Blood Levels of Homocysteine and Increased Risks of Cardiovascular Disease Causal Or Casual?
Total Page:16
File Type:pdf, Size:1020Kb
REVIEW ARTICLE Blood Levels of Homocysteine and Increased Risks of Cardiovascular Disease Causal or Casual? William G. Christen, ScD; Umed A. Ajani, MBBS; Robert J. Glynn, ScD; Charles H. Hennekens, MD Background: Accumulating data from epidemiologi- load) and/or a greater frequency of elevated homocyste- cal studies suggest that individuals with elevated blood ine level in persons with cardiovascular disease as com- levels of homocysteine have increased risks of cardio- pared with persons without cardiovascular disease. Re- vascular disease. We reviewed the currently available evi- sults of most prospective studies, however, indicated dence of an association between homocysteine and car- smaller or no association. The few prospective studies diovascular disease and examined whether the strength that reported a positive association between homocys- of the evidence varies according to study design. teine level and risks of cardiovascular disease included patients with preexisting vascular disease. Methods: We used a computerized MEDLINE litera- ture search, 1966 through September 1998, to identify Conclusions: In contrast to cross-sectional and case- all epidemiological studies that examined the relation- control studies, results of prospective studies indicated ship of homocysteine level with risks of coronary heart less or no predictive ability for plasma homocysteine in or cerebrovascular disease. Two measures of plasma ho- cardiovascular disease. Instead, elevated homocysteine mocysteine level and its association with risk of cardio- level may be an acute-phase reactant that is predomi- vascular disease were extracted: mean homocysteine level nantly a marker of atherogenesis, or a consequence of in cases and controls, and relative risk of cardiovascular other factors more closely linked to risks of cardiovas- disease for elevated homocysteine level. cular disease. Randomized trials are necessary to test re- liably whether lowering homocysteine levels will de- Results: A total of 43 studies were reviewed. Most cross- crease risks of cardiovascular disease. sectional and case-control studies indicated higher mean homocysteine levels (either fasting or after methionine Arch Intern Med. 2000;160:422-434 OMOCYSTINURIA is a rare level include vascular endothelial dysfunc- autosomal recessive con- tion,6-9 promotion of oxidation of low- dition usually resulting density lipoprotein cholesterol,10,11 vascu- from homozygous defi- lar smooth cell proliferation,12 and ciency of cystathionine coagulation abnormalities.13-15 Hb-synthase, an enzyme required in me- Accumulating data from epidemio- thionine metabolism for the conversion of logical studies suggested that individuals homocysteine to cystathionine. Patients with even moderately elevated levels of ho- with homocystinuria commonly have fast- mocysteine (eg, fasting blood levels exceed- ing plasma total homocysteine (sum of free ing approximately 16 µmol/L) have small plus protein-bound forms) levels exceed- to moderate increased risks of cardiovas- ing 250 µmol/L, compared with a refer- cular disease (CVD).16 Elevated homocys- ence range of 5 to 15 µmol/L in healthy teine levels are common in the general popu- 1,2 From the Division of Preventive subjects, and are prone to premature ath- lation; 21% of elderly participants in the Medicine, Department of erosclerosis and thromboembolism of the Framingham Study had plasma homo- Medicine, Brigham and extracranial and intracranial cerebral ar- cysteine levels exceeding 15.8 µmol/L.17 Women’s Hospital and Harvard teries and veins, the coronary arteries, and These can result from heterozygous de- Medical School (Drs Christen, the peripheral arteries and veins.3,4 While ficiency of cystathionine b-synthase Ajani, and Glynn), and other enzymatic defects can also produce or methylenetetrahydrofolate reductase Department of Biostatistics, homocystinuria, the observation 3 de- (an enzyme involved in remethylation of Harvard School of Public cades ago that vascular disease was com- homocysteine to methionine), or from Health (Dr Glynn), Boston, Mass. Dr Hennekens is now mon regardless of the source of the de- suboptimal intake of nutritional factors Visiting Professor of Medicine, fect suggested that homocysteine may be (folate, vitamin B6, vitamin B12) required for 18-23 and Epidemiology and Public responsible for the vascular abnormal- homocysteine metabolism. Whether Health, University of Miami ity.5 Plausible mechanisms to mediate a high homocysteine level is a cause of CVD School of Medicine, Miami, Fla. deleterious effect of high homocysteine is not yet clear. ARCH INTERN MED/ VOL 160, FEB 28, 2000 WWW.ARCHINTERNMED.COM 422 ©2000 