Rates and causes of maternal sepsis vary widely between developed and developing countries. In developing countries, infectious agents such as HIV and malaria account for a significant proportion of infection coincident with pregnancy while in developed countries these agents are much less common. The incidence of early pregnancy complications such as septic abortion vary with access to care and use of antibiotic prophylaxis, and again are far more common in developing countries. Although the obstetric anesthesiologist may be called upon to participate in the care of women with serious postpartum infection such as intraabdominal abcesses, infected episiotomies, and wound debridement, these infections have much in common with typical polymicrobial postsurgical deep tissue infections. Therefore, this chapter will focus on unique features of intrapartum fever (maternal temperature >38°C) and maternal infection related to obstetric causes. In addition, the chapter will discuss the treatment of maternal sepsis, which remains a leading cause of maternal death (1), and how maternal physiologic changes may alter the typical classification, triage, and treatment.

Rates of intrapartum fever vary widely with patient risk factors for infection, parity, and use of epidural analgesia. The lowest rates of intrapartum fever are seen in low risk, predominantly parous populations with lower rates of epidural analgesia. Traditionally, chorioamnionitis is diagnosed based on the three following criteria (2):

Epidural analgesia is often established relatively early in labor prior to the onset of maternal fever; therefore, uterine tenderness is not a particularly helpful clinical sign. Vaginal discharge has always been subjectively defined and is imprecise. Other ancillary signs such as maternal or fetal tachycardia are closely correlated to maternal and fetal hyperthermia and can be present whenever maternal fever occurs (3). It is well recognized that epidurals are associated with an increase in core temperature in labor, though the reason for this remains unknown. This is discussed in greater detail in a following section. Maternal white blood cell counts increase with duration of labor and can be markedly elevated without any evidence of infection (4). Further, white blood cell counts cannot be used to reliably distinguish between chorioamnionitis and epidural associated fever, although significant shift toward immature forms (bandemia) is concerning at any absolute white cell count. Ultimately, there is no reliable way to distinguish between infectious and non-infectious (epidural related) fever; intrapartum fever likely represents a combination of both of these types (5). Due to the potential neonatal sepsis risk, a conservative approach, where all maternal intrapartum fever is an indication to administer maternal antibiotics, is recommended.

Chorioamnionitis is defined as inflammation of the placenta, membranes, amniotic fluid, maternal decidua and in many cases, the fetus, usually due to bacterial infection. Risk factors for chorioamnionitis are largely associated with longer labors or prolonged rupture of membranes (6,7). Maternal bacterial vaginosis and colonization with Group B streptococcus are also weak risk factors (8,9,10). In addition, labor at <37 weeks’ gestation itself is a significant risk factor, with a rate of intrauterine infection estimated at 25% to 45% (11). Although the number of vaginal examinations during labor had been thought to be associated with infection, more recent investigations suggest that this is not an independent risk factor when the length of labor is scrupulously controlled for in multivariable analysis (12). The risk of neonatal sepsis rises with increasing maternal temperature: Approximately 2% with maternal temperatures <38.6°C and 6% with temperatures ≥38.6°C (13). Neonatal sepsis evaluations are recommended in infants born to women diagnosed with chorioamnionitis; however, the scope of sepsis evaluation and treatment varies locally.

Maternal treatment for chorioamnionitis should include broad spectrum coverage. Prior to maternal screening and treatment for maternal Group B Streptococcus carriage, maternal intrapartum antibiotic treatment reduced the risk of neonatal sepsis by up to 86% (14,15). Changes in screening practices and antibiotic prophylaxis may have altered the efficacy of treatment of suspected maternal chorioamnionitis in the prevention of neonatal sepsis; little modern data is available. Group B streptococcus and other organisms associated with obstetric infections are shown in Table 20-1. Standard therapy includes a combination of ampicillin and gentamicin. While 8-hour dosing of gentamicin is most common, there is some evidence to support daily dosing as being equally effective (16). Alternate antibiotic regimens include clindamycin or cefoxitin. Labor should be managed routinely and cesarean delivery reserved for the usual indications. Truncating fetal exposure to infection has not been associated with improved outcomes and maternal antibiotic treatment should result in therapeutic fetal levels of antibiotics. Further, cesarean delivery in the setting of chorioamnionitis is associated with increased maternal morbidity. Chorioamnionitis is a risk factor for maternal uterine atony (RR 2.5; 95% CI 2.2–2.8,17).
Therefore, it is prudent to have uterotonic agents immediately available prior to delivery in these women.

There is little definitive evidence to guide anesthetic management of women with suspected chorioamnionitis. While known sepsis is generally suggested as a contraindication to regional analgesia, the rate of maternal bacteremia is estimated at 5.2% to 9.2% in the setting of clinical chorioamnionitis (18). Two retrospective cohort studies describe a total of 850 women with chorionamnionitis who received epidural analgesia without infectious complications (19,20). Further, only 19% (166/850) received antibiotic treatment prior to epidural analgesia. Therefore, while there is a theoretical risk of seeding the epidural or intrathecal space during the placement of regional anesthesia, the actual risks are exceedingly low. There is no data from obstetric populations to guide a recommendation of the relative safety of epidural versus spinal anesthesia. Overall it seems reasonable to perform regional analgesia in women with a clinical diagnosis of chorioamnionitis in the absence of overt signs of sepsis. Ideally, antibiotics should be initiated prior to regional analgesia.

