Prediction of Subclinical Chorioamnionitis After Cervical Cerclage

Objective: To explore non-invasive indices for predicting subclinical chorioamnionitis following cervical cerclage. Methods: We performed a retrospective analysis of 80 singleton pregnant women who underwent cervical cerclage surgery in our hospital. Eighty patients were divided into either a histological chorioamnionitis group (n=57) and non-histological chorioamnionitis group (n=23). Gestational age before cervical cerclage, cervical dilation size, vaginal microbiota, cervical microbial colonization, and inammatory indicators related to peripheral blood cells (white blood cell count, neutrophil count, lymphocyte count, platelet count, CRP, neutrophil/lymphocyte ratio, platelet/white blood cell ratio, etc.) at pre-surgery time and onset of labor post-surgery were compared, and an independent sample t-test and multivariate logistic regression analysis were performed to study the risk factors associated with histological chorioamnionitis. Histological chorioamnionitis was used as the outcome variable, and receiver operating characteristic (ROC) curve analysis was performed to determine the cut-off value and evaluate the predictive value of these indicators for chorioamnionitis. Results: white blood cells and platelet/white blood cell during onset of labor) obtained from multi-factor analysis were grouped according to the cut-off value. A joint screening system was established that specied that two or more indicators are positive as joint screening, with a screening sensitivity of 87.5%, and specicity of 45.5%. Conclusions: After cervical cerclage, the number of white blood cells and the ratio of platelet/white blood cell during onset of labor combined with preoperative number of platelets have predictive value for potential histological chorioamnionitis in pregnant women with cervical cerclage.


Introduction
Cervical insu ciency is typically manifested clinically as painless cervical dilation with or without amnion sac. Cervical cerclage reinforces the cervix by application of special sutures at the utero-vesical pouch to achieve a supporting effect, protecting fetal membranes from pathogens, and effectively extending gestation. Existing evidence indicates that pregnant women with cervical cerclage have subclinical chorioamnionitis, and even infection in the amniotic cavity. However, such women lack the typical manifestations of clinical chorioamnionitis, such as fever (body temperature reaching or exceeding 38 °C) or accompanied by maternal tachycardia (heart rate exceeding 100 beats per minute), fetal tachycardia (more than 160 beats per minute), uterine tenderness, odor of amniotic uid, increased maternal white blood cells, increased C-reactive protein, and increased procalcitonin. Therefore, di culties in diagnosing subclinical chorioamnionitis increases the risk of miscarriage and causes adverse outcomes for mothers and children.
When pregnant women with cervical cerclage exhibit signs of possible premature delivery, understanding how to predict potential chorioamnionitis highly impacts the clinical anti-infection therapies, the appropriate time of termination of pregnancy and prognosis of mothers and children. At present, there is no ideal clinical method that can predict and monitor potential chorioamnionitis. The blood cell count and its derivative parameters such as platelet/white blood cell ratio have been widely recognized as markers of in ammatory diseases. This study uses retrospective analysis of cases with cervical cerclage to explore whether relevant indicators of peripheral blood cell analysis have an improved predictive role on potential chorioamnionitis during onset of labor.

Study population
This retrospective study included pregnant women who received transvaginal cervical cerclage in our hospital from 2018 to 2019 and gave birth between 20 weeks, 0 days and 36 weeks, 6 days of gestation.
We received written informed consent from all included patients. The Ethical Review Committee of Fujian Maternity and Child Health Hospital approved the collection and use of samples and clinical information for this study (No. 2019014).

Inclusion criteria:
Patients were required to exhibit the following indications for cervical cerclage, including: 1) singleton pregnancy, 2) cervical length less than or equal to 25 mm during transvaginal ultrasound scan (with or without funnel-like dilatation of the internal cervix), 3) have a history of one or more incidences of miscarriage in the second trimester of pregnancy or premature delivery, 4) the cervix is dilated during physical examination, the amniotic sac protrudes into the vagina, or the fetal membrane is visible through the cervix.
Exclusion criteria: The following conditions were used in screening for exclusion in the study: 1) multiple pregnancy, 2) multiple cerclages in the current pregnancy, 3) full-term delivery after cervical cerclage, and 4) fetal malformations.
Cervical length and/or funnelling were examined by vaginal ultrasound. The administration of prophylactic antibiotics and spinal anesthesia were administered before cerclage. The type of cerclage performed was the McDonald method, using No. 2 Ethibond (Ethicon, UK) to suture four ports at the cervico-vaginal junction with purse-string sutures for cerclage.
Clinical indicators included: 1) General information: maternal age, number of gestations, parity, number of fetal losses in the previous second trimester, number of miscarriages, preoperative cervical dilatation, gestational week of cervical cerclage, gestational week of fetal membrane rupture after cervical cerclage, gestational week of cerclage removal, gestational week of childbirth, and peripheral blood cell analysis results including platelet/white blood cell ratio, neutral granulocyte/lymphocyte ratio, preoperative Creactive protein (CRP) results at pre-surgery and onset of premature labor (no patients were treated with corticosteroids or antibiotics when taking a peripheral blood sample). 2) Maternal and child outcomes: delivery methods, delivery complications including premature delivery, miscarriage, premature rupture of membranes, placental abruption, sepsis, etc., use of uterine contraction inhibitors and drugs accelerating lung maturation, and placental pathology results. Newborn birth weight, Apgar score, the number of days the newborn stays in the ICU, and the complications related to premature delivery including necrotizing enterocolitis, intraventricular hemorrhage, and retinopathy were also outcomes documented.
After delivery, tissue samples were obtained from the placenta and placental membranes, and the tissue samples were embedded in para n and sectioned into thin slices for diagnosis of chorioamnionitis.

