Antithrombin III activity was measured prospectively in 127 pregnant women within four weeks of delivery, and the results were tabulated with respect to the clinical diagnosis of their hypertensive disorder. Plasma antithrombin III activity was significantly lower than controls in women with preeclampsia and in women with chronic hypertension and superimposed preeclampsia (P less than .001). In contrast, women with chronic hypertension alone had antithrombin III activities similar to controls. Based on discriminant analysis, an antithrombin III activity of less than 70% was selected as indicative of the preeclampsia-eclampsia syndrome. The sensitivity and specificity of plasma antithrombin III for preeclampsia was 76 and 91%, respectively. Importantly, an antithrombin III activity in excess of 70% accurately predicted the absence of preeclampsia in 89% of study patients. Although gestational ages at delivery were similar in preeclamptic women with antithrombin III activity above and below 70%, women with antithrombin III activity above 70% delivered larger infants and experienced less fetal distress during labor. These findings suggest antithrombin III measurement may be useful in the management of hypertensive pregnant patients who are unresponsive to bedrest.