The association between epidural analgesia and progressive increase in maternal temperature was first described in 1989 (3). Since then, an increased risk of intrapartum fever in women receiving epidural analgesia has been consistently confirmed in randomized studies (21,22,23,24). The primary clinical risk factor for developing fever after epidural analgesia is duration of exposure; therefore, multiparous patients are rarely at increased risk. The widespread introduction of epidural analgesia may be an important influence on the temporal incidence rates of intrapartum fever. Historical rates of intrapartum fever were generally reported to be 1% to 5% (2). Current rates of intrapartum fever in nulliparas have been reported between 13% and 33% (Table 20-2) while rates in multiparous patients are generally not increased (27).

The rate of increase in maternal temperature in women with epidural analgesia is controversial. Initial studies (3,28) presented mean rises in temperature on the assumption that the mechanism for epidural related hyperthermia was thermoregulatory and therefore logically would affect all women equally. However, more recent studies have suggested that temperature response to epidural analgesia may not be uniform. Nulliparous women who ultimately remain afebrile over the entire course of labor have no increase in temperature in the first 4 hours following epidural analgesia while women who ultimately become febrile have an immediate response that is significantly within 1 hour (Fig. 20-1). In women with a predisposition to hyperthermia, temperature increases can be rapid, averaging 0.33°F/h (29). This is consistent with observational studies that demonstrate a significant increased risk of maternal fever >38°C after 4 to 6 hours of exposure to epidural analgesia (25).

The etiology of epidural related fever also remains controversial. Proposed etiologies include a perturbation of maternal thermoregulation, acquired intrapartum infection, and non-infectious inflammation (30,31). Since the majority of women with epidural analgesia do not experience any increase in temperature with epidural analgesia (29), it is difficult to support a thermoregulatory mechanism based on the physiologic effects of epidural analgesia. In contrast, several studies have identified pre-epidural maternal inflammation, measured by maternal serum interleukin-6 (IL-6) levels, as a significant risk factor for subsequent fever (32,33). Women with early labor IL-6 levels in the highest quartile have a markedly increased rate of subsequent fever (Fig. 20-2). There is no evidence that underlying maternal serum IL-6 levels are higher in women choosing epidural analgesia (33). Therefore, there does not appear to be a selection bias for an increased risk of inflammation/fever in women selecting epidural analgesia. Women with the tumor necrosis factor (TNFα) Δ308 polymorphism, which increases
levels of this pro-inflammatory cytokine, have an increased risk of intrapartum fever (24.4%) compared with controls (RR 3.3; 95% CI 1.3–7.1,34). Perhaps the most compelling evidence for an inflammatory etiology is that prophylactic administration of maternal corticosteroids immediately prior to placement of epidural analgesia reduces the risk of subsequent fever by more than 90% (35). The source of the maternal inflammation is not well understood although an increase in placental inflammation has been observed in women with fever following epidural analgesia (36) and placental inflammation may account, in part, for the unique temperature response to epidural analgesia observed in pregnancy.

While acquired infection undoubtedly accounts for a small portion of observed intrapartum fever at term, epidural analgesia itself should not be associated with a significant increased risk of acquired maternal infection. Placentas of women receiving epidural analgesia do not show significant rates of infection, whether associated with fever or not (33). Critically, epidural analgesia should have a minimal effect on the risk of prolonged labor, the most powerful risk factor for infectious fever. In most summary analyses, the effect of epidural analgesia on the duration of labor is small (37), although an increased rate of oxytocin use is also required. Again, while vaginal examinations may occur more often in women with superior pain control, this is not associated with increased risk of infection (38). Finally, antibiotic prophylaxis for group B streptococcus is not associated with decreased rates of fever following epidural analgesia, again arguing against an infectious etiology.

Cytokines can cause fever directly through stimulation of prostaglandin synthesis in the preoptic area of the hypothalamus. Several other factors likely modulate the degree of temperature elevation in women following epidural analgesia. Epidural analgesia has been associated with maternal shivering and rigors and in turn maternal shivering has been associated with fever following epidural analgesia (39,40). Shivering and shaking rigors are common in systemic infection and likely to be cytokine mediated. The administration of IL-6 to healthy volunteers induced shivering that was dose dependent (41). Therefore, one alternate pathway through which elevated levels of maternal inflammation may result in subsequent hyperthermia is through heat generating rigors. However, rigors are unlikely to be the sole pathway to hyperthermia as paralysis only slightly reduces the febrile response to interleukin-2 (IL-2) in non-pregnant subjects (Fig. 20-3) (42). Maternal opioids have also been associated with a small decrease in maternal temperature following epidural analgesia in some but not all studies (28,43,44). This effect may be mediated through a decrease in shivering (45) and/or a direct inhibition of cytokine release (46). Finally, alterations in sweating and hyperventilation with epidural analgesia may have minor effects on maternal temperature regulation.