This study was approved by the China Ethics Committee of Registering Clinical Trials (ChiCTR1900026606) and was registered in the Chinese Clinical Trial Registry. After registering the study, and obtaining written informed consent from all patients, we started this prospective, randomized, controlled, non blinded study. The trial was conducted between October 2019 and February 2020.

The inclusion criteria were healthy nulliparity, age ≥ 20 years, singleton cephalic presentation, term pregnancy (≥37 weeks), cervical dilatation of 1–5 cm, and request for labor neuraxial analgesia. Parturients who had a baseline temperature ≥ 37.5 °C, contraindications to labor neuraxial analgesia, nonreassuring fetal heart rate tracings, or high-risk pregnancies (preeclampsia, placenta previa, known fetal anomalies, severe system comorbidities) were excluded. In addition, parturients who required cesarean delivery within 2 h after analgesia or experienced failed labor neuraxial analgesia were also excluded from this study.

Parturients were randomly allocated to two equal groups using computer-generated random numbers: the CSEA group and the EA group. The chief anesthesia resident with more than 3 years of obstetric anesthesia experience administered labor analgesia. The parturients were coloaded with 500 ml lactated Ringer’s solution and placed in the left lateral position. A convenient lumbar epidural space was located and identified by a 17-gauge Tuohy needle using a loss of resistance to air technique. In the CSEA group, a 26-gauge pencil-point spinal needle perforated the dura using the needle-through-needle technique, and a subarachnoid injection with bupivacaine 2.5 mg with fentanyl 15 μg was performed. After intrathecal injection, a multiorifice epidural catheter was inserted 3–5 cm into the epidural space and secured. If a bilateral T-10 sensory level was not attained or VAS score > 3 within 20 min, CSEA was defined as failed, and the patient was excluded from the study. For the successful CSEA, when the sensory level block down to the T10, a test dose of 3 mL 1.5% lidocaine with epinephrine 1:200,000 was administrated into the epidural space to ensure that an intravascular or subarachnoid puncture was not observed. A bolus of 8–10 ml 0.1% ropivacaine and 0.5 μg/ml sufentanil was administered, and a continuous infusion at a rate of 8 ml/h was established to maintain the sensory block at T10 or higher. If a bilateral T10 sensory level was not attained within 20 min after the epidural injection, an additional 5–10 ml was administered. If a bilateral T10 sensory level was still not attained after a total of 20 ml was administered, CSEA was also defined as failed.

In the EA group, a multi-orifice epidural catheter was directly inserted 3–5 cm into the epidural space, and a test dose of 3 mL 1.5% lidocaine with epinephrine 1:200,000 was administrated. After ensuring that the epidural catheter was not located in the intravascular or subarachnoid space, a bolus of 8–10 ml 0.1% ropivacaine and 0.5 μg/ml sufentanil was administered, and a continuous infusion was maintained at a rate of 8 ml/h to maintain the sensory block at T10 or higher. If a bilateral T10 sensory level was not attained within 20 min after the epidural injection, an additional 5–10 ml was administered. If a bilateral T10 sensory level was still not attained after administering a maximum of 20 ml, epidural analgesia was defined as failed.

After labor neuraxial analgesia, routine maternal and fetal monitoring was continued. The delivery room temperature was maintained between 20 °C and 22 °C. Maternal temperature was measured before analgesia and hourly after analgesia with an electronic thermometer inserted into the ear canal near the tympanum until delivery. The anesthesia nurse was followed up every 1–2 hours and collected the data. Intrapartum fever was defined as an tympanic temperature ≥ 38 °C at least once during analgesia. Parturients with fever routinely received physical cooling and antibiotics. Demographic data such as maternal age, height, weight, gestational age, baseline cervical dilatation, white blood cell count, and neutrophils were recorded. Obstetric characteristics such as the duration of analgesia, duration of analgesia to full dilatation and delivery, duration from rupture of the membranes (ROM) to delivery, cesarean section rate, and oxytocin augmentation rate were collected. Neonatal outcomes such as neonatal weight, antibiotic usage, Apgar scores at 1 min and 5 min were also recorded. The primary outcome was the incidence of intrapartum fever, and the secondary outcomes were the duration of analgesia, analgesia to full cervical dilation and analgesia to delivery.

The sample size was estimated based on the incidence of intrapartum fever in prior studies. The risk of intrapartum fever in epidural analgesia has been reported to range from 15 to 42% [6, 17]. There were only two studies reported the risk of intrapartum fever in CSE analgesia among 7 to 14% [18, 19]. We presumed that the incidence of intrapartum fever in the epidural analgesia group was 22% and that in the CSE analgesia group was 10%. To maintain a power analysis with α = 0.05 and β = 0.2, and considering a 20% dropout rate, we calculated that the sample size to be 179 parturients per group. Finally, 200 parturient women were recruited in each group.

Statistical analysis was performed with SPSS 20.0, and P < 0.05 was considered significant. Continuous variables that were normally distributed are presented as the mean ± standard deviation (SD) and were analysed by Student’s t-test, while nonparametric data were analysed by Mann-Whitney U tests. Categorical variables are presented as rates and were compared by the chi-square or Fisher’s exact test. Repeated variables were analyzed by repeated measures within groups and ANOVA. Maternal temperatures were considered repeated variables and analysed with repeated measures analysis of variance.