Epidural Analgesia and Neonatal Fever

To the Editor. We read with interest the report of Lieberman et al1 regarding the relationship of the duration of epidural analgesia and fever in labor. We have had a long-held hypothesis regarding this relationship based on both the study of Fusi et al2 in 1989 and a retrospective study we performed and presented as a poster at the Society for Pediatric Research in 1990.3Our prior study closely agrees with the conclusions in the Lieberman study. We analyzed computerized data on all 1909 infants admitted to the special care nurseries for a 30-month period, delivered by our obstetrical service with 60% epidural use. Fever in labor was found in 32% of the mothers with epidurals and in 12.5% of the mothers not given epidurals (χ2 94.11, P < .00001). Mothers with epidurals who had fever were given antibiotics more often (P < .05), and term infants born to mothers with epidural and fever received significantly more lab tests and antibiotics (P < .001) without any increase in culture-proven sepsis. Also, the length of stay for these latter infants averaged 6.7 days, compared with 3.0 days for children born to mothers who remained afebrile after epidural. Because our standard practice was to admit all infants of mothers with fever to the special care nurseries, the incidence of fever in the 8400 live births (all levels of gravidarity) was calculated as 5.7% with an incidence of approximately 2.5% in those without epidurals and 7.5% of those with epidural analgesia.3 These numbers, although less dramatic than the 1% and 14.5% found by Lieberman et al1 in their population of 1657 primigravidas, confirm the strong correlation of fever and epidural analgesia. Lieberman et al1 also concluded that there were more sepsis evaluations and antibiotic use in the offspring of women with epidural analgesia, with or without fever in the mother.Currently, our practice is to evaluate labor room records carefully when there is fever in labor, with extra concern about the real need for sepsis evaluations and antibiotic treatment in the infants of these mothers. We carefully evaluate the history and course of the fever if epidural analgesia is used and especially look for intense shivering by maternal history. Our experience is that shivering is so commonplace that it is rarely noted in the anesthesiologist's or delivery room nurse's records. Yet, shivering before temperature rise and a rapid defervescence after discontinuation of the epidural is the most common pattern among our patients without an infectious etiology. Shivering is thought to be the major mechanism of temperature elevation in parturition by Gleeson et al,4 and shivering occurs in approximately 30% of labors under epidural.5 One mechanism for shivering after epidural analgesia is probably related to the presence of thermosensitive tissue in the spinal canal deduced from the prompt induction of shivering after injection of cold anesthetic solutions.6 Ephedrine, the vasoconstrictor commonly given to women in labor to counteract hypotension, increases heat production in animals via the oxidation of carbohydrate7 and, therefore, might also contribute to fever in labor. Other drugs used with epidural analgesia may alter temperature control in other ways such as epidural narcotics that abolish shivering89 and would, therefore, tend to lower temperatures.Like the authors, we too believe that the relationship warrants further study and have also collected concurrent data that is being prepared for publication.To the Editor. We read with interest and some concern the article “Epidural Analgesia, Intrapartum Fever, and Neonatal Sepsis Evaluation,” by Lieberman et al.1-1 As the authors acknowledge, this is a retrospective offshoot of an original study regarding the active management of labor, and, as such, has some inherent problems. Epidural use was not randomized, and as Table 1 in their article indicates, the women who received epidurals are significantly different from those who did not; specifically they had larger infants and more than double the rate of induction of labor. See Figure (page 491) for more information.Most of the septic workups (63%) occurred in infants of mothers who did not have temperature elevations. This fact is mentioned in the article, although discussion or explanation is not offered. Would the authors comment on why these infants received septic workups? It should be emphasized that despite the large number of septic workups, the neonatal infection rate is exceedingly low (4/1657), and is totally unrelated to epidural usage.The data regarding the number of vaginal examinations, the duration of ruptured membranes, and the baseline temperature are absent. The protocol for the active management of labor study mandated early amniotomy and vaginal examinations “at least every 2 hours to ensure prompt detection of inefficient uterine action.”1-2 Could these techniques of labor management, which are not currently the standard of practice here and in most hospitals in North America, result in some of the neonatal sepsis workups for reasons other than related to maternal fever?In Table 4 of the Lieberman article, the phrases “expected cases with 0% epidural use” and “% attributable to epidural use” imply a causal relationship between epidural use and the particular outcome in question; ie, the Table implies that had the woman not received the epidural, then the fever, sepsis workup, or antibiotic treatment would not have occurred. It also assumes that the epidural and no epidural groups are similar, which, as noted above, they are not. Hence, the message of this Table, that the outcomes discussed can be altered by decreasing epidural use, is misleading.Certainly the large number of septic workups is concerning due to the cost and potential morbidity involved, and we agree with Dr Lieberman et al that further study is required to clearly elucidate the factors leading to these tests. This study calls for reevaluation of our policies concerning sepsis workups in neonates, but not our current indications for epidurals.Both the American Society of Anesthesiologists and the American College of Obstetricians and Gynecologists continue to endorse epidural analgesia as the most effective method of pain relief in labor. We agree that patients should be fully informed regarding epidural anesthesia, including the increased risks of the alternative, general anesthesia for an urgent cesarean section, should one become necessary