From the Institutes for Research in Extramural Medicine and
Endocrinology, Reproduction, and Metabolism, and the Department of
Epidemiology and Biostatistics, Vrije Universiteit, Amsterdam, Netherlands.
Correspondence to Johannes Ruige, Institute for Research in Extramural Medicine, Vrije Universiteit, Van der Boechorststr 7, 1081 BT, Amsterdam, The Netherlands. E-mail jb.ruige.emgo{at}med.vu.nl
Methods and Results—Articles were identified by means of a
MEDLINE and Embase search and citation tracking. Eligible studies were
prospective population-based cohort studies and nested case-control
studies on the relationship between, on the one hand, fasting or
nonfasting insulin levels and, on the other hand, myocardial
infarction, death from coronary heart disease, and/or ECG
abnormalities. Data were extracted pertaining to insulin measurements,
type of outcome studied, adjustment for confounding, sex, mean age of
the study population, follow-up period, insulin assay, and ethnic
background (white or nonwhite). Associations of insulin and CVD were
reexpressed in a uniform manner, an estimate of relative risk (RR) and
95% CI, to be used in meta–regression analyses. Twelve of 17
potentially eligible articles provided sufficient information. Overall,
a weak positive association was found. The meta-analysis
resulted in an estimated summary RR (95% CI) of 1.18 (1.08 to 1.29)
for differences in insulin level, equivalent to the difference between
the 75th and the 25th percentiles of the general population in the
Netherlands. Ethnic background and type of insulin assay modified the
relationship between insulin and CVD with borderline significance.
Conclusions—Hyperinsulinemia is a weak risk
indicator for the occurrence of CVD. The relationship between
hyperinsulinemia and CVD was modified by ethnic
background and by the type of insulin assay involved.
Standardization of Study-Specific Associations
Sources of Heterogeneity and Summary
Estimates
Overall, this meta-analysis showed weak positive associations
between insulin levels and CVD. An increase of 50 pmol/L fasting
insulin resulted in a summary RR (95% CI) of 1.18 (1.08 to 1.29)
before stratification (Table 3
The question of whether insulin itself increases the risk for CVD,
independent of insulin resistance, cannot be answered by this
meta-analysis, because none of the included studies measured
both insulin and insulin resistance. Cross-sectional results of the
IRAS study, in which insulin levels and insulin resistance both were
measured in
The importance of ethnic background for the pathogenic mechanism is
suggested by studies showing a stronger longitudinal relationship in
white than in nonwhite populations. In studies of nonwhite populations,
however, the outcome (CVD) was assessed only by means of
electrocardiography, and the mean age of the
study population was generally lower. ECG abnormalities reflect
ischemia or previous myocardial infarction, whether or not
clinically manifest, but obviously they do not reflect sudden
coronary heart death. Clinical myocardial infarction and
coronary heart death, by definition, reflect CVD.
Theoretically, it is possible that the pathophysiology involved might
be slightly different between these two outcomes. Pooling individual
patient data from different studies would probably not reveal the
responsible determinant either, because in various studies the
distribution of determinants is highly influenced by the study design,
resulting in an unresolved correlation between
determinants.14 22 23 24 25 26 27 28 39 Ethnic background as a
modifier is further supported by the previously mentioned IRAS
study9 and by a discrepancy between low rates of
CVD and high rates of non–insulin-dependent diabetes mellitus in Pima
Indians.43 In our study, the type of insulin
assay involved turned out to be another potentially important modifier.
In contrast to expectations, studies with a specific insulin
assay27 28 had a high RR, despite recent findings
that showed the importance of proinsulin and split products in
accelerating
atherosclerosis.44 45 Again,
correlation with other study characteristics made it impossible to draw
definite conclusions on this issue. In future research, however,
detailed measurements of specific insulin, proinsulin, and insulin-like
molecules with specific assays44 45 are needed,
as well as specific measurements of insulin resistance.
