The failure to find anticardiolipin antibodies a risk factor for ischemic stroke may have more to do with the design of the Physicians' Health Study than the actual absence of the antibodies in stroke patients [1].

Rosove and Brewer [2] reported that the initial site of thrombosis in patients with antiphospholipid antibodies tended to predict the site of subsequent events. Patients with venous thrombosis tended to have recurrent venous thrombosis, whereas patients with arterial events such as stroke tended to have recurrent stroke and not a venous event. Patients were excluded from the Physicians' Health Study if there was a history or stroke or transient ischemic attack but not for a past venous thrombotic event [1].

The long interval between the drawing of the plasma sample and the events may also be important. Although anticardiolipin antibodies were not present in these patients 3 or 4 years before their stroke, they could have been present at the time of the stroke. Given the exclusion criteria of the study and the dictum that arterial events beget arterial events and venous events beget venous events, elevated anticardiolipin antibodies would be less likely to have been present at entry in patients who would eventually develop stroke than in patients likely to develop deep venous thrombosis.

Stroke is a heterogeneous disorder. In the age group studied, there are many possible causes that tend to dilute the effect of any one cause. The study sample size may not have had sufficient power to document a link between stroke and elevated anticardiolipin antibodies. Moreover, antiphospholipid antibodies are present in a larger proportion of young patients with strokes [3, 4]. A study in which younger men (in their 20s and 30s) were included at entry might have revealed an association between anticardiolipin antibodies and stroke.