Laboratory inspection index:
According to observed pathology of the placenta following delivery, 80 patients were divided into histological chorioamnionitis (n = 57) and non-histological chorioamnionitis groups (n = 23 cases). The in ammatory indicators related to peripheral blood cells (leukocytes, neutrophil counts, lymphocyte counts, platelet counts, C-reactive protein, neutrophil/lymphocyte ratio, platelet/white blood cell ratio etc.) were compared between two groups of pregnant women with cervical cerclage at pre-operation and during onset of labor post-operation. To study the risk factors related to histological chorioamnionitis, receiver operating characteristic curve analysis was performed, we calculated the area under the curve (AUC), determined the cut-off value, and evaluated its predictive value for chorioamnionitis.

Statistical analysis:
All statistical analyses were performed using SPSS version 24.0 (IBM, Armonk, NY, United States), An independent sample t-test was used to compare differences, and logistic regression was used to evaluate risk factors related to histological chorioamnionitis. A receiver operating characteristic curve (ROC) curve was graphed to evaluate the optimal cut-off value of risk factors related to histological chorioamnionitis.
The results are expressed in odds ratio (OR) and 95% con dence interval (CI). A P value < 0.05 was considered statistically signi cant Results: 1. A total of 80 women underwent cervical cerclage (including 36 rescue cervical cerclage, 31 therapeutic cerclage, and 13 prophylactic cerclage). There were 57 women with histological chorioamnionitis and 23 with non-histological chorioamnionitis. Tables 1 and 2 provide general information concerning the pregnant women and neonatal outcomes. The pregnant women in the study had an average age of 31.61 ± 4.66 years, the average gestational age at cerclage was 22.04 ± 3.79 weeks, the average gestational age at cervical cerclage removal was 28.61 ± 5.24 weeks, and the average gestational age at delivery was 28.98 ± 4.79 weeks. The average prolonged gestational age was 47.76 ± 32.56 days, and 53 pregnant women (66.25%) were treated for accelerating fetal pulmonary maturation. The average gestational age of preterm premature rupture of fetal membrane after cervical cerclage was 30.61 ± 4.22 weeks, and the average incubation period of premature rupture of fetal membrane to delivery was 5.45 days. Among these patients, 2 women (2.6%) developed sepsis, and 21 (26.3%) exhibited placental abruption.  1. 2. Regardless of whether the women were diagnosed with histological chorioamnionitis, the cases were grouped. The routine blood analysis distribution between the two groups was analyzed and the results are shown in Table 3. The platelet counts pre-surgery in patents with histological chorioamnionitis, total number of white blood cells and the total number of neutrophils and platelets during onset of labor were higher than those with non-histological chorioamnionitis patients, and the ratio of platelet/white blood cell during onset of labor was lower than those with non-histological chorioamnionitis. The difference was statistically signi cant.
2. 3. In groups based on diagnosis with histological chorioamnionitis, univariate analysis of meaningful variables with single factor, receiver operating characteristic curve (ROC) analysis, and cutoff value with the maximum value of Youden index determination results are shown in Table 4 and Fig. 1. The area under the curve of the number of platelets before the operation, the total number of white blood cells, neutrophils, platelets, and platelet/white blood cell ratio during onset of labor were statistically signi cantly increased in women with histologic chorioamnionitis. The cut-off values were 229.5 × 109/L; 10.245 × 109/L; 9.86 × 109/L; 240.5 × 109/L; 1.022 respectively. 1. 4. According to the cut-off value, the number of platelets prior to surgery, the total number of leukocytes, neutrophils, and platelet/white blood cell ratio during onset of labor, cerclage gestational age, cervical dilatation, vaginal microbiota, and cervical microbial settings are independent variables. Chorioamnionitis is the dependent variable, and a univariate logistic regression was determined. We found that preoperative platelets, the number of white blood cells, the number of neutrophils, and the number of platelets and the ratio of platelet/white blood cell during onset of labor are risk factors for histological chorioamnionitis. The risk of histological chorioamnionitis for patients with preoperative platelets ≥ 229.5 × 109/L, leukocytes ≥ 10.245 × 109/L, neutrophils ≥ 9.86 × 109/L, platelet counts ≥ 240.5 × 109/L, platelet to white blood cell ratio < 1.02 during onset of labor, is 7.692, 6.750, 4.800, 9.333, 6.529 times as high (respectively) as those with preoperative platelets < 229.5 × 109/L, leukocytes < 10.245 × 109/L, neutrophils < 9.86 × 109/L, platelet counts < 240.5 × 109/L, and the ratio of platelet/white blood cells ≥ 1.022 during onset of labor. The regression method (P = 0.15) for single-factor meaningful variables was used to t the multi-factor logistic regression. Preoperative platelets, the number of white blood cells and platelet to white blood cell ratio during onset of labor were signi cantly elevated in individuals with histological chorioamnionitis (Table 5). 2. 5. The predicted probability was calculated by multi-factor analysis which was analyzed again by receiver operating curve (Table). The area under the curve was 0.795 (P = 0.000), and its predictive effect is greater than that of single routine blood analysis. 3. 6. The sensitivity and speci city of combined screening Following multivariate analysis, the statistically signi cant variables (preoperative platelets, white blood cells and platelet to white blood cell ratio during onstet of labor) were grouped according to cut-off value.
Then, a joint screening system was established to specify that 2 or more indicators were positive as a joint screening is positive, with a sensitivity of 87.5%, and a speci city of 45.5%.