This meta-analysis is based on a limited number of articles, of
which a substantial proportion provided insufficient information. A few
articles provided means (and SDs) of cases and controls to show the
relationship between insulin and CVD. Unfortunately, procedures to
estimate RRs (with 95% CI) from means (and SDs) for these studies
could not be performed, because insulin distributions are typically
highly skewed.11 Except for the
MRFIT34 and the Edinburgh
study,30 an attempt to gather additional
information from these studies by direct correspondence with the
authors provided no additional ways of calculating RRs. Another
limitation was "presentation bias," ie, some articles
provided more information on statistically significant than on
statistically nonsignificant associations. For example, although the
relevant data were collected,25 the "15-year
follow-up" article of the Paris Prospective Study did not provide
enough information on the nonsignificant association between fasting
insulin and CVD to reexpress the RR (95% CI). A more general
limitation in review of the literature is that it is prone to
publication bias. We explicitly investigated this by plotting the
number of cases versus effect magnitude, which resembled a
funnel,46 indicating the relative absence of
publication bias (data not shown). Estimates with a small sample size
were spread out over a wide range, and estimates with a large sample
size were spread out over a smaller range; no discontinuity could be
found. Publication bias might be limited for the above-mentioned topic,
because in this area negative as well as positive results are generally
considered to be clinically relevant.7 In the
meta-analysis, no assessment of the methodological quality of
the studies was made, because this is highly
arbitrary.21 47 48
The advantage of this meta-analysis is that it clearly
identifies sources that could potentially modify the relationship
between insulin and CVD. Moreover, it reveals issues that vary greatly
between studies (eg, number and type of confounding or intermediate
variables) or are in need of improvement (eg, correct
presentation of significant as well as nonsignificant
data). In conclusion, both fasting and nonfasting
hyperinsulinemia seem to be weak risk indicators
for the occurrence of CVD. Sources that could potentially modify the
relationship between insulin and CVD are ethnic background and type of
insulin assay.
Received August 16, 1997;
revision received October 23, 1997;
accepted November 13, 1997.
© 1998 American Heart Association, Inc.
Clinical Investigation and Reports
Insulin and Risk of Cardiovascular Disease
A Meta-Analysis
Abstract
Top
Abstract
Introduction
Methods
Results
Discussion
References
Background—Our purposes were to
estimate the strength of the longitudinal relationship between
hyperinsulinemia and cardiovascular
diseases (CVD) from the available literature and to identify study
characteristics that modify this relationship.
Key Words: insulin • cardiovascular diseases • follow-up studies • meta-analysis • epidemiology
Introduction
Top
Abstract
Introduction
Methods
Results
Discussion
References
The effect of
hyperinsulinemia on the occurrence of
cardiovascular disease (CVD) has been studied in
various large prospective studies but, as yet, no unequivocal
relationship has been established.1 2 It is known
that hyperinsulinemia precedes type II diabetes and
that it is associated with an adverse cardiovascular
risk profile. Type II diabetes carries a strongly increased risk for
CVD, but the role of hyperinsulinemia itself in
this process is not clear.3 Recent
articles4 5 6 suggest that
hyperinsulinemia reflects a compensatory mechanism
of decreased insulin sensitivity of the peripheral tissues
to insulin. This "insulin resistance" might be essential in the
pathogenesis of CVD and of type II diabetes.7 At
the moment, ongoing epidemiological studies investigating this
mechanism are directed toward measuring specific levels of insulin and
insulin resistance.8 9 Nevertheless, sources of
the heterogeneity that could explain earlier
conflicting results remain obscure.10 Previous
reviews on this issue1 2 3 4 5 have been narrative in
nature. Therefore, we decided to perform a meta-analysis to
estimate the strength of the longitudinal relationship between
hyperinsulinemia and CVD and to identify study
characteristics that modify this relationship.
