Abstract
The results of this study are not definitive because the stroke patients were not representative of all such patients in the community, and there are unknown selective factors which may have introduced biases into the case[long dash]control comparisons. However, the findings should be useful in formulating more precise hypotheses for investigation in a larger study.