High Level of Complement Factor Ba Within 11 to 17 Weeks of Gestation Increases the Risk of Subsequent Gestational Diabetes: A Propensity Score-Matched Study

Background: Several studies have shown that the over activation of complement factor B(CFB) was related to obesity, insulin resistance(IR) and type 2 diabetes mellitus. This study was to assess whether circulating complement factor Ba (CFBa) within 11 to 17 weeks of gestation is associated with subsequent gestational diabetes mellitus (GDM) or not. Methods: Biochemical parameters and blood samples were collected from 399 pregnant women within 11 to 17 weeks of gestation. At 24 to 28 weeks of pregnancy, all participants underwent 75-g oral glucose tolerance test (fasting for more than 8 hours before blood sampling) and were assigned to GDM group(n=80) and normal control group(n=319). Perinatal data were collected after delivery. A propensity score-matched (PSM) analysis was performed to reduce the impact of confounding factors on glucose metabolism during pregnancy between the two groups. Results: Two groups of 74 well-matched patients who maintained balance in terms of baseline characteristics. The levels of CFBa in pregnant women who later developed GDM were signicantly higher than those in healthy pregnant women [0.4(0.1-0.8) vs. 0.2(0.2-0.3), P=0.031]. Logistic regression results conrmed that the level of CFBa was an independent inuencing factor for the occurrence of GDM (OR=1.52, 95% CI: 1.25-1.85, P=0.000). Further grouping according to the quartile of CFBa level, it was found that the incidence of GDM in category 3 was markedly higher than that in the rst and the second categories. Conclusions: High level of the CFBa within 11 to 17 weeks of gestation increased the risk of subsequent GDM, and maybe a biomarker for predicting GDM.


Background
Gestational diabetes mellitus (GDM) refers to the abnormal glucose metabolism found for the rst time during pregnancy without reaching the level of dominant diabetes [1]. The incidence of GDM uctuates widely due to differences in geographical area, ethnicity and diagnostic criteria [2,3]. Long-term uncontrolled blood glucose in patients with GDM may have serious adverse effects on the mothers and fetus, such as preeclampsia, cesarean section, macrosomia, premature rupture of membrane, neonatal hypoglycemia and so on [4,5]. At present, the exact pathogenesis of GDM is still not throughly clear, but the IR and the chronic subclinical in ammatory process caused by immune system dysfunction are considered as the main factors for the development of GDM [6]. In recent years, the innate immune system was shown to be closed related to the metabolic pathway [7].
The complement system is an important part of innate immune system. It can protect host against pathogens and connect innate and adaptive immune response [8]. Abnormal regulation of complement system or de ciency of complement can increase the susceptibility of chronic in ammatory diseases and infections [9]. The complement cascades are activated in three different ways: the classical pathway, the alternative pathway and the mannose binding lectin way. Among them, the alternative pathway has attracted much more attention because of its potential role in metabolic diseases such as hypertension, type 2 diabetes mellitus, as well as dyslipidemia, and CFB is critical for its activation [10][11][12]. CFB is a single chain glycoprotein, which consists of ve protein domains. After the alternative pathway of complement system was activated, factor B was cleaved into two fragments, Ba and Bb. The molecular weight of Ba fragment is about 30000 dalton, which is composed of 234 amino acids and comes from the N-terminal part of the precursor molecule of factor B [13,14]. At present, many studies have con rmed that excessive activation of CFB was closely related to obesity, IR and type 2 diabetes mellitus [15,16]. but its role in GDM has not been elucidated. The main purpose of this study was to assess the effect of circulating CHBa (the activation product of CHB) level within 11 to 17 weeks of gestation on the risk of subsequent GDM.

Participants
This is a nested case-control study of pregnant women. From February 2017 to April 2019, a total of 399 pregnant women were enrolled in the department of Obstetrics and Gynecology of Jinshan District Central Hospital of Shanghai Jiao-Tong University A liated 6th People's Hospital. At 24 to 28 weeks of pregnancy, all pregnant women underwent 75-g oral glucose tolerance test (fasting for more than 8 hours before blood sampling). According to criteria of the International Association of Diabetes and Pregnancy Study Groups (IADPSG) and the WHO [17,18], GDM can be diagnosed when the blood glucose meets or exceeds any of the following values: fasting blood glucose: 5.1 mmol/L, 1-hour blood glucose: 10.0 mmol/L, 2-hour blood glucose: 8.5 mmol/L. According to the above criteria, the pregnant women were divided into two groups including the GDM group (n = 80) and the normal control group (n = 319). The women who started their rst prenatal test in early pregnancy until delivery with complete data were included. Pregnant women who meet the following conditions should be excluded: (1)previous history of diabetes; (2)severe acute and chronic infections; (3)autoimmune diseases; (4)multiple pregnancy. The study was approved by the Ethics Committee of Jinshan District Central Hospital of Shanghai Jiao-Tong University A liated 6th People's Hospital. Written informed consent was signed by all participants.

