Perinatal outcomes and risk factors for epidural analgesia-associated intrapartum maternal fever: a retrospective study

Abstract Background Epidural analgesia (EA) increases the risks of maternal fever during labor, which is associated with adverse maternal and neonatal outcomes, while the risk factors for epidural-associated fever and strategies for minimizing these effects remain limited. Methods A total of 325 pregnant women were retrospectively analyzed who had attended our hospital for a vaginal in-hospital delivery, including 208 who voluntarily accepted EA and 117 who did not receive EA. During labor, 208 EA women were allocated to a fever group (n = 42, a tympanic temperature ≥37.5 °C during labor), and a no fever group (n = 166). The outcome measures included main maternal and neonatal outcomes, labor times, duration of EA and the total EA dosage administered. Results 42 out of 208 women given EA exhibited fever temperatures during labor, which were higher than in women who did not receive EA (20.19% vs. 0.85%). Maternal fever had an increased risks for conversion to surgery (adjusted odds ratio (AOR), 4.05; 95% CI, 1.44–11.39) and neonatal infections (5.13; 1.98–13.29) compared to the no fever group. While maternal fever did not increase the risks for assisted vaginal delivery, fetal distress or admission to the neonatal intensive care unit (NICU), it was predominantly associated with primiparity and lesser times of gravity. Frequent cervical examinations, the duration of first stage and total labor, and the duration of EA and its total dosage were positively correlated with the incidence of fever. Furthermore, after stratifying risk factors into subgroups, we found that more frequent cervical examinations (≥7 times) and longer duration of first stage (≥442.5 min), total labor time (≥490 min), EA (≥610.0 min) increased the risk for epidural-associated fever after adjustment for potential confounding factors. Conclusions EA increased the risk of intrapartum epidural-associated fever, which was correlated with adverse perinatal outcomes. Nulliparity, less times of gravidity, ≥7 cervical examinations, increased volume of the EA dosage, prolonged duration of EA and total labor time were risk factors for epidural-associated fever. The findings provide clinicians with insights and strategies to prevent epidural-associated fever more safely and effectively.

pregnant women who received epidural or spinal anesthesia for elective cesarean delivery. Even in the absence of bacteriological evidence of vaginal, uterine or urinary tract infections, pregnant women were found to be more easily predisposed to the development of pyrexia due to ERMF. Broad-spectrum antibiotic administration [6] or acetaminophen [8] did not reduce the incidence of ERMF, suggesting that ERMF is a sterile (noninfectious) inflammatory process [9]. The unidentified pathological process led us to explore the risk factors for ERMF. About 15-25% of a cohort of pregnant women who received EA presented with fever [4]. Numerous studies have demonstrated that maternal fever during labor is likely mediated by various circumstantial factors such as nulliparity, a difficult labor, and prolonged ruptured membranes, all collectively contributing to its development [10].
Whether there are other risk factors for ERMF, and its consequences for the mother and baby, and future strategies for minimizing these effects remain unknown. Hence, our interest mainly lay in examining latent contributing factors such as the duration of labor, frequency of cervical examinations, duration of EA, and the dosage of patient-controlled epidural analgesia (PCEA) administered. The aim was to investigate a number of clinical features related to ERMF and therefore provide a reference for clinicians to carry out appropriate interventions in pregnant women undergoing or predisposed to ERMF.

