Goetzl, Laura MD, MPH; Rivers, Jose MD; Zighelboim, Israel MD; Wali, Ashutosh MD; Badell, Martina MD; Suresh, Maya S. MD

The association between intrapartum epidural analgesia and maternal fever was first described in 1989¹ and has since been confirmed in multiple randomized controlled trials.^2–4 Risk is largely confined to first-time mothers, who have the longest duration of exposure. The absolute risk of intrapartum temperature greater than 100.4ºF in nulliparas with epidural analgesia has been reported to range from 14.5% to 33%. The leading hypothesis regarding the mechanism of epidural fever has been an alteration in maternal thermoregulation.⁵ Two observational studies have prospectively followed maternal temperature after epidural analgesia with hourly temperature measurements. Fusi et al¹ evaluated 18 women (13 nulliparas) after epidural analgesia and reported a significant rise in vaginal temperature after 6 hours. Camann et al⁶ evaluated 40 women (31 nulliparous) after epidural analgesia and reported a rise in mean maternal tympanic temperature of 0.6°C (1.1ºF) over baseline that was significant after 4 hours. Although these studies provided novel insights, the underlying assumption inherent in their study design was that maternal temperature response to epidural analgesia is uniform. The purpose of our study is to evaluate this assumption.

MATERIALS AND METHODS

Our study cohort consisted of full-term nulliparous women requesting epidural analgesia. The cohort was derived from the placebo arm of a recently published double-blind, placebo-controlled trial.⁷ Our investigations were approved by the Institutional Review Board at Baylor College of Medicine (Protocol H-11497). Inclusion criteria included nulliparity, full-term pregnancy (37 weeks or more), vertex presentation, singleton gestation, ability to provide informed consent, and request for epidural analgesia for control of labor pain. Subjects were excluded if their temperature at any time prior to epidural analgesia exceeded 99.4ºF. After epidural catheter placement maternal tympanic temperature was assessed hourly until delivery (Thermoscan Pro3000, Braun Medical, Kronberg, Germany). Room temperature was centrally maintained at 71.8±0.85ºF (TM99A, Cooper Instruments, Middlefield, CT), and independent temperature control was not available in labor rooms.

Statistical analysis was performed with SPSS 8.0 (SPSS Inc, Chicago, IL). Maternal temperatures were compared at each time point with the repeated measures analysis of variance. Rise in temperature over time was evaluated with repeated measures analysis of variance for a change over time. In addition, continuous variables were compared with the Student t test for normally distributed data and the Mann Whitney U test for nonparametric data. Rates were compared with the Fisher exact test. Two-tailed tests were used, and P<.05 was considered significant.

RESULTS

Among 101 nulliparous women, 99 had two or more tympanic temperature samplings before delivery and were included in the analysis. Fever (tympanic temperature greater than 100.4ºF) occurred in 22 (22.2%) at some point before delivery. Mean maternal tympanic temperature after epidural analgesia is shown in Figure 1. A rise in mean maternal tympanic temperature was observed over time, which became statistically significant 4 hours after initiation of epidural analgesia. The absolute rise in mean temperature over 8 hours was 1.3ºF. Of interest, the magnitude of the standard error bars indicates considerable variability. For subsequent analysis we stratified our cohort into those who eventually developed intrapartum fever and those who remained afebrile. The two groups were clinically indistinguishable with respect to their demographic characteristics, initial clinical indicators of infection, maternal temperature at epidural placement, white blood cell count, or group B streptococcus carriage rates (Table 1). Although not statistically significant, subjects who developed an intrapartum fever after epidural analgesia were on average 2.4 years older and carried infants who were 193 g heavier. During labor, intrapartum fever was most strongly associated with duration of exposure to epidural analgesia. Not surprisingly, fetal tachycardia was more frequent in febrile women because there is a strong known correlation between maternal temperature and fetal heart rate (r=0.91).¹ No neonate born to a febrile mother had documented evidence of infection on surveillance blood cultures.