Data and serum sample collection
All participants were required to complete a questionnaire to collect background information including family history of diabetes mellitus, previous disease history, current disease, reproductive history and drug treatment. During the examination, Height and weight were assessed in a standard form by the same physician. Body mass index (BMI) was calculated in the following formula: BMI = body weight(Kg)/Height(m) 2 . Perinatal data of 399 pregnant women were collected. Newborn with a birth weight of 4000 g or more is de ned as macrosomia. Premature rupture of membranes refers to natural rupture of membranes before labor. Blood samples were collected after fasting for more than 8 hours at night before the rst antenatal examination and the serum were store at -80℃.
Other indicators used to evaluate liver and kidney function, such as alanine aminotransferase(ALT), aspartate aminotransferase (AST), γ-glutamyltransferase(γ-GT), blood urea nitrogen (BUN), creatinine (Cr), uric acid (UA) and blood lipids including total cholesterol (TC), triglycerides (TG), low density lipoprotein cholesterol (LDL-C) were all measured by enzymatic method. Blood glucose was measured by glucose oxidase method. The frozen (-80℃) serum was quickly thawed in a 37℃water bath until just thawed, and then transferred to ice. Serum Ba concentrations were measured in duplicate using a MicroVue Ba Enzyme Immunoassay Kit (Quidel Corporation, USA, A033). The linear range of the standard was 0.033-3.239 ng/ml. The intra-assay and inter-assay coe cients of variation (CV) were 2.3% and 8.1% respectively. The control values were in the control range.

Statistical analysis
Continuous variables were expressed as mean ± standard deviation (SD) when they conform to normal distribution or as median (interquartile range [IQR] 25th percentile-75th percentile) when they do not. Categorical variables were presented as the number (percentage). Kolmogorov-Smirnov test was used to test whether continuous variables conform to normal distribution. Comparison between the two groups using the two-tailed Student's test or the Mann-Whitney U test depends on whether the continuous variables followed a normal distribution. Categorical variables were compared by chi-squared test or Fisher's exact test. A propensity score-matched (PSM) analysis was done to balance the distribution of baseline characteristics between the two groups. The included covariables were as follows: age, family history of diabetes, prepregnancy BMI, ALT, AST, γ-GT, BUN, Cr, UA, TC, TG, LDL, miscarriage and reproductive history. Pairs of GDM and the control groups were derived using 1:1 greedy nearest neighbor matching with a caliper of 0.05. This strategy resulted in 74 matched pairs in each group. To verify the impact of the level of CFBa on subsequent GDM, a logistic regression analysis performed. All statistical analyses were performed by SPSS 26.0 (SPSS Inc., Chicago, IL). A two-sided P-value < 0.05 was considered statistically signi cant.

The baseline clinical characteristics and biochemical indexes between two groups
The clinical characteristics of all pregnant women were shown in Table 1, including the GDM group (n = 80) and the control group (n = 319). There was no signi cant difference in AST, γ-GT, BUN, Cr between the two groups. The pregnant women in the GDM group had signi cantly higher age, incidence of family history of diabetes, prepregnancy BMI, HbA1c, ALT, UA, TC, TG, LDL, reproductive history, fasting blood glucose, 1-hour blood glucose, 2-hour blood glucose levels (all P < 0.05). After that, regular follow-up was carried out until delivery. The perinatal data of all pregnant women were shown in Table 2. Gestational age at delivery, the proportion of premature rupture of membranes, the rate of intrauterine distress, amount of postpartum hemorrhage, the weight of newborn, the incidence of caesarean section, apgar score showed no signi cant difference between the two groups. The incidence of macrosomia in GDM group was signi cantly higher than that in control group (11.25% vs. 4.39%, P = 0.037).   GDM gestational diabetes mellitus, BMI body mass index, HbA1c glycosylated hemoglobin, ALT alanine aminotransferase, AST aspartate aminotransferase, γ-GT γ-glutamyltransferase, BUN blood urea nitrogen, Cr creatinine, UA uric acid, TC total cholesterol, TG total triglycerides, LDL-C low-density lipoprotein cholesterol.  Comparison of clinical characteristics and biochemical indexes between two groups after PSM matching A PSM analysis was performed using age, family history of diabetes, prepregnancy BMI, ALT, AST, γ-GT, BUN, Cr, UA, TC, TG, LDL, miscarriage, reproductive history as covariates. As a result, there were 148 (74 in each group) well-matched participants who maintained balance in terms of baseline characteristics (P > 0.05) ( Table 3). We further detected the levels of CFBa in the two groups and found that the level of CFBa in the GDM group was higher than that in the normal pregnant women group [0.4(0.1-0.8) vs. 0.2(0.2-0.3), P = 0.031]. Binary logistic regression was then performed. The results con rmed that the level of CFBa was closely related to the incidence of GDM (OR = 1.52, 95% CI: 1.25-1.85, P = 0.000). The perinatal data of PSM matched pregnant women were shown in Table 4. There was no signi cant difference in gestational age at delivery, cesarean section rate, incidence of premature rupture of membranes, intrauterine distress rate, neonatal weight, proportion of macrosomia and apgar score between the two groups. The amount of postpartum hemorrhage in normal control group was signi cantly lower than that in GDM group [280.0(247.5-312.5) vs. 300(260.0-352.5), P = 0.042].  Continuous data were expressed as mean ± standard deviation or as median (interquartile range [IQR] 25th percentile-75th percentile). The two-tailed Student's test was used for UA, TC, LDL, OGTT 1hPG, OGTT 2hPG. Chi-squared test was used for family history of diabetes, miscarriage and reproductive history. Mann-Whitney U test was used to analyse the other parameters.
GDM gestational diabetes mellitus, BMI body mass index, HbA1c glycosylated hemoglobin, ALT alanine aminotransferase, AST aspartate aminotransferase, γ-GT γ-glutamyltransferase, BUN blood urea nitrogen, Cr creatinine, UA uric acid, TC total cholesterol, TG total triglycerides, LDL-C low-density lipoprotein cholesterol, PSM propensity score matching, CFBa complement factor Ba. Effect of complement factor Ba on the incidence of GDM ROC curve analysis showed that the area under curve (AUC) of CFBa for predicting GDM was 0.602 ( 95% CI, 0.499-0.704; P = 0.033). According to the quartile of CFBa level, we divide it into three groups, category 1(< 0.16 ng/ml, n = 39), category 2 (0.16-0.41 ng/ml, n = 73), category 3(> 0.41 ng/ml, n = 36), as shown in Fig. 1. The results indicated that the incidence of GDM in category 3 was markedly higher than that in the rst and the second categories.

