Authors’ Contribution:
Yu Han:Statistical analysis and manuscript writing
Wang Na:Data collection
Wang Chao:Data Collection
Hailan Yang:Project development and final manuscript editing
Analysis of risk factors for cervical insufficiency: a retrospective controlled study
Introduction
Cervical insufficiency (CI) is one of the main obstetrical diseases leading to preterm birth, which also afflicts 4% of patients with recurrent miscarriage.1CI refers to the abnormal remodeling, premature softening and shortening of the cervix, which is one of the major obstetric disorders leading to preterm birth and recurrent miscarriage in the second trimester, with an incidence of 1%-2% that is increasing.2 There are known risk factors for CI, such as acquired factors including dilatation and evacuation, dilatation and curettage, induced abortion, previous cervical laceration, LEEP surgery, cold-knife cone biopsy, or other cervical surgeries. Congenital factors include utero diethylstilbestrol exposure, collagen vascular disease, or Müllerian abnormalities.3
The Canadian Society of Obstetrics and Gynecology (SOGC) clinical guidelines state that polycystic ovary syndrome (PCOS) is an independent risk factor for CI.4 In addition, it has been shown that higher BMI is an independent risk factor for CI;5however, it has also been shown that obese patients have a lower risk of CI and premature delivery.6 Therefore, whether BMI is significantly different between CI patients and non-CI patients requires further elucidation. The Royal College of Obstetricians and Gynecologists (RCOG) clinical guidelines also consider multiple pregnancies a risk factor for CI.7 However, whether a previous history of multiple pregnancies is associated with the development of CI has not been shown. Additionally, previous studies have proposed that assisted reproductive technology (ART) is associated with the occurrence of CI.8 In recent years, though studies have gradually found that the incidence of CI in the ART population is high,9 whether ART serves as a risk factor for CI has not been clarified. Studies have shown that PCOS, impaired glucose tolerance and type II diabetes, and elevated BMI together constitute 70% of patients with recurrent miscarriage.1 At the same time, in our previous study, it was found that more than 40% of CI patients had gestational diabetes mellitus (GDM) and pregnancy with diabetes mellitus (PGDM).10 However, no relevant studies on whether the incidence of GDM and PGDM is different in CI patients versus non-CI patients, as well as whether diabetes can be used as a risk factor to predict the incidence of CI, currently exist.
In this paper, a single-center retrospective cohort study is conducted pertaining to the unspecified risk factors of CI by performing comparisons between the CI population and non-CI population, which may offer significant insight into the comprehensive management of CI patients during pregnancy and may clarify further research directions.
Methods
A total of 209 patients with a singleton pregnancy complicated by CI (CI group) hospitalized in the First Hospital of Shanxi Medical University from July 2013 to June 2020 were enrolled in this study. The control group (non-CI group) was randomly selected at a ratio of 1:2 and the inclusion criteria were: singleton pregnancy and non-cervical insufficiency. All patients were required to have records of the whole pregnancy. A total of 348 patients in the control group who met the criteria were collected. The general conditions, pregnancy complications (subclinical hypothyroidism, GDM and PGDM, PCOS, uterine malformation, uterine fibroids/adenomyoma, hypertensive disorders in pregnancy, anemia, vaginitis, and hyperlipidemia) of the two groups were then collected. A model of logistic regression was used to calculate the prediction probability, produce a combination of multiple indicators of new variables, draw ROC curves, and assess the predictive ability of risk factors for CI.
Statistical tools, such as Excel and SPSS 20, were used for data entry and analysis, while the quantitative data were analyzed by the t-test or analysis of variance and rank sum test. Qualitative data were analyzed using the chi-square test to explore the relationship between the variables, while logistic regression and ROC curves were used for multivariate analysis. α=0.05 was set as the test level.
Results
Comparison in the general conditions between women with CI and those without CI
As shown in Table 1 , there was no statistical difference in age (30.66 ± 3.66 VS 30.28 ± 4.56, P=0.284) and height (1.61 ± 0.05 VS 1.62 ± 0.06, P=0.058) between CI group and non-CI group. Compared with the non-CI group, patients in the CI group had a greater mean body weight (62.49±9.90 VS 58.83±9.17kg, P<0.001), a greater mean BMI (24.20±3.57 VS 22.63±3.63 kg/m2, P<0.001), and a greater proportion had a BMI of 24 kg/m2 or higher (50.25%,102/203 VS 29.85%,103/345, P<0.001). Compared with the non-CI group, the CI group had less mean weight gain during pregnancy (10.66±6.11 VS 13.36±5.35kg, P<0.001), and there was no significant difference in the mean BMI at delivery (27.96±3.76 VS 27.77±4.34kg/m2, P=0.602) and in the proportion of patients with BMI greater than 24 kg/m2 at delivery between the two groups (88.41%,183/207 VS 84.30%,290/344, P=0.051).
There was no significant difference in the gravidity(2.92±1.20 VS 2.14±1.18, P=1.903), uterine cavity procedures(0.45±0.87 VS 0.50±0.84, P=0.556), or medical abortions (0.09±0.36 VS 0.05±0.26, P=0.154)between patients in the CI and non-CI groups. Patients in the CI group had significantly fewer deliveries(0.26±0.54 VS 0.48±0.59, P<0.001)and more premature deliveries(0.16±0.40 VS 0.01±0.12, P<0.001)and spontaneous abortions (1.19±0.93 VS 0.10±0.39, P<0.001)than those in the non-CI group. The proportion of patients in the CI group who had a history of previous multiple pregnancies was significantly more than that in the non-CI group (7.66%,16/209 VS 0.29%,1/348, P<0.001), while the proportion of patients who conceived by ovulation induction/IVF-ET was also significantly higher in the CI group than in the non-CI group (23.44%,49/209 VS 4.89%,17/348, P<0.001).
2. Comparison in the comorbidities between women with CI and those without CI
As seen in Table 2 , in the CI group, the incidence of GDM/PGDM was found to be significantly higher than that in non-CI group (41.21%,82/199 VS 18.39%,64/348, P<0.001), while patients with PCOS accounted for a higher proportion in the CI group (10.55%,21/199 VS 0.86%,3/348, P<0.001). There were no significant differences in the incidence of subclinical hypothyroidism, uterine malformations, uterine fibroids/adenomyomas, hypertensive disorders in pregnancy, anemia, vaginitis, and hyperlipidemia between the CI group and non-CI group. According to Figure 1 , the distribution of various complications in the two groups of patients can be more clearly understood (the results marked with ★★★ are statistically significant).