Survey, a population-based random sample of
Midwestern adults included measurements of serum vitE, total
cholesterol (TC) and HDL in 859 nonfasting men and women
aged 25-84. Dietary intakes, smoking and dietary supplement use were
self-reported. Multiple regression assessed the association of vitE
with age, adjusting for total calories, ethnicity (82% white, 18%
black), and other covariates noted below. Serum vitE was higher at
older ages, independent of vitamin E in foods, supplements and other
covariates, see Table. Other independent positive associations with
serum vitE included TC-HDL, HDL, vitamin E supplement use, and dietary
intake of vitamin E, fiber, polyunsaturated and saturated fats.
Independent inverse associations included
monounsaturated fat intake, current smoker status,
and black ethnicity. These results suggest that less vitE is needed at
older ages to maintain serum vitE concentrations, perhaps from an
increase in absorption, choosing foods with highly bioavailable vitE
and/or lower metabolic turnover of serum vitE. Further
evaluation of changes in vitE pharmacokinetics with age are needed for
effective use in the prevention of CVD.
|
P86 Evidence that vitamin E in food, not supplements, protects Low-Density Lipoprotein (LDL) from oxidation in postmenopausal women.
Lori Mosca, Thomas Tarshis, Kathy Rhodes, Melvyn Rubenfire, Gilbert S. Omenn, The University of Michigan, Ann Arbor, MI
The role of antioxidant supplements in cardiovascular disease (CVD) prevention remains controversial. We studied the relation between the susceptbility of LDL to oxidation and the intakes of vitamin E and C, beta-carotene, and folic acid from food alone and from diet plus supplements in 54 postmenopausal women (mean age 67±8.7yrs). Reduced susceptibility of LDL to oxidation was estimated by an increased lag phase (prolonged time to maximal rate of conjugated diene production after addition of copper ions). Each subject had 2 fasting measures of lag phase and standard lipoproteins 1 week apart; the results were averaged. Dietary intake was evaluated using the average of 2 three day food records corresponding to lab draws and processed using Nutritionist IV. Our results show the average % energy intake from total fat was 27.2±6.7 and from saturated fat 8.4±2.9. There was a stronger Spearman correlation between lag phase and intake of vitamin E from food sources among non-supplements users (n=34, r=0.46, p<0.01) than for total intake in the entire sample (n=54, r=0.31, p=0.03). No correlation existed between vitamin E intake and lag phase in the supplement users only. In multiple regression models adjusted for BMI, total and LDL-cholesterol, triglycerides and exercise, vitamin E from food was a significant predictor of log lag phase (p=0.03, R2=.54). No significant association was found for total vitamin E intake and lag phase in regression models. Vitamin C, beta-carotene and folic acid intake were not predictors of lag phase in this study. We conclude that vitamin E from food sources is associated with a lower susceptibility of LDL to oxidation. This provides a mechanism for the observation that intake of foods rich in vitamin E are associated with lower rates of CVD in postmenopausal women.
P87 Effect of Garlic Supplementation on Plasma Lipids in Hypercholesterolemic Men and Women
Christopher D Gardner, Lorraine Chatterjee, Joseph Carlson, Stanford University, Palo Alto, CA
Background: Health claims regarding the cholesterol-lowering effect of garlic supplementation are widespread. However, the clinical trial data are inconsistent.
Objective: To evaluate the effect of two doses of a commercially available garlic supplement on plasma lipids, compared to a placebo, in moderately hypercholesterolemic adults (LDL-C 158 ± 19 mg/dL, mean ± SD, range 130 - 190 mg/dL).
Design: Double-blind, randomized, placebo-controlled, parallel design trial. Fifty-one men and women, 52 ± 8 years of age, were randomized to either: 1) Placebo, 2) 500 mg/day dried, powdered garlic (Half-Dose), or 3) 1,000 mg/day (Full-Dose). Plasma lipids were assessed every two weeks for 12-weeks.