Discussion:
Chorioamnionitis is de ned as in ammation or infection of the placenta and placental membranes after 20 weeks of gestation. It is one of the primary causes of perinatal morbidity and mortality, which seriously negatively affects the prognosis of the perinatal infant. According to the different stages of its clinical manifestations, it is divided into histological chorioamnionitis and clinical chorioamnionitis. Histological chorioamnionitis, also known as subclinical chorioamnionitis, has hidden clinical manifestations and requires con rmation by a thorough pathological examination of the placenta. Clinical chorioamnionitis is the nal stage of histological chorioamnionitis progression. At this stage, the fetus in utero has been exposed to the in ammatory environment for an extended time period, which greatly impacts the prognosis of the newborn.
Prior studies have reported the different incidence of chorioamnionitis in pregnant women with cervical cerclage. Lee et al. reported that 81% (42/52) of patients with cervical insu ciency have evidence of intraamniotic in ammation [1,2]. In the present study, the diagnosis rate of clinical chorioamnionitis is only 23.8% and tissue chorioamnionitis incidence is 70%. For pregnant women with cervical cerclage, it often indicates potential chorioamnionitis when signs of labor appear. The removal or retention of cervical cerclage suture and the time for expected continued treatment are clinically controversial issues. Clinical decision-making is necessary to balance the potential bene ts of prolonged pregnancy with risk of intrauterine infection. Amniocentesis is an invasive test, and if maternal serum markers can be obtained non-invasively to predict tissue chorioamnionitis, management after cervical cerclage could be improved.
At present, many studies describe predictive indicators of tissue chorioamnionitis, and the results of clinical studies are also quite different. There are currently no recommended clinical or laboratory indicators. In particular, the predictive value of amniotic uid in ammatory factors is highly debatable. Yoneda et al. reported that the IL-8 level in amniotic uid can more accurately predict tissue chorioamnionitis [3]. Park et al. reported that maternal plasma IL-6 can independently predict intraamniotic infection in preterm women. However, its diagnostic value is inferior to assessing IL-6 levels in amniotic uid and similar to that of serum CRP [4]. Horinouchi et al. found that the IL-6 and PCT levels in umbilical vein can predict histological chorioamnionitis [5]. Another prospective study found no connection between in ammatory markers in amniotic uid and preterm birth [6]. Kim et al. found that for patients with preterm premature rupture of fetal membranes, non-invasive parameters (serum CRP) and invasive parameters (amniotic uid IL-6 levels) were not signi cantly different in predictive ability for histological chorioamnionitis [7]. A recent systematic review and meta-analysis of diagnostic indicators for histological chorioamnionitis indicated that simply using CRP and maternal leukocytosis as predictive indicators shows low sensitivity and speci city. In a combination of 13 studies, CRP sensitivity was 68.7% (95% CI 58-77%) and speci city was 77.1% (95% CI 67-84%). A combination of 4 studies to evaluate the quantity of maternal white blood cells revealed that the combinatorial sensitivity was 51% (95% CI 40-62%) and speci city was 65% (95% CI 50-78%) [8]. In light of these ndings, current recommendations support a combination of maternal blood and amniotic uid biomarkers for improved accuracy in predicting histological chorioamnionitis [9].