Study population
This was a retrospective study that utilized data available in the electronic medical records system of the Maternity and Child Health Hospital of Songjiang District, Shanghai, China, and was based on in-hospital deliveries from Jan to Jun 2021. The inclusion criteria were: (1) women with a full-term singleton pregnancy (!37 weeks); (2) scheduled vaginal in-hospital delivery; (3) vertex presentation; (4) willingness to use PCEA and ability to provide informed consent. The exclusion criteria were: (1) women with pre-pregnancy hypertension, diabetes before pregnancy, multiple pregnancies, and stillbirth; (2) women had undergone a previous cesarean section; (3) pre-epidural temperature ! 37.5 C; (4) pregnant women with pre-epidural acute inflammatory diseases (such as genital tract acute and upper respiratory tract infections, etc.); (5) induction of labor with balloon and prostaglandin preparation; (6) history of myocardial infarction, heart failure, maternal hepatitis, or HIV infection; (7) taking paracetamol within 6 h before the study; (8) fetal coagulopathy; (9) hemoglobin < 8 g/L or blood platelet counts < 80 Â 10 9 /L; (10) lumbar disk herniation. Based on the above inclusion and exclusion criteria, data from 208 women who underwent EA were analyzed. A total of 117 pregnant women who did not receive EA (non-EA) and gave birth during the same period were randomly selected as controls. Also, 208 women who received were allocated to an ERMF group (n ¼ 42) or a no fever group (n ¼ 166). Ethics approval was obtained from the research ethics committee of the Child Health Hospital of Songjiang District and written informed consent was provided by all participants prior to participation in the study.

Body temperature measurement and treatment of fever
Body temperature measurement were recorded from the onset of labor once every 60 min (infrared ear thermometer, Braun pro4000) until delivery. ERMF was defined as an intrapartum fever (tympanic temperature ! 37.5 C) related to the administration of EA during labor [11]. If the parturient developed a fever, prophylactic antibiotics were used and fluid replacement was also accelerated to replenish the water lost as the tympanic temperature rose to 38 C-38.5 C. If the tympanic temperature rose above 38.5 C, ibuprofen was administered orally.

PCEA
An epidural catheter was introduced at L3-L4 when contractions became regular and cervical dilation reached 2-3 cm. Women were given a test dose of analgesia including 1:200,000 adrenaline and 1.5% lidocaine after the successful puncture. After 5 min, a 10 mL bolus of ropivacaine 0.075% (lot number: H20140763; Astra Zeneca, Sweden) plus 0.5 lg/mL sufentanil (batch number: H20054171; Yichang Renfu Pharmaceutical, China) was administered. Then, the epidural catheter was connected to a programmed intermittent epidural bolus (PIEB) pump with the following parameters: PIEB settings were a 9 mL bolus every 50 min; injection rate, 1 mL/h; PCEA (demand) dose, 8 mL; maximum dose, 30 mL/h a and lockout interval of 10 min. The analgesic level was assessed according to a patient's sensitivity to temperature (ice) after 15 min. Additional 5 mL boluses of 0.1% ropivacaine were given for breakthrough pain and a visual analogue scale pain score < 3 at any time during labor.

Data collection and independent variables
The hospital's electronic medical records database stores data on each birth delivered in the hospital. Maternal demographic, anthropometric characteristics and reproductive history were collected at the first prenatal visit between 9 and 14 weeks gestation. Analgesic data, including the duration of EA, total EA and PCEA doses were also collected. Body temperature and changes in the parturients, including intrapartum fever status were also recorded. The duration of labor, frequency of cervical examinations and conversions to obstetric forceps or cesarean sections were also analyzed. Maternal blood pressure and heart rate were continuously monitored during labor and the rate and volume of fluid administration carefully recorded.

Definition of outcomes
The main complications of pregnancy including gestational hypertension, preeclampsia (PE), gestational diabetes mellitus (GDM), intrahepatic cholestasis of pregnancy (ICP) were coded according to the 10 th revision of the International Statistical Classification of Diseases (ICD-10) and diagnosed based on our previous study [12]. Pregnancy anemia was defined as a hemoglobin concentration (Hb) <110 g/L at sea level during pregnancy [13,14]. Neonatal infections were defined as the systemic inflammatory response syndrome caused by infection due to various pathogens (including bacteria, viruses, protozoa, etc.), which was diagnosed when neonates presented with jaundice, vomiting, fever, dyspnea or other symptoms and met any 2 of the following criteria: (1) Neonates with white blood cell count !30 Â 10 9 /L at 6 h to 3 days old, or !20 Â 10 9 /L at ages ! 3 days old, or <5 Â 10 9 /L at any day of age; (2) C-reactive protein !3 mg/L within 6 h of birth or !5 mg/L between 6 and 24 h or !10 mg/L in neonates >24 h old; (3) Positive blood culture. Primary postpartum hemorrhage (PPH) was defined as bleeding from the genital tract of 500 mL or 1000 mL within the first 24 h after a vaginal or cesarean section delivery, respectively [15]. PPH was usually measured by gravimetric or visual protocols [16].