Results: Adherence, dietary intake, physical activity and body weight were comparable at baseline and end-of-study among the three treatment arms. No significant between-group plasma lipid differences were detected.
Conclusions: Garlic supplementation in a moderately
hypercholesterolemic sample of men and women did not
significantly alter plasma lipid levels. Garlic supplementation may be
warranted for other outcomes not assessed here, but was not efficacious
in this study for the treatment of elevated
LDL-cholesterol.
|
P88 A Multi-Center, Self-Controlled Study Of Cholestin(TM) In Subjects With Elevated Cholesterol.
James Rippe, Kermit Bonovich, Harry Colfer, Michael Davidson, Carlos Dujovne, David Fried, Mitchell Greenspan, Stephen King, Ronald Karlsberg, Craig LaForce, Marc Litt, J. Robert McGhee, Burns Clinic, Petosky, MI, Cardiovascular Research Institute of So. CA, Beverly Hills, CA, Future Scripts, Spokane, WA, Healthy Options, Abingdon, MD, Jacksonville Heart Center, Jacksonville Be, FL, Kansas Foundation for Clinical Pharmacology, Overlans Park, KS, Lifemark Medical Center, Sellersville, PA, Mediquest Research Group, Ocala, FL, North Carolina Clinical Research, Raleigh, NC, Omega Medical Research, Providence, RI, Rush-Presbyterian-St. Luke's Medical Center, Chicago, IL, Tufts University, Shrewsbury, MA
A traditional Chinese food and medicine, red yeast rice (Cholestin) was evaluated in a multi-center, self-controlled, open-label study in 187 subjects (116 men, 71 women) with mildly to moderately elevated total cholesterol (TC) and LDL cholesterol (LDL-c). Eligible subjects were placed on the AHA Step 1 Diet throughout the study. After 4-weeks of this diet, Cholestin (2.4 g/day) was given for 8 weeks. Lipid profiles were determined at Week 4 (4 weeks of diet), and at Week 8 and 12 (4 and 8 weeks of Cholestin treatment), and at Week 14 (2 weeks after discontinuing Cholestin). The diet alone for 1 month had no effect on serum cholesterol. The 8-week treatment with Cholestin reduced TC by 16.4%, LDL-c by 21.0%, triglycerides (TG) by 24.5%, and ratio of total:HDL cholesterol by 17.7%, and increased HDL cholesterol (HDL-c) by 14.6%. Discontinuation of Cholestin following the Cholestin treatment led to a relatively rapid return of serum lipids to pre-study levels at Week 14. Product-related adverse reactions were mild or moderate in severity. Thirty-four patients (18%) were judged to have adverse reactions possibly or probably related to Cholestin treatment. The reported adverse events were headache, abdominal bloating, and gas. This multi-center clinical trial at 12 U.S. sites confirmed that treatment with a traditional Chinese food, Cholestin, was well tolerated and was effective in reducing TC, LDL-c, TG, and ratio of TC:HDL-c, and in increasing HDL-c in patients with hyperlipidemia.