Methods
Top
Abstract
Introduction
Methods
Results
Discussion
References
Eligible Articles
Articles were identified by means of an Index Medicus
(MEDLINE) and Embase search and by citation tracking covering the
period 1966 to 1996. Key words were insulin, prospective, cohort,
follow-up, cardiovascular, myocardial infarction, and
electrocardiography. Eligible for inclusion
were, on the one hand, prospective population-based cohort studies and
nested case-control studies on insulin levels and, on the other hand,
myocardial infarction, death from coronary heart disease,
and/or ECG abnormalities. Data were extracted pertaining to fasting and
nonfasting insulin levels, type of outcome studied (myocardial
infarction, death from coronary heart disease, ECG
abnormalities), confounders for which adjustment was made, sex, mean
age of the study population, follow-up period, type of insulin assay
involved, and ethnic background (white or nonwhite). If researchers
presented different follow-up periods on the same study
population, the study with the follow-up period closest to the mean of
the other study populations in the meta-analysis was selected.
Articles had to provide enough information to estimate a relative risk
(RR) and a 95% CI or an approximation, such as an odds ratio. We
abstracted the estimated RR, adjusted for the highest number of
potentially confounding variables, from each of the original
articles.
The objectives of this meta-analysis were to obtain a
summary estimate of the effect of hyperinsulinemia
on CVD and to explore sources of heterogeneity among
the RRs of the various studies. Analyses were performed
separately for fasting and nonfasting insulin levels. Most studies
provided fasting as well as nonfasting insulin levels, but only one
association per study population could be used per analysis.
This stratified analysis therefore prevented arbitrary choices
from being made but made it possible to investigate the differences in
RRs for fasting and nonfasting insulin levels. We obtained the summary
estimate across the different study populations by first estimating a
coefficient (b) that represented the relationship between
insulin and CVD per study and subsequently estimating a weighted
average of the coefficients. The weight of each study was calculated
inversely to the variance estimate of the coefficient after
reexpression of the SE in a uniform manner.11 For
cohort and nested case-control studies, the b represents the
coefficient for the effect of one standard unit difference (to be
specified) in insulin level in a Cox proportional hazards, logistic
regression, or Poisson regression model.12 The
reexpressed RR of cardiovascular disease per specified
uniform difference in insulin level is therefore exp(b), assuming that
the RR is constant during follow-up and the absolute risk is small. In
the same manner, the 95% CI can be calculated as exp(b±1.96 SE). For
studies that did not directly supply data that allowed the calculation
of b and its SE, the computation methods described by Greenland were
used.11 This calculation of the RRs (95% CI)
shows a relative risk for a difference of 50 pmol/L (fasting insulin
levels) or 250 pmol/L (nonfasting insulin levels). This approximates
the difference between the 75th and the 25th percentiles in the 50- to
74-year-old general population in the
Netherlands13 for fasting and nonfasting insulin
levels, respectively. If a study provided an RR without sufficient
specification of the difference in insulin levels involved, tables and
figures from the same article were used to estimate the difference in
insulin levels at issue. Articles from the Kuopio
study14 and the Busselton
study15 did not provide enough information in
tables and figures to identify the difference in insulin level for
which the RRs were calculated. Therefore, data on insulin distribution
in a Dutch general population was used to estimate the difference in
insulin levels that the RR refers to in the Kuopio study. We postulated
that the distributions of the two populations were similar, because
mean insulin levels of the Dutch population13 and
the Kuopio study14 were very similar. In the same
way, data from the Helsinki study16 were used to
estimate the difference in insulin levels for the RR of the Busselton
study.15 The latter studies had in common that
they measured insulin levels 1 hour after an oral glucose tolerance
test. Sensitivity analyses were performed to evaluate the
influence of these estimations on the final
conclusions.17
Univariate and multivariate
meta–regression analyses were used to identify study
characteristics that could explain differences in the relationship
between insulin and CVD.11 18 With this approach,
the logarithm of the study RR is regressed on study characteristics of
interest. Fixed-effect linear regression models were fitted by weighted
least squares.11 The fit of the weighted
regression model was evaluated by comparing the residual sum of squares
to a 
2
distribution.11 19 A small probability value
indicates a poor fit. The importance of various study characteristics
was evaluated according to the size of the b value as well as its CI.
