The perinatal outcomes of women treated for Asherman syndrome: a propensity score-matched cohort study

Abstract

Do the perinatal outcomes of patients following hysteroscopic treatment for Asherman syndrome (AS) differ from that of a control population? Perinatal complications including placental issues, high blood loss, and prematurity in women after treatment for AS should be considered as moderate to high risk, especially in patients who have undergone more than one hysteroscopy (HS) or repeated postpartum instrumental revisions of the uterine cavity (Dilation and Curettage; D&C). The detrimental impact of AS on obstetrics outcomes is commonly recognized. However, prospective studies evaluating perinatal/neonatal outcomes in women with AS history are sparse, and the characteristics accounting for the respective morbidity of AS patients remain to be elucidated. We conducted a prospective cohort study utilizing data from patients who underwent HS treatment for moderate to severe AS in a single tertiary University-affiliated hospital (enrolled between 01 January 2009 and March 2021), and who consequently conceived and progressed to at least 22nd gestational week of pregnancy. Perinatal outcomes were compared to a control population without an AS history, retrospectively enrolled concomitantly at the time of delivery for each patient with AS. Maternal and neonatal morbidity was assessed as well as the characteristics-related risk factors of AS patients. Our analytic cohort included a total of 198 patients, 66 prospectively enrolled patients with moderate to severe AS and 132 controls. We used multivariable logistic regression to calculate a propensity score to match 1–1 women with and without AS history based on demographic and clinical factors. After matching, 60 pairs of patients were analysed. Chi-square test was used to compare perinatal outcomes between the pairs. Spearman’s correlation analysis was utilized to investigate the correlation between perinatal/neonatal morbidity and the characteristics-related factors of AS patients. The odds ratio (OR) for the associations was calculated by logistic regression. Among the 60 propensity matched pairs, the AS group more frequently experienced overall perinatal morbidity, including abnormally invasive placenta (41.7% vs 0%; P < 0.001), retained placenta requiring manual or surgical removal (46.7% vs 6.7%; P < 0.001), and peripartum haemorrhage occurrence (31.7% vs 3.3%; P < 0.001). Premature delivery (P < 0.001). However, no increased frequency of intra-uterine growth restriction or worsened neonatal outcomes were observed in AS group. Univariable analysis of risk factors for AS group morbidity outcomes revealed that the main factor related to abnormally invasive placenta was two or more HS procedures (OR 11.0; 95% CI: 1.33–91.23), followed by two or more D&Cs preceding AS treatment (OR 5.11; 95% CI: 1.69–15.45), and D&C performed postpartum as compared to post abortion (OR 3.0; 95% CI: 1.03–8.71). Similarly, two or more HS procedures were observed as the most important factor for retained placenta (OR 13.75; 95% CI: 1.66–114.14), followed by two or more preceding D&Cs (OR 5.16; 95% CI: 1.67–15.9). Premature birth was significantly associated with the number of preceding D&Cs (OR for two or more, 4.29; 95% CI: 1.12–14.91). Although the cohort of patients with AS was enrolled prospectively, a baseline imbalance was intrinsically involved in the retrospective enrolment of the control group. However, to reduce the risk of bias, confounding factors were adjusted for using propensity score matching. The limitation to the generalization of our reported results is the single institution design in which all patients were treated for AS in one tertiary medical centre. Within our search scope, our study represents one of the first and largest prospective studies of perinatal and neonatal outcomes in moderate to severe AS patients with a prospectively analysis of the risks factors of characteristics significantly influencing reported morbidities among patients with AS. The study was supported by the Charles University in Prague [UNCE 204065] and by the institutional grant of The General Faculty Hospital in Prague [00064165]. No competing interests were declared. N/A.