Six randomised controlled trials with 458 participants (548 stimulated cycles, 514 retrievals, 181 tubal transfers, and 207 intra-uterine transfers) were included in the analysis.
Heterogeneity was found in only one outcome measure, namely pregnancy rate per transfer (P<0.08).
Implantation rates.
There was a total of 72 (15%) documented gestational sacs in 481 replaced embryos in ZIFT, and 63 (12%) in 523 replaced embryos in IVF-ET. The implantation rate between the two transfer methods was not significantly different (OR 1.25, 95% CI: 0.87, 1.8).
Pregnancy rate per transfer.
Sixty-six pregnancies (36.5%) resulted from 181 transfers in ZIFT, and 65 (31.4%) from 207 pregnancies in IVF-ET; the difference was not significant (OR 1.23, 95% CI: 0.8, 1.89). However, a chi-square test showed heterogeneity (as above). After removing one of the studies that used cryopreserved embryos, no heterogeneity was observed, but the results remained non significant.
Ectopic pregnancy.
Two (3%) of the 66 pregnancies in ZIFT were ectopic, compared with one (1.5%) of the 65 pregnancies in IVF-ET. While there was a trend towards a two-fold greater chance of having an ectopic pregnancy in ZIFT than in IVF-ET, there was no significant difference between the methods (OR 2.05, 95% CI: 0.21, 20.22).
The rates of pregnancy per retrieval, miscarriage, multiple, and ongoing pregnancy were comparable between tubal and intra-uterine transfers.