Letters

Folate and risk of cardiovascular disease

Study results were misinterpreted

Authors' reply

Study results were misinterpreted

EDITORHung et al conclude that their cohort study on serum folate and coronary heart disease provides evidence against the view that folic acid prevents coronary heart disease.¹ We disagree and believe they have misinterpreted their results.

A meta-analysis of studies on homocysteine and cardiovascular disease, supported by others,^2-4 together with randomised trial evidence on folic acid dose and serum homocysteine reduction, shows that the maximal homocysteine lowering effect of folic acid occurs at a dose of about 0.8 mg/day (which increases serum folate by 20 µg/l).⁵ This homocysteine reduction lowers the risk of coronary heart disease by about 16%. The difference in average serum folate between the highest and the lowest folate group in the cohort study of Hung et al was about 7 µg/l, since the median of the highest folate group was about 8 µg/l (the 93rd centile was stated as 9 µg/l, so the 85th centile must have been less than this) and the median in the lowest folate group was about 1 µg/l (stated as <3 µg/l).

Since a 20 mg/l higher serum folate is associated with a 16% higher risk of coronary heart disease the 7 µg/l difference will be associated with a 5% higher risk, a relative risk of 1.05 (1.16^7/20) consistent with the 1.10 (men) and 1.14 (women) in their study. These results weigh in favour, not against, the view that folic acid reduces the risk of coronary heart disease. The narrow range of serum folate values in their cohort study limits its ability to show a significant effect.

Table 4 in the paper confirms the link between low serum folate and increased risk of coronary heart disease. Hung et al present data from six other cohort studies. Each shows a higher risk of coronary heart disease in the lowest folate group compared with the highest. This observation alone is highly significant. The probability that all seven studies have estimates above 1.0 by chance alone is ()⁷ or 1 in 128, a P value of 0.008. Their negative conclusion is inappropriate.

David S Wald, specialist registrar in cardiology. 
Southampton General Hospital, Southampton SO16 6YD davidwald{at}hotmail.com

Malcolm Law, professor. 
Joan Morris, senior lecturer. 
Nicholas J Wald, professor. 
Department of Environmental and Preventive Medicine, London EC1M 6BQ

Competing interests: None declared.

Authors' reply

EDITORWald et al oversimplify the relation between serum folate concentration and coronary heart disease, and they have ignored our results for red cell folate, a more reliable indicator of long term folate intake than serum folate. The reference interval for red cell folate in our cohort was 114-608 µg/l in men and 101-604 µg/l in women, indicating a wide range of folate values consonant to that reported in other general populations. The power of our study to detect a relative risk of 1.2 associated with a change of one standard deviation (130 µg/l for red cell folate) was about 90% for death from cardiovascular disease and 70% for death from coronary heart disease.

Wald et al state that serum folate needs to increase by 20 µg/l to lower the risk of coronary heart disease by 16% on the premise of lowering homocysteine concentrations by 3 µmol/l.¹ However, the Framingham heart study showed a difference >5 µmol/l in mean homocysteine concentrations across the population within a comparatively narrow range of serum folate values from 2 to 8 µg/l.² We found similar results in a general population.³

Furthermore, the effect of raising folate concentration can be underestimated by the reduction in mean homocysteine concentration. For example, a population increase in serum folate from 4.6 to 10.0 µg/l decreased the prevalence of high homocysteine (>13 µmol/l) from 18.7% to 9.8% but reduced mean homocysteine concentration by only 0.7 µmol/l.⁴

Folates should not be assumed to prevent coronary heart disease only through lowering homocysteine concentration as other mechanisms may exist.⁵ Hence we did not argue against public health efforts to raise folate consumption in the general population by appropriate dietary measures. We argued against the routine use of vitamin supplements to lower homocysteine concentrations in the general population until their benefit is proved by controlled clinical trials.

Joseph Hung, associate professor. 
jhung{at}dph.uwa.edu.au

John P Beilby, senior biochemist. 
Matthew W Knuiman, associate professor. 
Mark Divitini, research associate. 
School of Medicine and Pharmacology, University of Western Australia, Nedlands, WA 6009, Australia

Competing interests: None declared.