ABSTRACT
Several other small trials of progesterone therapy were reported over the next two decades. Recently, interest in this therapy has been re-invigorated, as evidenced by the recent publication of six review articles and an ACOG Committee Opinion.3-9 The origin of this new enthusiasm and interest in progesterone was sparked by the publication, in 2003, of two randomized trials; one used progesterone vaginal suppositories and the other 17 alpha hydroxyprogesterone capoate (17P) injections to prevent recurrent preterm delivery.10,11
The results of the early reported trials of progesterone were evaluated by three different meta-analyses. The first of these, by Goldstein et al12 published in 1989, gave the results of a meta-analysis of randomized controlled trials involving the use of progesterone or other progestogenic agents for the maintenance of pregnancy. Fifteen trials of variously defined high-risk subjects were felt to be suitable for analysis. The trials employed six different progestational drugs. The pooled odds ratios (OR) for these trials showed no statistically significant effect on rates of miscarriage, stillbirth, neonatal death or preterm birth. The authors concluded that ‘... progestogens should not be used outside of randomized trials at present’. The second metaanalysis, by Daya,13 and published at the same time in the same journal, evaluated progesterone efficacy in treating women with a history of recurrent miscarriage, and found that progesterone therapy was effective in
increasing the likelihood of pregnancies reaching at least 20 weeks of gestation [OR 3.09, 95% confidence interval (CI) 1.28-7.42]. In a separate response to Goldstein, Keirse14 presented, in 1990, the results of an analysis of a more focused selection of trials. This meta-analysis was restricted to trials that employed 17P, the most fully studied progestational agent, and included all placebo-controlled trials which used this drug. Pooled OR found no significant effect on rates of miscarriage, perinatal death or neonatal complications. However, in contrast to Goldstein’s review, the OR for preterm birth was significant (OR 95% CI 0.30-0.85), as was the OR for birthweight 2500 g (OR 0.46, 95% CI 0.27-0.80). Keirse remarked that the results demonstrated by these trials contrast markedly with the poor effectiveness of other efforts to reduce the occurrence of preterm birth, but that since no effect was demonstrated to result in lower perinatal mortality or morbidity, ‘… further well-controlled research would be necessary before it is recommended for clinical practice’.
