Since the advent of routine automated platelet counting we have observed unexplained periparturient thrombocytopenia (PPT) in an unexpected number of periparturient women, ie, during labor or within 24 hr postpartum. Mean +/- SD platelet count in 686 random blood donors was 236 +/- 50 X 10(9)/L and 1.02% had a platelet count less than 136 X 10(9)/L; in 2,204 random prenatal and postpartum women mean count was significantly higher (275 +/- 86 X 10(9)/L; p less than 0.001). Of 1,621 periparturient women, 74 (4.6%) had unexplained PPT (mean +/- SD platelet count 122 +/- 24 X 10(9)/L, range 21-135 X 10(9)/L, N = 74). Platelet count in PPT usually rose to normal within 1 week of delivery; in 10% thrombocytopenia persisted greater than 6 months. PPT occurred in successive pregnancies with normal intervening platelet counts. Nine of 34 newborns of mothers with PPT were thrombocytopenic; there was no correlation between mother's and baby's platelet counts. In no case of PPT was there excessive bleeding in mother or infant. Positive indirect platelet radioactive antiglobulin tests (PRAT) were seen in 11% of normal postpartum women and in 90% of 22 women with PPT; 65% of the positive tests in PPT were due to reactions with anti-C3 only. In contrast, pregnant women with autoimmune thrombocytopenic purpura (AITP) had positive PRAT primarily because of anti-IgG (+/- anti-C3); only 10% were positive only with anti-C3. Results were concordant in all of eight women with PPT tested by both indirect and direct PRAT. Amount of C3 bound per platelet in direct or indirect PRAT was not predictive of degree of thrombocytopenia, but there was correlation of fg C3 per platelet detected by the two assays in individual patients (r = 0.8). Mean levels of serum C3, C4, and factor B in women with PPT did not differ from normal; individual patients had abnormal serum complements but no characteristic pattern was observed. Increased immune complexes were observed in 6% of normal subjects and 33% of women with PPT. Etiology and mechanism of PPT is unclear. Despite lack of clinical evidence in women with PPT of syndromes associated with increased platelet destruction, the presence of preeclampsia cannot be absolutely excluded. Similarly, although the pattern of antiglobulin sensitization in PPT differed markedly from that seen in AITP, autoimmune disorder cannot be excluded. Alloantibodies did not appear to be responsible for PPT. While PPT is usually benign, some patients had a markedly reduced platelet count. Recognition of the phenomenon may be important in obstetrics.