A meta-analysis of cerebrovascular disease and hyperhomocysteinaemia

Abstract

Hyperhomocysteinaemia has been identified as a risk factor for stroke and cerebrovascular disease in several studies. To evaluate the evidence we performed a meta-analysis. We found 21 studies searching Medline from 1966-July 1999 using the key words homocysteine, homocystine and cerebrovascular disease or stroke combined with a search of Embase, Science Citation Index and Biological Abstract. In 17 of these studies the populations were comparable. The studies were divided into two groups, cross-sectional studies and longitudinal studies where a pre-insult plasma or serum total homocysteine was used. The reports on 8 cross-sectional and 4 longitudinal studies gave data on the mean and standard deviations of plasma or serum homocysteine for both cases and controls, and these studies were included in the meta-analysis. The results of the 5 excluded studies all pointed to a positive relationship between hyperhomocysteinaemia and cerebrovascular disease. For each study, the expected fractions of the cases with total homocysteine higher than the 95-percentile for the controls were calculated, using the means and standard deviations, assuming a log-normal distribution, and the odds-ratios for disease with total homocysteine above the 95-percentile were computed. The overall weighted odds-ratio for disease with a concentration of homocysteine in plasma or serum above the 95-percentile (95% confidence interval) for the cross-sectional studies was 4.12 (2.94-5.77), for the longitudinal studies 3.74 (2.53-5.54), and for all 12 studies 3.97 (3.07-5.12). In conclusion, the results support the case for a strong relation between hyperhomocysteinaemia and cerebrovascular disease.