Statistical analysis
Continuous variables are expressed as the mean ± SD. Independent-samples t-test, non-parametric and chi-squared tests were used to evaluate the statistical significance between the two groups. Categorical variables are presented as frequencies with proportions. Univariate logistic regression analysis was used to estimate odds ratios (ORs) with a 95% confidence interval (CI) into assessing potential associations between EA and the risk of maternal fever, and the associations between maternal fever and the risks for adverse maternal and neonatal outcomes. The ORs for the risk of maternal fever and adverse perinatal outcomes were also calculated. The frequency of cervical examinations, time of first stage, total labor time, amount of drugs administered and the total EA duration did not conform to normal distributions; therefore, these data were stratified into quartile value subgroups. Similarly, the absolute risks were estimated for maternal fever among women who underwent EA across the range of risk factor categories.
Multivariable analyses were adjusted for potential confounders. Maternal age ( 24, 25-29, 30-34, or !35 years), height ( 154, 155-164, 165-174, !175 cm), late pregnancy BMI ranges (BMI, 24.9, 25.0-29.9, and !30 kg/m 2 ), parity (once, or more than twice), gravidity (once, twice or more than three times), pregnancy complications GDM (no or yes), gestational hypertensive disease (no or yes), ICP (no or yes), cephalopelvic disproportion (no or yes), and pregnancy anemia (no or yes) were included as covariates and to assess any associations between EA and the incidence of maternal fever, maternal fever and the incidence of perinatal outcomes, and risk factors for the overall incidence of maternal fever. To reduce the bias caused by low sample capacity, when evaluated the association between the frequency of cervical examinations and the risk of maternal fever, women were classified into three subgroups as follows: <5, 5-6 and !7 times. A CI that contained a 1 meant that there was no significant difference between the event and the control in terms of risk. All statistical analyses were performed using SPSS package version 21.0 (SPSS Inc., USA). A two-tailed p-value <.05 was used as the threshold for statistical significance.

Baseline characteristics of the study population
We enrolled 208 women with a mean (± SD) age of 28.76 (± 3.76) years and 117 pregnant women with a mean (± SD) age of 30.03 (± 4.81) years who did not receive EA. During the study period, there were 42 (20.19%) maternal fevers among the 208 women who underwent EA. While only 1 out of 117 women not given EA presented with fever during labor. Women who underwent ERMF were more likely to be less gravid and nulliparous compared to the no fever group. Similar age and BMI at delivery, and the rates of pregnancy complications such as GDM, gestational hypertensive disease, cephalopelvic disproportion, pregnancy anemia are presented for the no fever and fever groups in Table 1.
Maternal fever and risks of adverse maternal and neonatal outcomes among women who underwent EA To determine the effect of epidural-associated fever, we calculated the baseline characteristics, maternal and neonatal outcomes in EA participants. Similar age, height, BMI and gestational weeks were found among the fever and no fever groups. Women with ERMF had higher white blood cell and neutrophil counts, while the concentrations of interleukin 6 and C-reactive protein were not significantly different between the fever and no fever groups (Table 2). Indeed, it was found that the amount of drugs given and the total EA duration in the fever group was obviously higher than for the no fever group. The average time between EA administration and the onset of fever was 338.95 ± 55.16 min in the fever group (Table 2).
Furthermore, for maternal outcomes, women with ERMF had longer first-stage labor and total labor time, more frequent cervical examinations and less volume of postpartum hemorrhage ( Table 2). Significant higher rates and risks of conversion to surgery were also found in the fever group compared to the no fever group (28.57% vs.7.53%; adjusted odds ratio (AOR), 4.05; 95% CI, 1.44 to 11.39). However, the rate and risk of forceps delivery, though slightly higher, showed no statistical difference compared with the control group (Table 3). For neonatal outcomes among women who were given EA, the Apgar score of 5-min was lower and rates of fetal distress higher in women with epidural-associated fever, while the risk of fetal distress was not significantly different after being adjusted for maternal age, BMI, parity, gravidity and pregnancy complications (Table 3). In addition, compared with the no fever group, the fever group exhibited a highly increased risk for neonatal infections (35.71% vs. 7.83%; AOR, 5.13; 95% CI, 1.98 to 13.29). No significant difference was found in the 1-min Apgar score or in the rate of neonatal intensive care unit (NICU) between the fever and no fever groups (Tables 2 and 3).
Considering the critical importance of how ERMF influenced obstetric management, we further compared frequent risks factors for cesarean section and forceps delivery to see whether these risks were associated with ERMF. The results revealed that ERMF increased the cesarean section rate that resulted from abnormal labor and fetal distress (Table S1).