Recent data suggest complete blood count (CBC) and its derived parameters including neutrophil to lymphocyte ratio (NLR), platelet to lymphocyte ratio (PLR), and platelet to white blood cell ratio are recognized in ammatory markers for low-grade in ammation [10,11]. These non-invasive in ammatory markers are highly correlated with neonatal outcomes, suggesting a poor neonatal prognosis [12]. Our research on the prediction of preterm birth found that the combination of three parameters of blood cell components such as NLR, PDW (platelet distribution width) and HGB (hemoglobin) can better predict preterm birth, and the predictive sensitivity of the combined diagnostic markers is 88.6% and the speci city is 40.5% [13,14]. Another of our studies on the prediction of preterm birth found that platelets and PLT/WBC as a potential marker have certain signi cance in predicting the occurrence of histological chorioamnionitis [15]. Using blood cell-related parameters to predict preterm labor requires no concomitant drug therapy to promote lung maturity. However, Winkler et al. suggest that the dose of corticosteroids used to prevent respiratory distress syndrome does not affect early prediction of asymptomatic infections in preterm labor [16].
The ndings presented here indicate that the number of platelets in patients with cervical cerclage before surgery is a risk factor for chorioamnionitis. The post-surgery risk of histological chorioamnionitis for pregnant women with cervical cerclage whose preoperative platelets ≥ 229.5 × 109/L was 7.692 times greater than those with preoperative platelets < 229.5 × 109/L. Multivariate regression analysis also indicated that preoperative platelet count is a potential serum marker of histological chorioamnionitis.
Multiple studies have con rmed that platelet activation participates in the pre-in ammatory response and can be used as a predictor of in ammation. In the course of in ammation, platelets rapidly participate in in ammatory pathogenesis through the secretion of cytokines, chemokines and other in ammatory mediators [17]. Our team's prospective study on spontaneous preterm birth also found that the platelet count of histological chorioamnionitis (HCA) patients was signi cantly higher than that of the non-HCA and control groups (P < 0.001). In addition, the AUC of the PLT count was 0.8095, indicating that the PLT count is a sensitive predictor of HCA in preterm patients.
The platelet/white blood cell ratio (PLT/WBC) as a hematological marker of systemic in ammation has been widely used to predict postoperative infection and disease prognosis [18,19]. We have found that PLT/WBC are sensitive biomarkers for the diagnosing HCA. Our data also attempts to verify that the PLT/WBC ratio can be used as an index to predict histological chorioamnionitis during onset of labor following cervical cerclage. The data in this study also con rms that the number of white blood cells and platelet/white blood cell ratio during onset of labor are risk factors for subclinical chorioamnionitis are also likely risk factors after multivariate logistic regression analysis. After further grouping the platelets before cervical cerclage, the number of white blood cells and the platelet/white blood cell ratio during onset of labor according to the cut-off values, a joint screening system was established. When these three indicators have 2 or more values greater than or equal to the corresponding cut-off value, the sensitivity for predicting subclinical chorioamnionitis is 87.5%, and the speci city is 45.5%.
Park et al. reported that non-invasive parameters, including maternal white blood cell count, CRP level, parity and gestational age, can strongly predict intraamniotic infection in women with preterm premature rupture of membranes (PPROM) [20]. The present study attempted to nd appropriate clinical and laboratory indicators to predict chorioamnionitis. However, no clinical indicators such as parity, cervical cerclage gestational age, and cervical dilatation were found to be risk factors for chorioamnionitis. This difference may be due to the different sample sources in different studies.
Our research has certain limitations that must be mentioned. One limitation is that this study is retrospective and was conducted in a single center. The small sample size limits the widespread application of our results. Secondly, we will further verify the effectiveness of the noninvasive prediction model through prospective studies in future clinical studies.

Declarations
Author contributions: Li Li, Mian Pan and Jianying Yan designed the study, collected the data, analyzed the data and wrote the manuscript; Xinxin Huang, Liyin Qiu, Mei Ma and Jun Zhang collected the data. All authors have read and approved the manuscript. Receiver operating characteristic curve of histological chorioamnionitis prediction probability