Subsequently, summary estimates for the effect of
hyperinsulinemia on CVD were provided, stratified
according to the study characteristics that significantly modified the
relationship. The importance of study characteristics identified by
meta–regression analyses was confirmed by two different
meta-analytic techniques: first, heterogeneity tests of
pooled studies, of which small values of P indicate
differences in the RRs (95% CI) of these
studies,11 and second, summary RRs (95% CI) of
pooled studies, calculated according to fixed- as well as
random-effects models.20 21 Summary RRs (95% CI)
calculated by fixed-effects models imply that differences in the RRs
(95% CI) of pooled studies are due to sampling error. Summary RRs
(95% CI) calculated according to random-effects models make allowance
for unidentified sources of heterogeneity beyond
sampling error.21 This incorporation of possible
unidentified sources of heterogeneity in the
random-effects models results, in general, in a greater contribution of
smaller studies to the overall mean in the random-effects models than
in the fixed-effects models. Differences between summary RRs (95% CI)
calculated according to both models indicate unidentified sources of
heterogeneity, in which case the RR (95% CI) of the
random-effects model is the more appropriate. Otherwise, only RRs (95%
CI) of the fixed-effects model are
presented.20 21 Study characteristics
that were consecutively included in meta-regressions as possible
sources of heterogeneity were the type of outcome
studied (myocardial infarction and death from coronary heart
disease and/or ECG abnormalities), adjustment for confounding (both the
number of confounding variables and the presence or absence of
control for a specific confounder), sex (male versus female or mixed
population), mean age of the study population, length of follow-up
period, insulin assay (specific insulin assay versus potential
cross-reactivity with proinsulin-like molecules), and ethnic background
(white versus nonwhite). Analyses were performed with the
SPSS-PC software package, version 5.0.
Results
Top
Abstract
Introduction
Methods
Results
Discussion
References
Twenty-two potentially eligible articles with data on insulin
levels and CVD were identified.14 15 16 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Twelve articles describing 17 different studies were included in the
meta–regression analyses and are listed in Table 1
.14 15 22 23 24 25 26 27 28 30 34 39
The coefficient, corresponding SE, and unit increment (specified
difference in insulin level) derived from the original articles, before
reexpression, are shown. Subsequently, RRs and 95% CIs are
presented, after reexpression for a difference of 50 pmol/L
fasting or 250 pmol/L nonfasting insulin. The
Busselton,15 Helsinki,16 23
and British Regional Heart27 studies provided
data on different follow-up periods of the same study populations. One
study per population was chosen, the closest to the mean follow-up
period of the remaining studies. Sensitivity analyses revealed
that using a study on any of the other reported follow-up periods had
no substantial influence on the final
conclusions17 (data not shown). Eleven articles
did not provide sufficient quantitative information for this
meta-analysis. Two studies30 34 could be
included after additional details were provided by direct
correspondence and are listed in Table 1. Five articles described study
populations that had already been
included,35 36 37 38 40 and 4 other articles are
listed in Table 2.29 31 32 33 Of
these 4 articles, 1 found a negative association between
hyperinsulinemia and CVD,32
and 3 did not find an association.29 31 33
View this table:
[in a new window]
Table 1. Summary of Results of Prospective Studies Included
in Meta–Regression Analyses of Insulin and
Cardiovascular Disease
View this table:
[in a new window]
Table 2. Summary of Results of Prospective Studies Not
Included in Meta–Regression Analyses of Insulin and
Cardiovascular Disease ). However,
the meta-analyses revealed heterogeneity across
studies of nonfasting insulin and CVD (probability value of
heterogeneity test, P=.007, Table 3). The
heterogeneity might be explained by a difference
between studies involving white and nonwhite populations. Separate
summary RRs were 1.04 (0.93 to 1.16) for studies of nonwhite
populations and 1.42 (1.23 to 1.65) for studies of white populations,
respectively. Unfortunately, it was not possible to identify the study
characteristic responsible for this heterogeneity,
