Case Study

A healthy nulliparous patient at term entered labor spontaneously and was admitted to the labor floor. At 3 cm of cervical dilation, an uncomplicated labor epidural was initiated to successfully provide analgesia with a patient-controlled regimen of 0.1% bupivacaine with fentanyl 2 µg/ml. The patient’s membranes were ruptured artificially 2 hours later, and oxytocin augmentation of labor was begun. Six hours after initiation of epidural analgesia, the patient was dilated 6 cm. Her temperature measured orally was increased to 38.0°C (100.4°F), after having been admitted with a normal temperature of 37.0°C (98.6°F). Antibiotics were started by her obstetrician. Two hours later, she was still 6 cm dilated, and she was taken to the operating room for cesarean delivery. A male infant was delivered with Apgar scores of 5 and 9 at 1 and 5 minutes, respectively. He required brief bag-mask ventilation for treatment of a low heart rate immediately after delivery but recovered promptly and was discharged home with the mother on postdelivery day 4.

Key Points

• 
  

  This patient experienced an increase in temperature during administration of epidural labor analgesia.

  
  
• 
  

  While the etiology of intrapartum fever is diverse, neuraxial analgesia has emerged as a leading cause of predominately noninfectious but inflammatory fever.

  
  
• 
  

  Her slow labor progress and fever led her obstetrician to deliver her by cesarean, and her baby experienced transient respiratory depression after birth.

  
  
• 
  

  The precise etiology and consequences of epidural-associated fever remain incompletely characterized but are a concern for the obstetric anesthetist, obstetrician, and neonatologist.

Discussion

Women receiving labor epidural analgesia experience a greater incidence of clinical fever. Two decades ago, observational studies demonstrated a gradual rise in temperature in women receiving epidurals compared with those receiving no analgesia or systemic opioids alone.^{1, 2} Although originally ascribed to altered thermoregulation, it is now evident that this represents an artifact from averaging afebrile women’s temperatures and those developing clinical fever.^{3, 4} Substantial evidence suggests that true clinical fever develops more often in women with epidurals than in those without. The data come from retrospective observations (which attempt to statistically control for confounding factors such as longer labors, prolonged rupture of membranes, and number of cervical examinations), sentinel event studies (in which epidural analgesia suddenly becomes available to a population), and randomized, controlled trials.^5–7 Once dismissed as a physiologic curiosity or artifact of selection bias, this phenomenon can no longer be ignored.

The mechanism of epidural-associated fever remains unclear. Earlier, investigators noting the very slow gradual rise in temperature on average hypothesized that thermoregulation might be altered in laboring women with epidurals. For example, by inhibiting sweating and hyperventilation, the neuraxial block might impair heat dissipation. Other investigators suggested that opioids given to women without epidurals might be suppressing temperature elevation rather than epidurals causing it. There is some evidence from nonlaboring volunteers to support each of these mechanisms. However, the observation that “epidural fever” is actually clinical fever in some women averaged with normal temperature in others leads to a search for an explanation of more frequent overt fever in some women with epidural analgesia. Some evidence suggests that placental inflammation (chorioamnionitis) is more common in febrile women with epidural analgesia.⁸ Importantly, in a nonrandomized comparison, bacterial infection in the placenta was not found to be associated with epidural analgesia, although inflammation was.⁹ Furthermore, prophylactic treatment with broad-spectrum antibiotics did not inhibit maternal fever or placental inflammation.¹⁰ Moreover, high-dose methylprednisolone (but not acetaminophen) given to laboring women with epidurals blocks the febrile response.¹¹ Unfortunately, this treatment also leads to a substantial increase in asymptomatic bacteremia in the exposed babies, making it untenable as a clinical strategy. Thus the mechanism of epidural-associated fever is most likely sterile inflammation. Recently, some have suggested that bupivacaine itself may trigger noninfectious inflammation.¹² However, epidural-related fever has been observed with similar frequency when other drugs are used or even when solely intrathecal opioids are used, making this explanation incomplete.

