Background: The majority of studies in reproductive medicine only report on surrogate outcomes, such as the occurrence of ovulation or a positive pregnancy test. The primary aim of subfertility treatment is to help couples to get a healthy child. It has recently been proposed that the live birth of a healthy child should be the primary outcome in randomised controlled studies and meta analyses. However follow-up until live birth generates additional efforts and costs for researchers thus limiting the size and number of trials that are done.
Objective: We hypothesised that conclusions drawn from analysis of clinical and ongoing pregnancy are similar to conclusions drawn from analysis of live birth.
Methods: We searched the MSDG database for reviews with RCTs that reported both live birth and clinical pregnancy outcomes. Only those RCTs that reported on both outcomes were included. Sixty-two systematic reviews were identified which reported on both clinical pregnancy and live birth. The results were reported as relative risks with 95% confidence intervals and using adjusted estimates when available.
Results: Within those 62 reviews, 649 RCTs were incorporated. In the treatment groups, 30.3% of trials reported pregnancy and 24.6% reported live birth, a 5.7 % decrease in events reporting live birth. In the control groups, 27.1% of trials reported pregnancy, but only 21.4% reported live birth, again a 5.7% decrease in events reporting live birth.
Conclusions: There was a 5.7% difference between live birth and clinical pregnancy rates in both treatment and control groups. Live birth must be regarded as the primary outcome in fertility studies.