P89 Elderly Patients with Primary Hyperlipidemia Benefited from Treatment with A Monascus purpureus Rice Preparation: A Placebo-Controlled, Double-Blind Clinical Trial
Shucun Qin, Weiqiang Zhang, Peng Qi, Meiling Zhao, Zhennan Dong, Yongchang Li, Xiaoman Xu, Xin Fang, Lei Fu, Jia-Shi Zhu, Joseph Chang, Beijing No. 22 Taipinglu Clinic, Beijing, People's Rep of China, Pharmanex, Inc., Simi Valley, CA, PLA General Hospital, Beijing, People's Rep of China, The Clinic of Beijing Institute of CAD, Beijing, People's Rep of China
This double-blind, placebo-controlled clinical trial evaluated the ability of a natural product, a Monascus purpureus (Red Yeast) rice preparation, to regulate blood lipids in elderly patients with primary hyperlipidemia. Seventy patients were randomly assigned to two groups: a M. purpureus rice preparation group and a placebo control group, 35 patients each group. Sixty-five patients completed the trial, 34 in the M. purpureus rice group and 31 in the placebo group. After eight weeks, Monascus purpureus rice preparation (1.2 g/day) significantly reduced serum total cholesterol by 25.9% and LDL-cholesterol by 32.8% (both p<0.001), whereas the control group (1.2 g/day placebo)showed 6.5% and 7.9% reductions, respectively. Notably this M. purpureus rice preparation, after 8 weeks, lowered serum triacylglycerols by 19.9% (p=0.02), this was significantly different (p=0.006) from the 2.3% increase observed in controls. When the overall effect of M. purpureus rice preparation was assessed relative to causing a change in the value of one or more lipid risk factors, 91.2% of patients in the treatment group showed a significant improvement in their lipid profile. Few side effects were observed during the 8-week treatment with M. purpureus (Red Yeast) rice preparation. We concluded that use of M. purpureus rice preparation as a dietary supplement is safe in elderly and may represent an effective, novel modality to manage elevated serum cholesterol and triacylglycerols that influence cardiovascular health.
P90 Hypo-cholesterolemia associated with superior working memory in healthy adult men
Matthew F. Muldoon, Christopher M. Ryan, Janine D. Flory, Karen A. Matthews, Stephen B. Manuck, University of Pittsburgh, Pittsburgh, PA, University of Pittsburgh School of Medicine, Pittsburgh, PA
This investigation explored the association between hypo- and
hypercholesterolemia and cognitive
performance. Subjects were healthy Caucasian adults aged 24-60
years recruited into one of three cholesterol groups based
upon low-density lipoprotein cholesterol concentration: low
(100 mg/dl, N=77), average (101 to 159 mg/dl, N=64), and high (160
mg/dl, N=153). None had any neurological disorder, or was a receiving
cholesterol-lowering or psychotropic medication. All were
administered a 90-minute neuropsychological test battery which assessed
five cognitive functioning domains — attention, psychomotor speed,
mental flexibility, working memory, and memory retrieval. Based upon
multivariate analysis of variance with age as a
covariate, cholesterol group was associated with working
memory (omnibus F4,574=3.4, p=.009). Subsequent
univariate analyses indicated that Associative
Learning scores (recall of unrelated word pairs) were inversely related
to cholesterol group in men but not women. (See table.)
This finding persists upon adjustment for years of education. These
results are consistent with, and extend, previous reports of
superior verbal knowledge in individuals with low
cholesterol levels.
|
P91 Do Changes in LDLc Through Menopause Predict Coronary and Aortic Atherosclerosis? Observations From the Healthy Women Study
Lewis H Kuller, Karen Matthews, Daniel Edmundowicz, Kim Sutton-Tyrrell, Clareann Bunker, University of Pittsburgh, Pittsburgh, PA, University of Pittsburgh, Pittsburgh, UT, University of Pittsburgh, Pittsburgh, PA
There is a very strong association between premenopausal LDLc measured
at age 47 and the amount of coronary and aortic calcification
based on EBCT 11 years later at the 8th postmenopausal visit (age
59,n=168)in the Healthy Women's Study. Approximately 9% of the women
with premenopausal LDLc <100 and 8% with LDLc 100-130mg% as compared
to 32% with LDLc 130-159, and 30% with LDLc >160 had high
coronary calcium scores 101 (P=.000). The LDLc increases
substantially during the pre- to postmenopausal period, 16 mg to the
1st postmenopausal visit and 20 mg to the 8th postmenopausal visit. The
change in LDLc from 1st to 8th postmenopausal visit was unrelated to
the extent of coronary or aortic calcification at the 8th post
visit. We then divided the baseline LDLc into four categories: <100
(n=66), 100-130 (n=67), 130-159 (n=22), and 160 (n=10), and then
their 8th postmenopausal LDLc into these same four categories. The
amount of coronary calcification was strongly related to the
LDLc at baseline but not to the changes in LDLc within each category.
