Ursodeoxycholic acid does not improve outcomes for intrahepatic cholestasis of pregnancy

Référence

Chappelll LC, Bell JL, Smith A, et al. Ursodeoxycholic acid versus placebo in women with intrahepatic cholestasis of pregnancy (PITCHES): a randomized controlled trial. Lancet 2019;394:849-60.

Sommaire

Oral urodeoxycholic acid is a traditional treatment for cholestasis of pregnancy of unproven benefit. This report is of a multicenter double-blinded RCT undertaken in the UK to determine whether it reduces adverse perinatal outcomes and assess other maternal and perinatal and infant outcomes, as well as health care resource utilization. Cholestasis of pregnancy was defined as maternal pruritus occurring from 20 weeks gestation onward together with an increase in serum bile acid concentration above the upper limit of normal (per local laboratory criteria). Additional inclusion criteria were singleton or twin pregnancy, no known fetal anomaly, and age 18 years or older. Women were excluded if a decision for delivery within the next 48 hours had already been established, for known allergy to the active or other ingredient in the tablets, or higher order pregnancy. Placebo tablets contained the same ingredients as the active tablets except for the urodeoxycholic acid. Randomization was stratified on gestational age (< 34 weeks, 34 - <37, 37 or more) singleton or twin gestation, and highest recorded bile acid concentration prior to randomization (<40 micro mols, 40 or more). Women (n=605, 37 of whom had twin gestation) started with 500 mg tablets twice daily, which could be titrated up to 1000 mg twice daily if needed for better symptom control or reduced to once daily at provider discretion (e.g. gastrointestinal side effects). The primary outcome was a composite of perinatal death, preterm delivery before 37 weeks gestation, or neonatal unit admission of at least 4 hours (infants counted only once) and did not differ significantly between groups (74/322 [23%] with active treatment vs 85/318 [27%], RR 0.85, CI 0.62-1.15, P = 0.28). There were, likewise, no differences between groups on the components of the composite outcome, though worth mentioning that there were 2 deaths in the placebo group and 1 death in the active treatment group. Proportions of vaginal and cesarean births were also similar between groups. Maternal pruritus was assessed using a 100-mm visual analog scale was reduced to a statistically significant degree in the active treatment group, but modest. The mean difference that was less than the 30 mm that the authors considered to be clinically significant (mean difference 5.7 mm, CI 1.7-9.7, P=0.0054). There were no significant side effects attributed to treatment. Multiple additional assessments of secondary maternal and infant outcomes, including costs, as well as subgroup analyses revealed no important additional findings.