One versus two inseminations per cycle in intrauterine insemination with sperm from patients' husbands: a systematic review of the literature

Abstract

Objective: To study the efficacy of performing two inseminations per cycle in IUI with husband's sperm compared with one insemination per cycle.

Design: Meta-analysis.

Setting: Randomized and prospective trials comparing two inseminations vs. one insemination per cycle in IUI with husband's sperm, retrieved by MEDLINE and Cochrane Library searches (1966-2001) and a manual search of the abstracts of the European Society of Human Reproduction and Embryology and American Society for Reproductive Medicine annual meetings (1990-2001).

Patient(s): A total of 865 patients underwent 1156 cycles of IUI with husband's sperm.

Intervention(s): After different ovarian stimulation protocols, one or two inseminations were performed.

Main outcome measure(s): Pregnancy rate per cycle. Detected studies were tested for homogeneity. Because heterogeneity was observed, DerSimonian-Laird relative risk with alleatory effects was used.

Result(s): Six randomized and prospective trials involving 865 patients and 1156 cycles were identified. There was remarkable heterogeneity among the different studies concerning methodology, especially regarding ovarian cycle management and the timing of inseminations. Although the pregnancy rate per cycle was somewhat higher in the two-inseminations-per-cycle group (14.9% vs. 11.4%), there were no statistically significant differences (relative risk = 1.34; 95% confidence interval 0.90-1.99).

Conclusion(s): No significant differences were observed when two inseminations per cycle were performed, compared with one insemination. There was great heterogeneity concerning ovarian management and insemination timing. This heterogeneity hampered the analysis. We detected a better pregnancy rate with two inseminations vs. one insemination when clomiphene citrate with or without gonadotropins and 5000 IU of hCG were used. More studies are necessary to ascertain whether this is true or merely an artifact from the multiple subgroups analysis.