doi: 10.1161/01.CIR.0000142057.63388.A6
(Circulation. 2004;110:e305-e306.)
© 2004 American Heart Association, Inc.
Correspondence |
Do Meta-Analyses of Association Studies of Endothelial Nitric Oxide Synthase Variants and Ischemic Heart Disease Provide Conclusive Answers?
Department of Internal and Experimental Medicine, University of Padova, Padova, Italy
To the Editor:
A meta-analysis suggested that the Asp298 (894T) and an intron-4 single-nucleotide polymorphism (SNP), but not the promoter –786C allele, in the endothelial NO synthase (eNOS) gene would increase the risk of ischemic heart disease (IHD).1 Although based on studies of >23 000 subjects, this conclusion is amazing. The authors stated an increased susceptibility to cleavage of the Asp298-encoded eNOS enzyme,1 but this was disproved.2 Thus, no evidence for a functional role of these SNPs exists. In contrast, the –786C allele bears functional consequences, because it creates a binding site for a RPA-1 protein that blunts transcription3; it is also in linkage disequilibrium and interacts with the Asp298 in causing endothelial dysfunction.3 Serendipity or linkage disequilibrium with functional variants, as the –786C allele, in the eNOS and/or other genes might therefore explain the association of these SNPs with IHD.1
The meta-analysis was underpowered to detect an association of the –786C allele with IHD, because of the few available studies.1 Regrettably, two large studies that reported an association of this SNP with coronary artery disease were overlooked.4,5 Therefore, had the authors meta-analyzed these studies, their conclusions might have been altogether different.
There are several other caveats. First, because determinants and mechanisms of myocardial infarction and coronary artery disease differ considerably, a definition of IHD combining these phenotypes might be disagreeable. Second, the genotyping with RFLP performed in the meta-analyzed studies is affected by a high misgenotyping rate and should be replaced by more accurate methodologies based on allele-specific fluorescent resonance energy transfer probes.3 Third, a population-stratification bias should be considered when meta-analyzing studies of heterogeneous ethnicity. Fourth, by definition, cross-sectional studies of potentially fatal phenotypes, such as IHD, entail "survivors," thus exposing researchers to underestimation of the association of SNPs with IHD. Equally important, cardiovascular risk factors were generally more prevalent in IHD cases than in controls, thus introducing an important confounding factor when estimating this association. To circumvent this problem, in most quoted studies,1 the authors analyzed their data with multivariate logistic regression analyses. Unfortunately, this requires the use of individual patients’ data, rather than summary data, from each study and was unfeasible in the meta-analysis.1 Thus, although meta-analyses have furnished important information when an adequately large number of summary data were available and conclusions were drawn with caution, the use of multivariate techniques to meta-analyze high-quality genetic and clinical individual data are crucial to provide solid information on association of SNPs with IHD.
References
1. Casas JP, Bautista LE, Humphries SE, et al. Endothelial nitric oxide synthase genotype and ischemic heart disease: meta-analysis of 26 studies involving 23 028 subjects. Circulation. 2004; 109: 1359–1365.
2. Fairchild TA, Fulton D, Fontana JT, et al. Acidic hydrolysis as a mechanism for the cleavage of the Glu(298)
Asp variant of human endothelial nitric-oxide synthase. J Biol Chem. 2001; 276: 26674–26679.
3. Rossi GP, Taddei S, Virdis A, et al. The T-786C and Glu298Asp polymorphisms of the endothelial nitric oxide gene affect the forearm blood flow responses of Caucasian hypertensive patients. J Am Coll Cardiol. 2003; 41: 938–945.
4. Rossi GP, Cesari M, Zanchetta M, et al. The T-786C endothelial nitric oxide synthase genotype is a novel risk factor for coronary artery disease in Caucasian patients of the GENICA study. J Am Coll Cardiol. 2003; 41: 930–937.
5. Hirashiki A, Yamada Y, Murase Y, et al. Association of gene polymorphisms with coronary artery disease in low- or high-risk subjects defined by conventional risk factors. J Am Coll Cardiol. 2003; 42: 1429–1437.