because ethnic background, mean age of study population, and the type
of outcome studied were highly correlated within these studies. All
studies of nonwhite populations had younger subjects and used only ECG
abnormalities as outcome, in contrast to studies of white populations,
which had older subjects and used clinical myocardial infarction or
death from coronary heart disease as outcome (exception: see
Reference 2828 ). Another source of heterogeneity across
studies, identified with meta-regression analyses, was the type
of insulin assay involved. Although the probability value of the
heterogeneity tests (P=.09, fasting studies
in whites; P=.11, nonfasting studies in whites; Table 3)
does not confirm strongly that this study characteristic induces
heterogeneity, meta–regression analyses are
generally regarded to be more sensitive in revealing sources of
heterogeneity.11 In the fasting
and nonfasting studies, only one study involved a specific insulin
assay, and they both had higher RRs. Again, however, interference with
other study characteristics could not be excluded. One of the
studies27 measured nonfasting insulin without
using an oral glucose tolerance test, and the
other28 had a nested case-control design. A study
that involved a nonspecific insulin assay and nested case-control
design had a population selection of middle-aged men with a high-risk
profile.34 Adjustment for confounding varied
greatly across studies for both type and number for which adjustment
was made, as is shown in Table 4. In
general, most studies adjusted only for a limited number of
confounders. More than nine studies adjusted for age, body mass index,
smoking, blood pressure, glucose level, cholesterol, and
triglycerides. In our meta-regression, neither the number
of confounding variables for which adjustment was made, the
presence or absence of control for one specific confounder, length of
follow-up period, nor sex differences in the study populations modified
the association between insulin and CVD.
View this table:
[in a new window]
Table 3. Meta-Analysis of Relationship Between
Insulin and Cardiovascular Disease
View this table:
[in a new window]
Table 4. Summary of Confounding Variables Adjusted for in
Prospective Studies on Insulin and Cardiovascular
Disease
Discussion
Top
Abstract
Introduction
Methods
Results
Discussion
References
In contrast to previous narrative
reviews,1 2 3 4 5 this meta-analysis provides
a quantitative estimate of the strength of the longitudinal
relationship between insulin and CVD and systematically investigates
which study characteristics could be responsible for the
heterogeneity of this relationship. The overall
relationship between insulin and CVD turned out to be weak, and ethnic
background as well as type of insulin assay involved were identified as
potentially important study characteristics. The strength of the
relationship was therefore presented for the different
categories of these characteristics. However, both ethnic background
and type of insulin assay involved were correlated to other study
characteristics, and this might also be a cause of the modification
effect. Thus, definite conclusions cannot yet be made on the basis of
this meta-analysis alone,41 but it
stresses the importance of further research on these issues. Another
remarkable finding was the great variety across studies in the number
and type of confounding variables for which adjustment was made.
Although the absolute number of confounders adjusted for in the
individual studies does not appear to be a modifier of the relationship
between insulin and CVD, more research into the role of individual
confounders or intermediates is also clearly
needed.2 The ability of this
meta-analysis to identify important study characteristics was
limited by the fact that some investigators adjusted for a particular
study characteristic of interest, whereas other investigators excluded
the same study characteristic by design (eg, sex, glucose intolerance).
In future, a more uniform approach would facilitate meta–regression
analyses to identify sources of heterogeneity
and thus contribute to insight into pathogenic mechanisms. 
1400 subjects, did not show an independent association
between insulin level and CVD.9 A recent
experimental study also found no evidence of a role of exogenous
insulin in accelerating
atherosclerosis.42
Acknowledgments
We thank T.J. Orchard, G.A. Grandits, and R.A. Riemersma for
their contributions in providing additional details on the MRFIT and
the Edinburgh Study.