Risk factors for ERMF
We further analyzed the risk factors which were related to the development of ERMF. In Table 4, it is shown that maternal pathogen infection and different rupture patterns of membranes did not increase the risk for maternal epidural-associated fever. Women who were nulliparous and with less gravidities showed higher rates of epidural-associated fever. Compared to a first pregnancy, women with !2 times gravidities had a lower risk of developing epidural-associated fever (2 times: 11.67% vs. 29.65%, respectively; AOR, 0.30 [95% CI, 0.12-0.76]; !3 times: 9.52% vs. 29.65%, respectively; AOR, 0.27 [95% CI, 0.08-0.91]).
The results presented in Tables 2 and 3 show timeand dose-dependent correlations between EA and maternal fever during delivery. In order to define feasibility values for clinicians to prevent the epidural-associated fever, the values of frequent cervical examinations, time of first stage, total labor time, amount of medicine, and total EA duration were stratified into quartile value subgroups. The cutoff points of frequency of cervical examinations at the 50th (5 times) and 75th (7 times) percentiles were defined according to the assayed values among women enrolled. The cutoff points of time of the first stage at the 25th (168.8 min), 50th (260 min) and   The association between risk factors and epiduralassociated fever in the four subgroups were also analyzed ( . The rates and risks of developing ERMF were unequivocally positively correlated with the amount of drugs administered. However, women who received more medicine showed no significant increase in risk of developing maternal fever after adjustment for potential confounding factors (Table 4).