Response
Centre for Clinical Pharmacology, Department of Medicine, British Heart Foundation Laboratories at University College London, London, UK, a.hingorani{at}ucl.ac.uk
Centre for Cardiovascular Genetics, Department of Medicine, British Heart Foundation Laboratories at University College London, London, UK
Department of Epidemiology and Population Health, London School of Hygiene and Tropical Medicine, London, UK
Department of Population Health Sciences, University of Wisconsin Medical School, Madison, Wis
We disagree that a functional role for Glu298Asp has been disproved or that a functional role for –786T>C has been demonstrated unequivocally. Fairchild et al1 suggested that preferential proteolytic cleavage of Asp298 observed in vitro was an artifact of nonspecific acid hydrolysis during sample preparation, but selective proteolysis of Asp298 in human vascular tissues, with a reduction in catalytic activity, has since been seen in samples prepared at neutral pH.2 Promoter-reporter studies of –786T>C have also been equivocal.3 Linkage disequilibrium between –786T>C and Glu298Asp polymorphisms is at best moderate.4
Rossi et al5 and Hirashiki et al6 published after our January 2003 deadline for data extraction and were not overlooked. We stated our meta-analysis was unlikely to detect a relative risk for the –786T>C polymorphism of 
1.2. Updating the meta-analysis to March 2004 (five and four additional studies of the Glu298Asp and –786T>C polymorphisms, respectively, including Rossi et al5 and Hirashiki et al6) added 3456 cases and 2559 controls. We limited analysis to an additive model for –786T>C because of incomplete genotype information in two studies. The updated random-effect summary-odds ratio (OR) for the –786T>C-ischemic heart disease (IHD) association (CC versus TT; 4882 cases and 9366 controls) was 1.30 ([95%CI: 1.01 to 1.66], P=0.04). Significant heterogeneity was observed (P=0.002), partially accounted for by Fatini et al.7 When excluded, the summary OR decreased to 1.21 (0.96 to 1.53; P=0.11). No publication bias was detected (P=0.32). These data suggest an effect of the –786T>C variant, but considerable uncertainty remains. In a cumulative meta-analysis, the –786T>C-IHD association became positive only in 2004, but with wide confidence intervals. The summary OR does not yet appear stable to additional data. In contrast, the summary OR for Glu298Asp has been stable since 2001, and the confidence intervals have narrowed (data not shown).
In response to the additional points raised, results were homogeneous for coronary artery disease and myocardial infarction, and ethnic background was not a significant source of heterogeneity. No study has compared genotyping methods, and thus the optimal assay remains to be defined. Survival bias is only likely to be large if the polymorphism influences lethality. There was no evidence for this in 3100 subjects followed for 8 years, although power may have been low.4 Confounding by other cardiovascular risk factors is an unlikely explanation of our findings. Inheritance of genetic variants is subject to a random assortment of maternal and paternal alleles at gamete formation, so the distribution of Glu298Asp and –786T>C variants should be independent of other risk factors.3,4 Indeed, one advantage of genetic over classical observational studies is that the problem of classical confounding is overcome.8 Heterogeneity between studies (and not confounding within studies) was evaluated by meta-regression, not "multivariate logistic regression analysis." Whether genetic risk is explained or modified by –786T>C or intron-4 variants requires further clarification. The data available currently provide the strongest support for an association between Glu298Asp and IHD.
References
1. Fairchild TA, Fulton D, Fontana JT, et al. Acidic hydrolysis as a mechanism for the cleavage of the Glu(298)Asp variant of human endothelial nitric-oxide synthase. J Biol Chem. 2001; 276: 26674–26679.
2. Persu A, Stoenoiu MS, Messiaen T, et al. Modifier effect of ENOS in autosomal dominant polycystic kidney disease. Hum Mol Genet. 2002; 11: 229–41.
3. Tsujita Y, Baba S, Yamauchi R, et al. Association analyses between genetic polymorphisms of endothelial nitric oxide synthase gene and hypertension in Japanese: The Suita Study. J Hypertens. 2001; 19: 1941–8.[CrossRef][Medline] [Order article via Infotrieve]
4. Jeerooburkhan N, Jones LC, Bujac S, et al. Genetic and environmental determinants of plasma nitrogen oxides and risk of ischemic heart disease. Hypertension. 2001; 38: 1054–61.
5. Rossi GP, Cesari M, Zanchetta M, et al. The T-786C endothelial nitric oxide synthase genotype is a novel risk factor for coronary artery disease in Caucasian patients of the GENICA study. J Am Coll Cardiol. 2003; 41: 930–937.
6. Hirashiki A, Yamada Y, Murase Y, et al. Association of gene polymorphisms with coronary artery disease in low- or high-risk subjects defined by conventional risk factors. J Am Coll Cardiol. 2003; 42: 1429–1437.
7. Fatini C, Sofi F, Sticchi E, et al. Influence of endothelial nitric oxide synthase gene polymorphisms (G894T, 4a4b, T-786C) and hyperhomocysteinemia on the predisposition to acute coronary syndromes. Am Heart J. 2004; 147: 516–21.[CrossRef][Medline] [Order article via Infotrieve]
8. Smith GD, Ebrahim S. Mendelian randomization: prospects, potentials, and limitations. Int J Epidemiol. 2004; 33: 30–42.
This article has been cited by other articles:
 |
|
 |
|
  J. P. Casas, G. L. Cavalleri, L. E. Bautista, L. Smeeth, S. E. Humphries, and A. D. Hingorani Endothelial Nitric Oxide Synthase Gene Polymorphisms and Cardiovascular Disease: A HuGE Review Am. J. Epidemiol., November 15, 2006; 164(10): 921 - 935. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. P. Rossi, G. Maiolino, M. Zanchetta, D. Sticchi, L. Pedon, M. Cesari, D. Montemurro, R. De Toni, S. Zavattiero, and A. C. Pessina The T-786C Endothelial Nitric Oxide Synthase Genotype Predicts Cardiovascular Mortality in High-Risk Patients J. Am. Coll. Cardiol., September 19, 2006; 48(6): 1166 - 1174. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L C Jones and A D Hingorani Genetic regulation of endothelial function Heart, October 1, 2005; 91(10): 1275 - 1277. [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||