References
Top
Abstract
Introduction
Methods
Results
Discussion
References
This article has been cited by other articles:
 |
|
 |
|
  J. Frystyk, R. A. Ritzel, J. Maubach, M. Busing, R. Luck, J. Klempnauer, W. Schmiegel, and M. A. Nauck Comparison of Pancreas-Transplanted Type 1 Diabetic Patients with Portal-Venous Versus Systemic-Venous Graft Drainage: Impact on Glucose Regulatory Hormones and the Growth Hormone/Insulin-Like Growth Factor-I Axis J. Clin. Endocrinol. Metab., May 1, 2008; 93(5): 1758 - 1766. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Ashikaga, M. Nishizaki, H. Fujii, K. Ihara, S. Niki, T. Murai, S. Maeda, N. Yamawake, Y. Kishi, and M. Isobe Coronary Endothelial Dysfunction and Impaired Microcirculation Response to Atrial Natriuretic Peptide in Hyperinsulinemia Journal of Cardiovascular Pharmacology and Therapeutics, March 1, 2008; 13(1): 58 - 63. [Abstract] [PDF]
|
|
|
|
 |
|
 F. E. Hirai, S. E. Moss, B. E.K. Klein, and R. Klein Relationship of Glycemic Control, Exogenous Insulin, and C-Peptide Levels to Ischemic Heart Disease Mortality Over a 16-Year Period in People With Older-Onset Diabetes: The Wisconsin Epidemiologic Study of Diabetic Retinopathy (WESDR) Diabetes Care, March 1, 2008; 31(3): 493 - 497. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Amano, T. Matsubara, T. Uetani, M. Nanki, N. Marui, M. Kato, T. Yoshida, K. Arai, K. Yokoi, H. Ando, et al. Abnormal Glucose Regulation Is Associated With Lipid-Rich Coronary Plaque: Relationship to Insulin Resistance J. Am. Coll. Cardiol. Img., January 1, 2008; 1(1): 39 - 45. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Sarwar, N. Sattar, V. Gudnason, and J. Danesh Circulating concentrations of insulin markers and coronary heart disease: a quantitative review of 19 Western prospective studies Eur. Heart J., October 2, 2007; 28(20): 2491 - 2497. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. J.G. Hanley, L. E. Wagenknecht, A. Festa, R. B. D'Agostino Jr., and S. M. Haffner Alanine Aminotransferase and Directly Measured Insulin Sensitivity in a Multiethnic Cohort: The Insulin Resistance Atherosclerosis Study Diabetes Care, July 1, 2007; 30(7): 1819 - 1827. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. H. Oterdoom, A. P. J. de Vries, R. T. Gansevoort, P. E. de Jong, R. O. B. Gans, S. J. L. Bakker, and for the PREVEND Study Group Fasting insulin modifies the relation between age and renal function Nephrol. Dial. Transplant., June 1, 2007; 22(6): 1587 - 1592. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. A. Rettenbacher, M. Hummer, A. Hofer, S. Baumgartner, C. Ebenbichler, M. EdLinger, G. Kemmler, M. Lechleitner, and W. Wolfgang Fleischhacker Alterations of glucose metabolism during treatment with clozapine or amisulpride: results from a prospective 16-week study J Psychopharmacol, June 1, 2007; 21(4): 400 - 404. [Abstract] [PDF]
|
|
|
|
|
|
A. Avogaro, C. Giorda, M. Maggini, E. Mannucci, R. Raschetti, F. Lombardo, S. Spila-Alegiani, S. Turco, M. Velussi, E. Ferrannini, et al. Incidence of Coronary Heart Disease in Type 2 Diabetic Men and Women: Impact of microvascular complications, treatment, and geographic location Diabetes Care, May 1, 2007; 30(5): 1241 - 1247. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. S. Johansson and H. J. Arnqvist Insulin and IGF-I action on insulin receptors, IGF-I receptors, and hybrid insulin/IGF-I receptors in vascular smooth muscle cells Am J Physiol Endocrinol Metab, November 1, 2006; 291(5): E1124 - E1130. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C.