The aim of this prospective study was to analyze the characteristics of intrahepatic cholestasis of pregnancy (ICP) in a French population. From 1989 to 1995 we studied 50 consecutive pregnant women with ICP (41 single, 7 twin, and 2 triplet pregnancies) referred for hepatologic consultation. All patients suffered from pruritus and/or jaundice associated with elevated fasting serum levels of total bile acids (mean 49 micromol/L, range 7-290). No patients had concomitant liver disease and all recovered normal liver function after delivery. Overall prematurity rate was 60%: 100% in multiple pregnancies and 41% in single pregnancies. Three of 61 babies died. Systematic clinical interviews revealed that 34 patients had been treated with oral micronized natural progesterone (200-1,000 mg/d) during the current pregnancy for risk of premature delivery, including at least 32 (64%) before the onset of pruritus. Onset of pruritus was statistically earlier in patients previously receiving progesterone than in patients not receiving progesterone (217 +/- 21 vs. 240 +/- 26 days, P < .01). This was also found in the single pregnancy subgroup of patients (222 +/- 19 vs. 240 +/- 26 days, P < .05). Pruritus disappeared before delivery in 10 of 50 patients, i.e., after withdrawal of progesterone in 7 patients (only one concurrently treated with cholestyramine), after decrease in dose of progesterone in 1 patient, and spontaneously in 2 patients. During the same period, the percentage of pregnant women without ICP who had been treated with progesterone during pregnancy was statistically lower than the percentage of patients treated with progesterone before the onset of pruritus in our group of patients with ICP (36% vs. 64%, P < .01, odds ratio 3.16, 95% CI:1.29-7.80). These results suggest that orally administered progesterone might be an exogenous factor which triggers ICP in predisposed women.