Uterine adenomyosis and in vitro fertilization outcome: a systematic review and meta-analysis

Is adenomyosis associated with IVF/ICSI outcome in terms of clinical pregnancy rate? In a meta-analysis of published data, women with adenomyosis had a 28 reduction in the likelihood of clinical pregnancy at IVF/ICSI compared with women without adenomyosis. Estimates of the effect of adenomyosis on IVF/ICSI outcome are inconsistent. A systematic literature review and meta-analysis were conducted. A Medline search was performed to identify all the comparative studies published from January 1998 to June 2013 in the English language literature on IVF/ICSI outcome in women with and without adenomyosis. Two authors independently performed the literature screening, scrutinized articles of potential interest, selected relevant studies and extracted data. Studies were categorized based on research design. Of the 17 articles assessed in detail, 9 were finally selected based on diagnosis of adenomyosis at magnetic resonance imaging or transvaginal ultrasonography. The quality of studies was evaluated by means of the Newcastle-Ottawa scale. A total of 1865 women were enrolled in the 9 selected studies, 665 of whom in 4 prospective observational studies, and 1200 in 5 retrospective studies. The dichotomous data for clinical pregnancy and secondary outcomes were expressed as risk ratios (RR) with 95 confidence intervals (CIs) and were combined in a meta-analysis using the random-effects model. The heterogeneity Cochranes Q and the I-2 statistics were calculated. Eggers approach to testing the significance of funnel plot asymmetry was also used. The clinical pregnancy rate achieved after IVF/ICSI was 123/304 (40.5) women with adenomyosis versus 628/1262 (49.8) in those without adenomyosis. The RR of clinical pregnancy ranged from 0.37 (95 CI, 0.150.92) to 1.20 (95 CI, 0.582.45), with a significant heterogeneity among studies (I-2 56.8, P 0.023). Pooling of the results yielded a common RR of 0.72 (95 CI, 0.550.95). A funnel plot showed no indication of asymmetry among studies (Eggers test, P 0.696). In a meta-regression model, no association was observed between prevalence of endometriosis and the likelihood of clinical pregnancy. Three studies reported the pregnancy rate per cycle. The common RR was 0.71 (95 CI, 0.510.98; I-2 78.1, P 0.010). The RR observed in a study with donated oocytes was 0.90 (95 CI, 0.751.08). The number of miscarriages per clinical pregnancy was reported in seven studies. A miscarriage was observed in 77/241 women with adenomyosis (31.9) and in 97/687 in those without adenomyosis (14.1). The RR of miscarriage ranged from 0.57 (95 CI, 0.152.17) to 18.00 (95 CI, 4.0879.47) (I-2 67.7, P 0.005). Pooling of the results yielded a common RR of 2.12 (95 CI, 1.203.75). Qualitative and quantitative heterogeneity among studies was high. At sensitivity analysis, I-2 statistic regarding the main outcome was reduced under the 50 threshold removing one trial, but the resulting confidence interval crossed unity. Also the confidence interval of the common RR of the four studies reporting only one IVF/ICSI cycle included unity. Only part of the studies could be included in the assessment of secondary outcomes. Adenomyosis appears to impact negatively on IVF/ICSI outcome owing to reduced likelihood of clinical pregnancy and implantation, and increased risk of early pregnancy loss. Screening for adenomyosis before embarking on medically assisted reproductive procedures should be encouraged. The potentially protective role of long down-regulation protocols needs further evaluation. In future studies on the association between adenomyosis and IVF/ICSI outcome, a matched casecontrol design should be adopted, live birth should be the default primary outcome and only the results regarding the first cycle should be considered. None.