Statistical analyses
Participants were categorised according to quintiles (Q1 was the most pro-inflammatory group, whereas Q5 was the most anti-inflammatory group) and further stratified into two groups based on the presence or absence of endometriosis. Maternal characteristics were summarised according to the presence of endometriosis. T-tests were conducted to compare continuous variables, and Chi-square tests or Fisher’s exact tests were used to compare categorical variables, as appropriate. The extended Mantel-Haenzel Chi-square test for linear trend was used to analyse the trend in proportion. Adjusted odds ratios (aOR) and 95% confidence intervals (CI) for each obstetric outcome were calculated using a multiple logistic regression model, accounting for maternal age, maternal smoking status, maternal education, household income, maternal prepregnancy BMI, myoma uteri, adenomyosis, and ART pregnancy. To examine the interaction between endometriosis and infertility therapy, endometriosis women were further grouped into pregnancy with ART or pregnancy without ART. Then, the aOR and 95% CI for the same obstetric outcome were calculated by multiple logistic regression model. We accomplished this by using dummy variables for categorical variables composed of more than three categories. SPSS version 24 (IBM Corp., Armonk, NY) was used for the statistical analyses. A p value <0·05 indicated statistical significance.