Thromboxane A2 and prostacyclin levels in molar pregnancyDear Sir. Roy et al. [Br J Obstet Gynaecol(l984) 91,908-9121 recently reported an interesting study on prostacyclin (PG12) in molar pregnancy showing that molar intravesicular fluid had very high levels of 6-keto-PGFI,, a major €'GIz metabolite. They reported no corresponding increase in maternal plasma 6-keto-PGF1,. Since they did not determine normal maternal plasma PGT, metabolite levels with their assay, this claim is hard to substantiate, as the variation between different radioimmunoassays (RIA) for PGIz metabolites is enormous. RIA is of great value in studying comparative values of PGI, metabolites using the same assay, but cannot give absolute values and therefore comparisons between groups studied by different RIA methods, as Roy et al. (1984) have done for their normal maternal plasma control range, is not possible. They also report that a series of normal amniotic fluids had a mean PG12 metabolite level of 34 (SD 17) pg/ml. suggesting that PGIl is normally present in amniotic fluid in the second trimester. We have measured PGI, metabolites in amniotic fluid and maternal plasma in 27 pregnancies between 18 and 22 weeks gestation. All patients were undergoing amniocentesis for genetic analysis or investigations for a possible neural-tube defect. We used a RIA previously described (Belch et al. 1983;Greer et al. 1985) with a lower limit of sensitivity of 5 pgiml, interand intra-assay variations were both less than 10% and recovery of added 6-keto-PGF,,was 95 (SD 9)%.Of the 27 pr_egnancies 26 were normal and in one the fetus had a myelomeningocoele. In all the normal patients amniotic fluid PGI, metabolities were undetectable whereas the normal maternal plasma level was 15.8 (SEM 0.7) pgiml, this compares with our previously reported mean level of 15.5 (SEM 1.05) pgiml for normal second trimester pregnancy (Greer et al. 1985). In the one abnormal pregnancy, however, amniotic fluid PGI, metabolites were 62 pgiml and the maternal plasma level was 170 pg!nil, the highest we have ever recorded in pregnancy. The reason for the high level is unclear but may be related to exposure of damaged highly vascular, neural tissue to amniotic fluid, and to placental transfer of the PG12 metabolites. While further studies on pregnancies complicated by neural-tube defects are required to substantiate this finding, it raises the possibility that high PGIz metabolite levels may be a marker for such defects. In addition it suggests that PG12 metabolites can cross into maternal plasma. We would therefore urge Roy et al. (1984)