Effectiveness of assisted reproductive technology (ART).

Objectives: We reviewed the evidence regarding the outcomes of interventions used in ovulation induction, superovulation, and in vitro fertilization (IVF) for the treatment of infertility. Short-term outcomes included pregnancy, live birth, multiple gestation, and complications. Long-term outcomes included pregnancy and post-pregnancy complications for both mothers and infants.

Data sources: MEDLINE and Cochrane Collaboration resources.

Review methods: We included studies published in English from January 2000 through January 2008. For short-term outcomes, we excluded non-randomized studies and studies where a pregnancy or live birth rate per subject could not be calculated. For long-term outcomes, we excluded studies with fewer than 100 subjects and those without a control group. Articles were abstracted for relevant details, and relative risks or odds ratios, with 95 percent confidence intervals, were calculated for outcomes of interest for each study.

Results: We identified 5294 abstracts and (for the three questions discussed in this draft report) reviewed 1210 full-text articles and included 478 articles for abstraction. Approximately 80 percent of the included studies were performed outside the United States. The majority of randomized trials were not designed to detect differences in pregnancy and live birth rates; reporting of delivery rates and obstetric outcomes was unusual. Most did not have sufficient power to detect clinically meaningful differences in live birth rates, and had still lower power to detect differences in less frequent outcomes such as multiple births and complications. Interventions for which there was sufficient evidence to demonstrate improved pregnancy or live birth rates included: (a) administration of clomiphene citrate in women with polycystic ovarian syndrome, (b) metformin plus clomiphene in women who fail to respond to clomiphene alone; (c) ultrasound-guided embryo transfer, and transfer on day 5 post-fertilization, in couples with a good prognosis; and (d) assisted hatching in couples with previous IVF failure. There was insufficient evidence regarding other interventions. Infertility itself is associated with most of the adverse longer-term outcomes. Consistently, infants born after infertility treatments are at risk for complications associated with abnormal implantation or placentation; the degree to which this is due to the underlying infertility, treatment, or both is unclear. Infertility, but not infertility treatment, is associated with an increased risk of breast and ovarian cancer.

Conclusions: Despite the large emotional and economic burden resulting from infertility, there is relatively little high-quality evidence to support the choice of specific interventions. Removing barriers to conducting appropriately designed studies should be a major policy goal.