Association between adenomyosis and placenta accreta and mediation effect of assisted reproductive technology on the association: A nationwide observational study

Abstract Objectives To investigate the association between adenomyosis and placenta accreta spectrum (PAS) and to evaluate the effect of assisted reproductive technology (ART) in mediating this association. Methods We retrieved data for singleton women from the Japanese nationwide perinatal registry between 2013 and 2019, excluding women with a history of adenomyomectomy. To investigate the association between adenomyosis and PAS among women, we used a multivariable logistic regression model with multiple imputation for missing data. We evaluated mediation effect of ART including in vitro fertilization and intracytoplasmic sperm injection on the association between adenomyosis and PAS using causal mediation analysis based on the counterfactual approach. Results Of 1 500 173 pregnant women, 1539 (0.10%) had adenomyosis. The number receiving ART was 489/1539 (31.8%) and 117 482/1 498 634 (7.8%) in women with and without adenomyosis, respectively. The proportion of women who developed PAS was 21/1539 (1.4%) in women with adenomyosis and 7530/1 498 634 (0.5%) in women without adenomyosis. Adenomyosis was significantly associated with PAS (odds ratio [OR] 1.95; 95% confidence interval [CI] 1.26–3.00; P = 0.002). Mediation analysis showed that OR of the total effect of adenomyosis on PAS was 1.98 (95% CI 1.13–3.04), OR of natural indirect effect (effect explained by ART) was 1.15 (95% CI 1.01–1.41), and OR of natural direct effect (effect unexplained by ART) was 1.72 (95% CI 0.86–2.82). The proportion mediated (natural indirect effect/total effect) was 26.5%. Adenomyosis was also significantly associated with PAS without previa (OR 1.96; 95% CI 1.23–3.13, P = 0.005). Conclusion Adenomyosis was significantly associated with PAS. ART mediated 26.5% of the association between adenomyosis and PAS.