Discussion
In the present study, it was demonstrated that EA increased the risk of intrapartum fever, while epiduralassociated fever increased the risk for conversion to surgery and neonatal infections, but exhibited no increase in the risk for postpartum hemorrhage and assisted vaginal delivery, fetal distress or NICU after adjustment for potential confounding factors. The fact needs to be mentioned that the mechanisms by which EA produces maternal hyperthermia during labor remain unclear. The results from a few randomized controlled trials (RCTs) which compared women who received or did not receive EA also found an increased incidence of fever in the epidural groups [17][18][19]. Several explanations such as the role of thermoregulatory factors, systemic opioid administration, and inflammation have been advanced to explain the findings. Infection is believed to be an attributable factor for epidural-related fever due to prolongation of labor and higher the incidence of chorioamnionitis, and the latter condition being observed in several studies [20]. In contrast, cumulative evidence found that epidural-related fever was associated with noninfectious systemic inflammation, with an elevation of serum interleukin 6 (IL-6) and interleukin 8 (IL-8) [9,21]. The theory was further proven indirectly utilizing placental histology [22].
In this study, although we established detailed inclusion and exclusion criteria to eliminate potential maternal factors that may have led to suspected infection, the incidence of maternal fever was still up to 20.19% (42/208), while there < 1% in women with who did not receive EA. Among 42 patients with fever, 35 patients had unexplained epidural-related fever, with neither chorioamnionitis nor elevation of white blood cells, indicating the possible role of noninfectious causes in the pathophysiological process of ERMF. Further studies deciphering the effects of inflammatory cytokines and downstream signaling pathway on systemic noninfectious inflammation will be of critical importance to understand the underlying mechanisms leading to adverse pregnancy and neonatal outcomes.
EA-associated fever may have significant consequences for both the mother and the fetus/neonate. The rates of conversion to surgery in the ERMF group were higher than for the no fever group (28.57% vs. 7.23%), and further risk factor analysis revealed that ERMF likely increased the cesarean section rate resulting from abnormal labor and fetal distress. Whether EA leads to a prolonged labor duration is an interesting yet controversial topic at present. Several retrospective studies have reported the effect of EA in prolonging total labor duration as well as the different stages of labor [23,24], findings of which are in consistent with our study results. Zhang et al. [25] reported similar results that the duration of the active phase of labor was unchanged; on the other hand, the second stage of labor was likely to be prolonged. However, Aneiros et al. [26] found no significant difference between epidural and combined spinal-EA given for labor analgesia in nulliparous women, in terms of the duration of labor, mode of delivery or the total local anesthetic dosage.
Similarly, in addition to the raised risk of conversion to surgery, a significant increase in neonatal infections was found in women with intrapartum maternal fever in our study. Epidemiologic investigations have shown an increase in neural-tube defects in the babies of women exposed to external heat sources in early pregnancy. Hence, we should focus on the effect of intrapartum maternal fever on the neonates or children, and long-term follow-up studies should be carried out on those offspring whose mothers had a history of intrapartum fever, especially in children with neonatal infections. Interestingly, we found the fever group exhibited a lower incidence of postpartum hemorrhage compared to the no fever group, which may due to the increased rates of cesarean section, resulting in oxytocin administration during the earlier stages of labor compared to the no fever group.
Since epidural-related intrapartum fever inevitably develops during many labors, exploring the risk factors is important for those women undergoing EA. The most profound risk factor for ERMF was the duration of EA, considering that body temperature starts to rise after epidural placement. A recent study showed a total EA duration exceeding 6.3 h was associated with an increase in the duration of maternal intrapartum fever and early measures, such as oxytocin administration to accelerate labor, findings in good agreement with our study results. These suggest the role of the duration of EA is of great importance for the development of maternal fever [27]. There is conflicting evidence whether the frequency of cervical examination contributes to the development of ERMF. To date, cumulative studies have indicated that cervical examination is a potential risk factor for maternal fever, findings in good agreement with the results of the present study. We found that women who underwent more cervical examinations (!7 times) had a greater propensity to develop fever during labor [11]. However, one previous retrospective cohort study reported opposite results to ours and concluded that the number cervical examinations performed was not associated with an increased risk of developing ERMF [28]. Maternal fever was also found to be associated with primiparity, and less times of gravity. Marx and Leow [29] reported that temperature progressively increased over time with epidural use, averaging 1.5 C in nulliparous and 0.5 C in parous women. Intrapartum maternal fever was time-and dosedependently correlated with the frequency of cervical examinations, the times of first stage and total labor and the duration of EA. Logistic regression analysis revealed that parturients with more frequent cervical examinations (!7 times), longer time of first stage (!442.5 min), total labor time (!490 min), and a duration of EA (!610.0 min) were more likely to have an increased likelihood of developing fever.
Our results suggest that nulliparous, less gravidity, !7 times cervical examinations, prolonged duration of EA and total labor time are risk factors for epiduralassociated fever, which provides some insights into interventions that will prevent epidural-associated fever in a safer and more effective way.
There are a number of limitations to our study which are worth mentioning. First, the investigators had no control over the quality of the recorded temperature. Second, the study was a retrospective analysis of a limited cohort of patients. Therefore, a further study with an expanded sample sizes will be necessary to confirm our results. Lastly, the study failed to explore the underlying mechanisms because of limited laboratory facilities.

Conclusions
EA was found to be associated with maternal fever and an important risk factor for adverse perinatal outcomes, including higher rates of cesarean section and neonatal infection. Nulliparity, less times of gravidity, !7 times cervical examinations, a prolonged duration of EA and total labor time were risk factors for epidural-associated fever. Limiting the number of cervical examinations, the duration of EA and total labor time are safe and effective strategies to prevent epiduralassociated fever to a certain degree. However, further study is required on how epidural-related intrapartum fever increases the risk of neonatal infections and the mechanisms involved.