We investigated the value of the ankle-brachial systolic pressure index (ABPI) as a test for the diagnosis of peripheral arterial occlusive disease (PAOD) in general practice. ABPI measurements on 231 legs of 117 subjects performed in three general practice centers (GPC) were compared with the diagnostic conclusions of a Vascular Laboratory. The optimum cutoff value for the ABPI, its accuracy and diagnostic value were estimated. In a subpopulation of 51 subjects for whom repeated measurements were available, we checked whether taking the mean of three consecutive ABPIs for test outcome would enhance diagnostic performance. Receiver Operating Characteristic analysis showed that overall performance of the GPC ABPI was good (area under the curve approximately 0.9). Performing repeated ABPI measurements was superior to performing a single measurement. The optimum cutoff value for the ABPI was 0.97, associated with a diagnostic odds ratio (OR) of 17 and an accuracy of 81%. In a somewhat more selected subpopulation, the optimum cutoff value was 0.92 (OR 70, accuracy 90%). On the basis of our results, we suggest the following rule of thumb: if the ABPI < 0.8 or if the mean of three ABPIs < 0.9, it is highly probable that PAOD is present (PV+ > or = 95%); if the ABPI > 1.1 or if the mean of three ABPIs > 1.0, PAOD can be ruled out (PV- > or = 99%). In conclusion, in primary health care, the ABPI measurement can be a useful supplementary test in ambiguous diagnostic situations with regard to PAOD.