To the Editor:

The role of anticardiolipin antibodies (aCL) as risk factors for stroke was examined in an incident case-referent study nested within the MONICA and Västerbotten Cohort Project.1 The authors conclude that while aCL were associated with future stroke, they did not constitute an independent risk factor. Their data do not support a causative role for IgA-aCL and IgG-aCL but neither do the data justify dismissing IgM-aCL as an independent risk factor.

On multivariate adjustment for hypertension, diabetes mellitus, current cigarette smoking, and smokeless tobacco, the odds ratio for future stroke associated with IgM-aCL decreased from 1.34 to 1.24 and was no longer statistically significant. However, the 95% confidence interval was fairly narrow (0.98 to 1.56). It is highly likely that a type II error could explain this loss of statistical significance. The authors gave no indication of the power of the study to detect the independence of aCL as risk factors for stroke. The effective sample size would also have decreased because of missing values. Data from Table 1 of the article suggests that several patients and controls had no data on the covariates included in multivariate logistic regression model on which the authors based their conclusion. For example, 11 patients and 15 controls had no data on hypertensive status.

Smoking was not an independent risk factor in this study, either. However, it would be imprudent to suggest on the basis of this study that smoking is not a risk factor for stroke.2 3 As the authors indicate, this is the only prospective study to report an association between aCL and stroke. Furthermore, few studies have examined the role of IgM-aCL. In light of this, the consequences of a misleading conclusion are important.

Some of the risk associated with IgM-aCL was explained by the other classic risk factors included in the regression model, and the antibodies could be the result of endothelial cell damage induced by these factors.4 On the other hand, they may have a causal role, and this needs to be further investigated. IgM-aCL increased the risk of stroke by only 24% in this study. If this is a true indication of the strength of the association, then future studies will need larger sample sizes to be able to assess the independence of IgM-aCl as a stroke risk factor.

Copyright © 2001 by American Heart Association