Results were similar for measures of aortic calcification. Women on HRT
(about 50%) had a smaller increase in LDLc but premenopausal baseline
LDLc was still a significant predictor of extent of coronary or
aortic calcification. The LDLc levels measured at the 8th
postmenopausal visit were a much weaker predictor of the extent of
coronary calcification as compared to the premenopausal LDLc
levels. Postmenopausal LDLc levels are a weaker predictor of
atherosclerosis and classification for purposes of
therapies based on these LDLc's are also probably misleading. Finally,
effective therapy to reduce atherosclerosis among women
should probably begin in premenopause. Women can be classified into
their risk of atherosclerosis, even in premenopause at
age 47.
P92 Interaction of Blood Pressure and LDL Cholesterol in Early Atherosclerosis. The Los Angeles Atherosclerosis Study
Ping Sun, Kathleen M Dwyer, Noel Bairey Merz, Wei Sun, Lisa M Nicholson, Cheryl Nordstrom, James H Dwyer, Cedars Sinai Medical Center, Los Angeles, CA, University of Southern California, Los Angeles, CA
Background. The response to injury model of atherosclerosis has been investigated in animal models but not in epidemiologic studies. Relations between LDL cholesterol and carotid intima-media thickness (IMT) within levels of SBP provide a test of this model.
Methods. Data are from a longitudinal study of 573 randomly
sampled asymptomatic employees of a large company aged
40-60 years. IMT and change in IMT over 18 months (
IMT) were
determined sonographically in the common carotid artery. 497 subjects
were available for cross-sectional analysis. To investigate
interactive effects of SBP and serum LDL on IMT, linear slopes (ß±SE
in mm/mmol/L) of IMT regressed on LDL were computed within SBP
tertiles: Low 93-122, Middle 123-131, and High 132-175 mmHg.
Covariates were age, body height, sex, body mass index, ethnicity,
diabetes, smoking status, and treatment for hypertension or
hypercholesterolemia. Analysis of
IMT in 414 subjects was similar.
Results. In cross-sectional models, IMT was positively
related to LDL in the high SBP group (ß=0.028±.008, p=0.0006), but
not in the middle (ß=-0.005±.008, p=0.51) or low (ß=-0.003±.009,
p=0.78) SBP groups. These differences in slope between SBP groups were
statistically significant (p=0.004 for high vs middle, p=0.011 for high
vs low). Results were comparable for the longitudinal analysis:
IMT was significantly related to LDL in the high SBP group
(ß=0.013±.005, p=0.009), but not in the middle (ß=-0.006±.005,
p=0.18) or low (ß=-0.005±.005, p=0.31) groups. The differences in
slope between SBP groups were again significant (p=0.005 and p=0.010,
respectively).
Conclusion: These cross-sectional and longitudinal findings are consistent with the hypothesis that wall injury due to elevated SBP increases the susceptibility of the artery wall to LDL induced atherogenesis.
P93 Low-Density Lipoprotein Particle Size is a Risk Factor for Coronary Heart Disease Independent of Triglyceride and HDL Cholesterol: A Meta-Analysis of Three Prospective Studies in Men
Melissa A Austin, Alisa Kamigaki, John E Hokanson, University of Washington, Seattle, WA
Numerous studies have shown that decreased low-density (LDL) particle size is associated with increased risk of coronary heart disease (CHD). However, it is not clear whether this relationsip is independent of other lipoprotein risk factors. Thus, we applied the semi-quantitative methods of meta-analysis to 3 recent prospective studies: Physicians' Health Study (PHS),Stanford Five-City Project (5S), and Quebec Cardiovascular Study (QCS). Each study used a nested case-control design in primarily Caucasian men, aged 25-84 years, followed from 5-13 years and used gradient gel electrophoresis for LDL size.