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/30/2021 A recent review that combined and these data are presented sepa- measures of plasma homocysteine available data from cross-sectional, rately according to CVD type in this level and its association with risk of case-control, and a limited number review. If data in the original report CVD. of prospective studies concluded that were not presented separately for spe- 1. Mean homocysteine: This in- an elevated level of homocysteine was cific CVD end points, that study’s cludes mean level of homocysteine an independent risk factor for coro- overall results were included in the (micromoles per liter) (basal, fast- nary heart disease (CHD), cerebro- CVD category most frequently rep- ing, or post–methionine challenge) in vascular disease, and peripheral resented in the report (CHD or ce- cases and controls, and the associ- vascular disease.24 More recent pro- rebrovascular disease). Studies con- ated P value. In studies where over- spective data, however, tend to show ducted within a dialysis population all mean level of homocysteine for little or no association between ho- (see review25) or in patients with cases and controls was not pre- mocysteine levels and subsequent end-stage renal disease were not in- sented (eg, several studies presented CVD. Nevertheless, despite these cluded in this review. We also ex- only sex-specific values), we esti- more recent and apparently conflict- cluded 4 studies because of incom- mated the overall mean for cases and ing data, enthusiasm continues to be plete information,26 questionable controls from the reported data and expressed for screening of homocys- homocysteine values,27 or cases se- conducted statistical comparisons (t teine level as a regular component of lected on genetic or familial factors tests) by using a pooled estimate of CVD prevention, largely because the presumably related to homocyste- the variance. Because the raw data levels can be easily reduced by ad- ine level.28,29 were not available, more appropri- ministration of pyridoxine hydro- Studies were categorized as ate nonparametric tests (Wilcoxon chloride, cyanocobalamin, and folic cross-sectional, case-control, or pro- rank sum) could not be used for these acid. spective. Studies in which homo- comparisons. However, many stud- In this review we present the cysteine was measured at or after the ies reported the use of parametric tests currently available evidence regard- identification of cases were consid- (t tests) on loge-transformed data to ing homocysteine and CVD. We also ered either cross-sectional (if mem- compare homocysteine levels in cases consider the strengths and limita- bers of a defined population were ex- and controls, and several studies in- tions of completed cross-sectional, amined for the presence or absence dicated that the results for nontrans- case-control, and prospective stud- of the CVD end point) or case- formed and loge-transformed data ies, and suggest future directions. control (if a distinct control group were similar.23,30,31 was identified as a standard of 2. Elevated homocysteine level: METHODS comparison). Studies in which ho- For each study, we indicate the defi- mocysteine was measured before the nition of elevated homocysteine level We used a computerized MEDLINE identification of cases were consid- used by the investigators, the rela- English-language literature search, ered prospective and included pro- tive risk (RR) of CVD and 95% con- 1966 through September 1998, to spective cohort studies and case- fidence interval (95% CI) associated identify all epidemiological studies control studies nested within a with elevated levels of homocyste- that examined the relationship of ho- prospective cohort. In all study de- ine, and the covariates controlled for mocysteine level with risks of CHD signs, persons with the CVD end by matching or in analyses. When the or cerebrovascular disease. We also point were designated as cases and original report presented only the per- searched reference lists of all identi- those without the end point were centage of cases and controls that met fied articles for additional relevant designated as controls. the definition of elevated homocys- studies. Several components of homo- teine level, we estimated the RR of cysteine have been measured in the CVD by computing the odds ratio and Criteria for Inclusion included reports, and these compo- estimated the 95% CI by using the val- nents are indicated in the tables. Ho- ues of the 2 3 2 table according to There were 2 criteria for inclusion of mocysteine is a thiol-containing Woolf’s method.32,33 If the 2 3 2 table studies in the review: reporting of amino acid formed from the me- contained a value of 0 in any one of data for plasma homocysteine con- tabolism of methionine and is