Variations in epidural technique do not appear to have an important effect on the phenomenon. Increased fever has been observed with a variety of local anesthetics and opioid adjuncts, with the combined spinal-epidural technique as well as conventional epidurals, with intermittent versus continuous administration of local anesthetics, and even with continuous intrathecal opioid with no local anesthetic.¹³

Epidural-associated fever may have significant effects on the fetus and newborn. One of the earliest recognized effects was an indirect one, namely an influence on the practice of neonatologists. Lieberman⁵ demonstrated that babies born to mothers with epidurals underwent evaluation for sepsis four times more often than babies born to mothers electing for natural childbirth or systemic opioids. Actual sepsis was vanishingly rare and did not differ between epidural and no-epidural groups. This phenomenon may be a function of neonatology practice style, however, because other institutions have reported different results.⁸ Other adverse effects related to intrapartum maternal fever include increased need for bag-mask ventilation and increased incidence of otherwise unexplained neonatal seizures.¹⁴ Obstetricians may also be more likely to perform operative delivery in the setting of fever,¹⁵ although it is unclear whether this reflects an effect on the labor process itself or an effect on obstetrical decision making.

A far more worrisome possibility is that maternal fever may cause neonatal brain injury. Over 50 years ago, an association between cerebral palsy and maternal fever was first noted, but the observation was not investigated further until recently. Substantial epidemiologic evidence now confirms a four- to nine-fold increase in the risk of otherwise unexplained cerebral palsy in term and near-term infants exposed to maternal fever or clinical or pathologically diagnosed chorioamnionitis.^{16, 17} Other neonatal brain injuries have likewise been associated with maternal fever, including neonatal encephalopathy (which often has devastating lifelong neurologic consequences for the infant).¹⁸ Even babies not manifesting these profound impairments may nonetheless experience neurologic injury. Dammann¹⁹ found that cognitive deficits at age nine, as measured by the Kaufman Assessment Battery for Children, were four times as common in children whose mothers had fever at the time of delivery compared with control individuals.

The link between maternal fever and neurologic injury in the newborn is most likely inflammation. In experimental pregnant animal preparations, bacterial intrauterine infection causes white matter lesions in the fetuses. Yoon²⁰ injected Eschericia coli or saline into the cervix of pregnant rabbits. Thirty-six percent of infected dams delivered fetuses with white matter lesions (6 percent of all fetuses of infected mothers had such lesions) compared with 0 percent in the saline control group. Similarly, Rodts-Palenik²¹ demonstrated white matter lesions and enlarged cerebral ventricles in pups of infected pregnant rats more frequently than in saline-treated rats. Importantly, these lesions were not seen if the mothers were treated with the anti-inflammatory interleukin 10 (IL-10). Although such experiments do not definitively demonstrate that maternal intrapartum fever causes neonatal brain injury via fetal inflammation in humans, Yoon²² has documented increased IL-6 and IL-8 (pro-inflammatory cytokines) in amniotic fluid in a cohort of pregnancies resulting in babies with cerebral palsy compared with controls with normal brain development. In our own laboratory, we have demonstrated similar brain injury in fetuses of mothers exposed to a noninfectious inflammatory stimulus (IL-6) injected for a few hours before delivery, mimicking the clinical situation of epidural fever.²³

Many questions remain. First, it is far from clear how epidural analgesia is associated with maternal inflammatory fever. If the neuraxial block itself causes inflammation, the mechanism would be unique and startling. Conversely, if neuraxial analgesia blocks pathways that might otherwise oppose inflammation in susceptible mothers, then it may be possible to block inflammation in selected women. Second, it is not clear yet whether fever itself can cause injury or whether inflammation causes both fever and brain injury. Hyperthermia itself can worsen the effects of other brain-injuring stimuli, so the two effects may be additive. Third, it is not known whether epidural-associated fever is specifically associated with brain injury. Fourth, it is unknown whether epidural-associated fever can be safely blocked. Trials of various anti-inflammatory strategies are ongoing. These questions will likely be the subject of intense investigation in the near future.