As shown in the table, univariate ORs for a 10 angstrom decrease in LDL size and CHD risk were statistically significant for 2 of the 3 studies, with a summary OR of 1.6 that was also statistically significant. In multivariate analyses adjusting for triglyceride (TG), HDL cholesterol and other covariates, the summary OR was reduced to 1.3, but remained statistically significant. Due to differences in LDL size variances, this result reflects primarily the findings of the PHS and S5 studies.
Therefore, decreased LDL particle size is associated with a 30%
increased risk of CHD in men, independent of TG and HDL
cholesterol.
|
P94 Lifestyle Determinants of HDL-Cholesterol: The NHLBI Family Heart Study
R. Curtis Ellison, Yuqing Zhang, Sarah Knox, Donna K. Arnett, Michael A. Province, Boston University School of Medicine, Boston, MA, DECA, NHLBI, Bethesda, MD, University of Minnesota, Minneapolis, MN, Washington University, St. Louis, MO
Alcohol is known to increase HDL-cholesterol (HDL), but physicians often suggest other lifestyle changes as alternatives. We evaluated alcohol, exercise, & smoking as determinants of HDL in 2,309 randomly chosen subjecs in the NHLBI Family Heart Study.
Adjusting for age, BMI, education, estrogen use, and the other two
lifestyle factors, HDL increased in women from 54.3 (non-drinkers) to
58.0, 61.3, 63.7, and 67.6 mg/dl, respectively, for categories of 1,
1.1-4, 4.1-7, and >7 drinks/week. For men, with the highest category
divided into 7.1-14 and >14 drinks/week, HDL was 40.8 for non-drinkers
and 42.0, 45.3, 44.9, 49.4, and 49.8 mg/dl, respectively, with
increasing alcohol intake.
With total activity levels of <10, 10-30, 31-59, and 60 minutes/day,
similarly adjusted HDL values were 55.7, 58.4, 58.5, and 59.0,
respectively, for women and 42.9, 43.1, 43.0, and 45.9 mg/dl,
respectively, for men. Moderate and strenuous levels of activity,
assessed separately, gave similar results. In women, adjusted values of
HDL decreased from 58.2 for non-smokers to 56.0 for 1-19 and 47.7 mg/dl
for 20 cigarettes/day, respectively; for men in these categories, HDL
was 43.4, 42.8, and 41.3 mg/dl, respectively.
Overall, 17.6% of total variance of HDL was explained by the full model in women, and 15.3% in men. Alcohol, smoking, and exercise accounted for 29%, 9%, and <1% of this variance in women and 48%, 4%, and <1% in men.
These results suggest that exercise is associated less strongly with HDL than alcohol in both men and women; smoking has a large effect on HDL in women. In addition to advice to stop smoking, for appropriate subjects (e.g., older subjects without previous abuse and with no religious or medical reason not to drink), light drinking might be considered as a way for increasing HDL and decreasing risk of cardiovascular disease.
P95 Association Between Elevated Plasma Fibrinogen and the Small, Dense Low Density Lipoprotein Phenotype Among Postmenopausal Women.
