Is GIFT (gamete intrafallopian transfer) the best treatment for unexplained infertility?

Dear Sir, The report of Murdoch et al. (1991) comparing GIFT with no treatment among 76 women with unexplained infertility was a thoughtful and well-analysed study. During 894 months of observation without treatment the average fecundability was 2.1% compared with 14% after 99 cycles of GIFT. Although GIFT seemed seven times better than no treatment, the authors were careful not to infer efficacy from the results of this comparison. Their caution is justified, because the different lengths of follow-up (a mean of 1.7 cycles of GIFT compared with 12.2 months of observation) introduce an unexpected bias that is related to better fecundability in the earlier months of observation. When the follow-up of a group of infertile couples is analysed, those who are more likely to get pregnant will conceive early in the observation period. Unless new couples are added, the remainder will have fewer pregnancies during the later months of observation. Thus, selecting couples according to their length of follow-up can affect the observed fecundability and the cumulative pregnancy rate. The cumulative pregnancy rates for livebirths among 340 untreated