-H. Tseng Exogenous Insulin Use and Hypertension in Adult Patients With Type 2 Diabetes Mellitus. Arch Intern Med, June 12, 2006; 166(11): 1184 - 1189. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. K. Rutter, J. B. Meigs, L. M. Sullivan, R. B. D'Agostino Sr, and P. W. Wilson Insulin Resistance, the Metabolic Syndrome, and Incident Cardiovascular Events in the Framingham Offspring Study Diabetes, November 1, 2005; 54(11): 3252 - 3257. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Meyer, B. P. McGrath, J. Cameron, D. Kotsopoulos, and H. J. Teede Vascular Dysfunction and Metabolic Parameters in Polycystic Ovary Syndrome J. Clin. Endocrinol. Metab., August 1, 2005; 90(8): 4630 - 4635. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A.A. Lteif, K. Han, and K.J. Mather Obesity, Insulin Resistance, and the Metabolic Syndrome: Determinants of Endothelial Dysfunction in Whites and Blacks Circulation, July 5, 2005; 112(1): 32 - 38. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Alssema, J. M. Dekker, G. Nijpels, C. D.A. Stehouwer, L. M. Bouter, and R. J. Heine Proinsulin Concentration Is an Independent Predictor of All-Cause and Cardiovascular Mortality: An 11-year follow-up of the Hoorn Study Diabetes Care, April 1, 2005; 28(4): 860 - 865. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. M. Stuveling, S. J. L. Bakker, H. L. Hillege, P. E. de Jong, R. O. B. Gans, and D. de Zeeuw Biochemical risk markers: a novel area for better prediction of renal risk? Nephrol. Dial. Transplant., March 1, 2005; 20(3): 497 - 508. [Full Text] [PDF]
|
|
|
|
|
|
T. Takamiya, W. R. Zaky, D. Edmundowics, T. Kadowaki, H. Ueshima, L. H. Kuller, and A. Sekikawa World Health Organization-Defined Metabolic Syndrome Is a Better Predictor of Coronary Calcium Than the Adult Treatment Panel III Criteria in American Men Aged 40-49 Years Diabetes Care, December 1, 2004; 27(12): 2977 - 2979. [Full Text] [PDF]
|
|
|
|
 |
|
 S Jadhav, J Petrie, W Ferrell, S Cobbe, and N Sattar Insulin resistance as a contributor to myocardial ischaemia independent of obstructive coronary atheroma: a role for insulin sensitisation? Heart, December 1, 2004; 90(12): 1379 - 1383. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. C. Eringa, C. D. A. Stehouwer, G. P. van Nieuw Amerongen, L. Ouwehand, N. Westerhof, and P. Sipkema Vasoconstrictor effects of insulin in skeletal muscle arterioles are mediated by ERK1/2 activation in endothelium Am J Physiol Heart Circ Physiol, November 1, 2004; 287(5): H2043 - H2048. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. M. Punjabi, E. Shahar, S. Redline, D. J. Gottlieb, R. Givelber, and H. E. Resnick Sleep-Disordered Breathing, Glucose Intolerance, and Insulin Resistance: The Sleep Heart Health Study Am. J. Epidemiol., September 15, 2004; 160(6): 521 - 530. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. A. Kronmal, J. I. Barzilay, R. P. Tracy, P. J. Savage, T. J. Orchard, and G. L. Burke The Relationship of Fasting Serum Radioimmune Insulin Levels to Incident Coronary Heart Disease in an Insulin-Treated Diabetic Cohort J. Clin. Endocrinol. Metab., June 1, 2004; 89(6): 2852 - 2858. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. B. Meigs, F. B. Hu, N. Rifai, and J. E. Manson Biomarkers of Endothelial Dysfunction and Risk of Type 2 Diabetes Mellitus JAMA, April 28, 2004; 291(16): 1978 - 1986. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Z. T. Bloomgarden Definitions of the Insulin Resistance Syndrome: The 1st World Congress on the Insulin Resistance Syndrome Diabetes Care, March 1, 2004; 27(3): 824 - 830. [Full Text] [PDF]
|
|