Kevin C. Maki, Michael H. Davidson, Mary Sue Cyrowski, Ann C. Maki, Chicago Center for Clinical Research, Chicago, IL
A predominance of small, dense low density lipoprotein particles (LDL
subclass pattern B) has been associated with a 2- to 3-fold increase in
the risk for coronary artery disease. Recently published data
have shown a positive association between the levels of small, dense
LDL particles and plasma fibrinogen. This study investigated the
prevalence of LDL subclass pattern B in relation to fibrinogen
concentration among 258 postmenopausal women. The prevalence of LDL
subclass pattern B was 41.9% in the highest fibrinogen tertile (354
mg/dL), compared with 27.9% and 24.4% in the first and second
tertiles, respectively (global X2=6.8, p=0.03). The crude
odds ratio (OR) for LDL pattern B among women in the highest tertile
compared to those in the lower tertiles was 2.03 (95% confidence
interval 1.18-3.51, p=0.01). Multivariate logistic
regression analyses were run to adjust for potential
confounders. After adjustment for age and lipid variables (LDL
cholesterol, high density lipoprotein
cholesterol, loge triglycerides)
the OR for high (tertile 3) versus low (tertiles 1 and 2) fibrinogen
was 2.14 (95% CI 1.17-3.96). After adjusting for age, lipid
variables, and 14 non-lipid variables (including anthropometric
indicators, aspects of carbohydrate homeostasis, alcohol and cigarette
use, antihypertensive medication use, physical activity, dietary
patterns, and history of atherosclerotic disease) the OR was 2.56 (95%
CI 1.27-5.27). All ORs were significant at p<0.05. These data suggest
that hyperfibrinogenemia and LDL subclass pattern B may be two
components of a common syndrome. Future studies investigating the
relationship between LDL subclass pattern and
cardiovascular disease should take fibrinogen
concentration into account.
P96 Alternate measures of lipoprotein(a) detect an association with coronary artery disease in Black and White males but not in females.
Herbert J. Marx, Henry F.C. Weil, Thomas A. Pearson, Charles K. Francis, Paul L. Jenkins, Bhavna Solanki, Roberta G. Reed, Bassett Healthcare, Cooperstown, NY, Harlem Hospital Center, New York, NY
Elevated plasma lipoprotein(a) [Lp(a)] (30mg/dl) has been
associated with coronary artery disease (CAD) in some studies
involving Whites (W) but not in Blacks (B). In a study of B and W males
and females undergoing coronary arteriography at Harlem and
Bassett Hospitals, we found no correlation between presence of
significant CAD (75% stenosis of a major coronary
artery) and Lp(a) level measured by an ELIZA using antibodies specific
for apo(a) (Strategic Diagnostics Macra(TM)). However, in
those subjects with Lp(a)30mg/dl there was a significant association
(Wilcoxin Rank Sum Test, p<0.01) between CAD and Lp(a)
cholesterol content [Lp(a)C] and moles of Lp(a) [NMLPA]
measured by an ELIZA format which employs a trapping antibody specific
for apo(a) and a detection antibody specific for the apoB protein of
Lp(a) [Sigma Diagnostics Apo-Tek Lp(a)(TM)]. In subgroup
analysis, this association persisted for the 79 B males and 55
W males, but not for the 66 B females or 34 W females even though both
B and W had no significant differences between genders in levels of
Lp(a), Lp(a)C, and NMLPA. We conclude that for males with
Lp(a)30mg/dl, both Lp(a)C and NMLPA appear to provide additional
useful discriminating clinical information and that Lp(a) appears to
have a more important relationship to CAD in males than females.