Discussion
This is the rst study to assess the effect of CFBa level within 11 to 17 weeks of gestation on the risk of subsequent GDM. In this study, a propensity score-matched analysis was used, which could balance the distribution of baseline characteristics among groups, reduce the in uence of bias and confounding variables in non-random studies and make more reasonable comparison between the experimental group and the control group [19]. Before PSM, the age, family history of diabetes, prepregnancy BMI, reproductive history, ALT, UA, TC, TG, LDL in GDM group were higher than those in normal pregnant group. After PSM, there was no signi cant difference of the above indexes between the two groups. We found that the levels of CFBa in pregnant women who later developed GDM were signi cantly higher than those in healthy pregnant women. Logistic regression results con rmed that the level of CFBa was an independent risk factor for the occurrence of GDM. Further grouping according to the quartile of CFBa level, it was found that the incidence of GDM in category 3 was markedly higher than that in the rst and the second categories.
At present, insulin resistance and low-grade in ammation caused by immune imbalance are the important reasons for the formation and development of GDM [20]. The chronic in ammation may lead to increased expression of complement gene and activation of complement pathway. Many studies have investigated the role of CFB in metabolic syndrome including type 2 diabetes mellitus, hypertension and insulin resistance. Moreno-navarrete JM et al. [21] found that the circulating CFB concentration was closely related to IR. The CFB gene expression is highly in omentum tissue, which is positively correlated with BMI, log fasting triglycerides and fasting glucose. Fujita T et al. [22] conducted a study including 32 obese patients with type 2 diabetes mellitus and 32 healthy people showed that the circulating level of CFB was signi cantly higher in type 2 diabetes mellitus than that in healthy people, and it was also associated with IR, atherosclerosis and diabetic microangiopathy. It is suggested that the activation of the alternative complement pathway can enhance the sustained mild in ammation and insulin resistance, thus inducing the activation of macrophages and accelerating tissue damage. In recent years, an animal experiment was carried out by Coan PM et al. [11] to con rm the role of serum CFB in metabolic syndrome. They knocked out the CFB gene in rats and found that the glucose tolerance and insulin sensitivity of rats were improved. As the activation product of CFB, CFBa has been rarely reported in the population.
GDM can lead to a series of maternal and fetal complications, such as macrosomia, preterm birth and cesarean section [23]. In our study, before PSM, the incidence of macrosomia in GDM group was higher than that in normal pregnant women group (11.25% vs. 4.39%, P = 0.037). After PSM, compared with the normal control group, the amount of postpartum hemorrhage of GDM group was more. The percentage of cesarean section in the GDM group was also higher than that in the normal pregnant women group, but it did not reach statistical signi cance. This may be due to the psychosocial factors, such as maternal anxiety about delivery or the mother's desire to have cesarean section without any medical indications [24][25][26], or it may be due to the small sample size of our study.
There are some limitations in the study. First, due to the lack of serum insulin data, it is impossible to calculate the homeostasis model assessment of insulin resistance index (HOMA-IR) and the homeostasis model assessment of beta cell insulin secretion (HOMA-β), so the relationship between CFBa and IR could not be further evaluated. In addition, the sample size of the study was relatively small. More patients are needed to enroll and follow up in the future and con rm these ndings.

Conclusion
In summary, this study revealed that the level of CFBa within 11 to 17 weeks of gestation in the GDM group was signi cantly higher than that in the normal pregnant women, which was an independent risk factor for subsequent GDM. Pregnant women with a high level of CFBa within 11 to 17 weeks of gestation were more likely to develop GDM.