Fig

Table 1

We subsequently reexamined mean maternal tympanic temperature over time stratified by ultimate intrapartum fever status (Fig. 2). We restricted our analysis to the first 4 hours after epidural analgesia to pinpoint thermoregulatory changes rather than acquired infection. In the 77 women (77.8%) who remained afebrile, epidural analgesia was not associated with any temperature increase during the 4 hours after catheter placement (P=.26). In contrast, in the subset of 22 women who subsequently became febrile, mean maternal temperature rose dramatically after initiation of epidural analgesia. This increase was significant as early as 1 hour after catheter placement (P<.05). Over the 4-hour interval, the rise in mean maternal tympanic temperature was 1.3ºF (0.33ºF per hour). In comparison to the unstratified temperature curve, the magnitude of the standard error bars indicates less variability within each subgroup.

Fig

DISCUSSION

To our knowledge, this study represents the largest prospective cohort with hourly evaluation of maternal temperature after epidural analgesia (MEDLINE search January 1966 to November 2006; keywords: “hyperthermia,” “temperature,” “fever,” “epidural analgesia,” “labor”). When the cohort is examined as a whole, our results closely resemble the findings of Camann et al (1.3ºF versus 1.1ºF over 8 hours).⁶ However, the larger size of our cohort allowed us to present a stratified analysis that reveals a bimodal temperature response among nulliparous patients receiving epidural analgesia. The majority of women do not demonstrate an increase in mean tympanic temperature after epidural analgesia. In contrast, a minority subset of women demonstrates a brisk temperature response of higher magnitude than has been previously reported (0.33ºF per hour). Our findings do not support the current theory of a universal perturbation in maternal thermoregulation after epidural analgesia. Hyperthermia appears to be an abnormal response that occurs only in a relatively small subset of women at term. Our recent randomized controlled trial demonstrating that prophylactic maternal methylprednisolone (100 mg intravenously every 4 hours) reduced the rate of intrapartum fever after epidural analgesia by more than 90% also suggests an inflammatory, rather than a thermoregulatory, etiology.⁷

Of equal interest, those women who respond to epidural analgesia with hyperthermia do so immediately. A significant increase in maternal temperature was observed just 1 hour after epidural placement. This early onset of maternal temperature response precedes the accumulation of other maternal exposures to known risk factors for chorioamnionitis (eg, vaginal exams, prolonged membrane rupture). This rapid temperature rise suggests that factors which mediate alterations in maternal temperature response are already present in early labor and that identification of women at risk of subsequent fever may be possible. Although the two subsets of women receiving epidural analgesia appear clinically similar, they may have subclinical alterations in serum or cerebral spinal fluid markers, including altered levels of pro- or anti-inflammatory cytokines. Of note, women with the tumour necrosis factor (TNF) α Δ308 polymorphism, which increases levels of this pro-inflammatory cytokine, have an increased risk of intrapartum fever (24.4%) compared with controls with a relative risk of 3.3 (95% confidence interval 1.3–7.1).⁸ Currently, our group is actively investigating underlying maternal inflammatory balance to develop a strategy for early identification of women at high risk for subsequent intrapartum fever.

Early identification is critical if intervention to prevent fetal exposure to intrauterine hyperthermia and inflammation is found to be beneficial. Recent reports of neonatal hypothermia to treat hypoxic encephalopathy have refocused clinical attention on the interrelationship between fetal and neonatal brain temperature and the extent of brain injury after a hypoxic insult.^9,10 Fetal hyperthermia at term (potentially in combination with intrauterine inflammation) is associated with a more than 4-fold increased risk of neonatal encephalopathy,¹¹ and a 4- to 9-fold increased risk of cerebral palsy.^12,13 Most recently, data has linked fetal cytokine polymorphisms, including heterozygous carriage of the Δ308 polymorphism of TNF α, to an increased risk of cerebral palsy in term infants.¹⁴ The fetal contribution to maternal temperature response is not known, but is an active area of research. Alternatively, fetal cytokine polymorphisms may reflect inherited maternal polymorphisms.

REFERENCES