P97 Patterns and Trends in Utilization of Lipid Lowering Medications from 1986 to 1995: The Atherosclerosis Risk in Communities (ARIC) Study
Rongling Li, Suzanne F Cook, Eyal Shahar, Jianwen Cai, Javier F Nieto, Andrew C Brown, Herman A Tyroler, GlaxoWellcome Inc., RTP, NC, Johns Hopkins University, Baltimore, MD, University of Mississppi, Jackson, MS, University of MN, Minneapolis, MN, University of NC, Chapel Hill, NC, University of North Carolina at Chapel Hill, Chapel Hill, NC
The NIH National Cholesterol Education Program Expert Panel established guidelines in 1989 and 1993 for the detection and treatment of serum cholesterol abnormalities. We assessed the patterns and trends of utilization of lipid lowering medications during 1986 to 1995, in the ARIC cohort, a random sample from four US communities. Participants in this study were 2,341 African-Americans and 8,599 whites, who attended all three visits, each three years apart. Use of lipid lowering agents (LLA) (during a 2-week period preceding each clinic visit) was more prevalent in the participants with high plasma LDL, pre-existing CHD, hypertension (HTN), and diabetes. Proportions (%) (95% CI) of LLA use adjusted for age, sex, education, LDL, CHD, HTN and diabetes were significantly higher in whites than African-Americans in each visit, 3.4 (2.9, 3.9) vs. 1.4 (1.0, 2.1) for visit 1, 7.6 (7.0, 8.2) vs. 3.8 (3.1, 4.7) for visit 2, and 8.5 (7.9, 9.3) vs. 3.5 (2.8, 4.4) for visit 3. Within race groups, there was no significant difference of LLA use between men and women. LLA use was not altered by different education levels nor income inequality in the four race gender groups. Mean LDL cholesterol (mg/dl) (95% CI), adjusted for age, sex, education, CHD, HTN, diabetes, LLA use and cholesterol lowering diet, significantly decreased from visit 1 to visit 3 in both African-Americans, 140.7 (139.0, 142.3), 136.7 (135.2, 138.3), 130.3 (128.9, 131.8) and whites, 136.4 (135.6, 137.3), 132.2 (131.4, 133.0), 125.9 (125.1, 126.6). In summary, an increased use of lipid lowering agents and improved profile of serum cholesterol were observed in the ARIC African-American and white participants from 1986 to 1995.
P98 Cholesterol Screening Misclassifies U.S. Children Due to Fixed Criteria
Darwin R Labarthe, Shifan Dai, Ronald B Harrist, School of Public Health, UT, Houston, TX, School of Public Health, UT Houston, Houston, TX
Background : Cholesterol screening under the
current NCEP guidelines results in misclassification of many children
and adolescents as not having an acceptable serum total
cholesterol concentration (TC). This results from use of
fixed absolute criteria to distinguish among acceptable (<170 mg/dl),
borderline (170 - 199 mg/dl), and high ( 200 mg/dl) TC for all ages
2-19 years, and both sexes, when there is actually marked fluctuation
in mean TC by age. Trajectories of TC by age also differ by sex. This
phenomenon was demonstrated previously by multilevel analysis
of data from Project HeartBeat!, a mixed longitudinal study of
dynamics of CVD risk factors from ages 8-18 years in Texas.
Hypothesis : We predicted consistent findings in the
US population as a whole, represented by the HANES III
data. Methods : Data were analyzed similarly for
this national sample of 2528 children aged 8-17 years: 741 black
females (BF), 694 black males (BM), 566 white females (WF), and 527
white males (WM). Results : TC trajectories of with age were
modeled for HANES III with significant effects of sex, race, and
age3 . For these 4 respective sex-race groups, over ages
8-17 years, the percentile ranks for TC = 170 mg/dl ranged by age
from 40-56 in BF, 37-65 in BM, 48-64 in WF, and 46-73 in WM; this value
was thought to approximate the 75th percentile when adopted
by the NCEP. The corresponding percentile ranks for TC = 200 mg/dl
ranged from 77-87 in BF, 75-92 in BM, 83-91 in WF, and 81-95 in WM;
this value was thought to approximate the 95th percentile
when adopted by the NCEP. Conclusion : Screening under the
NCEP criteria results in many more borderline and high TC than would be
identified by true 75th or 95th percentile
values; age- and sex-specific percentile values of TC should be the
screening criteria for TC in childhood and adolescence.
|
This article has been cited by other articles:
 |
|
 |
|
  M. L. Daviglus, K. Liu, L. L. Yan, A. Pirzada, D. B. Garside, L. Schiffer, A. R. Dyer, P. Greenland, and J. Stamler Body Mass Index in Middle Age and Health-Related Quality of Life in Older Age: The Chicago Heart Association Detection Project in Industry Study Arch Intern Med, November 10, 2003; 163(20): 2448 - 2455. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||