without an epidural catheter already in place.1-3To the Editor. The study by Lieberman et al2-1 is of doubtful scientific value because the trial was not randomized to assess the effect of epidural pain relief on maternal fever and neonatal sepsis. Because it was not a controlled study, it is difficult to accept the authors' conclusions as valid. There are many factors that can increase temperature in a woman in labor.2-2 There was no reference in the study to analgesic techniques in the nonepidural group. It is likely that opioids and other sedatives significantly impair thermoregulatory responses.2-3 Such impairment may be compounded by regional anesthetic techniques. In addition, the method and site of temperature measurement was not standardized, and no allowance was made for ambient temperature, diurnal variations in temperature, and the use of tepid sponging and fans.Selection bias and observer bias are other problems with interpreting this type of data. It is not clear from their article exactly which criteria the pediatricians chose in terms of evaluating the infants for infection. Maternal fever was present in only one third of mothers whose infants were subsequently screened for infection. Perhaps the group in Boston needs to reevaluate the indicators for performing these tests, because their 25% rate of neonatal sepsis evaluations, in a low-risk population, is at odds with other centers. Indeed, at the University of California, San Diego, the rate for neonatal sepsis evaluations is 12% in a population of term parturients at high risk for neonatal infection with access to an on-demand epidural service.It is important to remember the benefits enjoyed by women in labor from the excellent pain relief provided by an epidural. It is misleading to publish data so obviously flawed, because many women who would benefit from the considerable advantages of epidural pain relief may now forfeit them out of an understandable, but misguided, concern for their newborn.To the Editor. The article by Lieberman et al (Pediatrics.1997;99:415–419) raised a number of issues of interest to maternal-child care providers. This retrospective case study purports to show an excess number of intrapartum fevers, neonatal sepsis workups, and neonatal antibiotic therapy secondary to provision of epidural analgesia. Once again, a well-recognized association between epidural analgesia and maternal temperature is demonstrated.3-1-3-3 Causes of this relationship remain unknown, but may involve both anesthetic (reset of hypothalamic setpoint, impaired thermoregulation) and obstetric components (frequency of cervical exams, duration of labor, and/or rupture of membranes). Although this aspect of the study is not surprising, both the methodology and the conclusions reached are somewhat problematic. First, no effort is made to investigate the potential interaction between environmental temperature and epidural analgesia, despite the fact that epidural analgesia may be associated with impaired thermoregulation. The induction of epidural analgesia is often associated with maternal shivering (secondary to a complex interplay of multiple factors3-4), which often results in labor and delivery nurses covering the patient with warm blankets and increasing the room temperature—even when the patient denies being cold. In patients with impaired thermoregulation, such maneuvers can cause increased body temperature. Second, there appears to be both an earlier onset and more profound increase in maternal temperature than has been observed in prospective studies,3-1-3-3 which have investigated these phenomena. Typically in those studies, temperatures approaching 38°C are not even noted until ≥12 hours postepidural placement. In the current study, one wonders if patients in the epidural group were already experiencing upward trends in maternal temperature at epidural placement. Additionally, the study claims that labor length is not an important risk factor for maternal temperature despite the fact that only 8% of the nonepidural group had labors ≥12 hours, whereas 42% of the epidural group had labors ≥12 hours. Add to this the curious observation that approximately 50% of patients in the nonepidural group received active labor management, in contrast to the approximately 30% managed with this protocol in the epidural group. Clearly, there is a bias toward prolonged labor in patients receiving epidural analgesia (read-in patients with elevated temperatures). Additionally, it is reasonable to believe that prolonged labor would result in both a greater number and frequency of cervical exams, a potential risk factor for maternal temperature that was not even considered.Epidural analgesia is also implicated as a major factor in excess neonatal morbidity measured as neonatal sepsis workups and antibiotic therapy. Although it may be correct to describe the above factors as morbidity, the concerning measure of morbidity—documented neonatal sepsis—is nearly identical in both the epidural and nonepidural groups (0.3% and 0.2%, respectively). This is not a surprise as at least one prospective study showed that maternal temperature increases after epidural analgesia were not associated with bacteriologic signs of infection in either the parturient or infant.3-1 What is surprising is the large excess of neonatal sepsis workups and antibiosis in both groups of parturients (especially notable in the epidural group). This would seem to indicate that the current criterion guiding these workups (a maternal temperature ≥38°C) is grossly inadequate to the task. Fortunately, a review of current obstetrical and neonatology literature indicates that this shortcoming is realized, and work continues to identify better diagnostic criteria for both chorioamnionitis and neonatal sepsis. Finally, although this study focused on infectious morbidity, it would have been interesting for the authors to note the noninfectious morbidity associated with delivery in the nonepidural group (eg, administration of naloxone and/or intubation secondary to neonatal narcosis).Epidural analgesia provides the best pain control for the parturient during labor. Apart from maternal pain control, there are other benefits that accrue to the mother and fetus secondary to epidural analgesia. These include improved blood flow to the fetus especially in preeclampsia,3-5 decreased metabolic demands on the mother,3-6 decreased circulating levels of stress hormones,3-7 and decreased fetal metabolic acidosis.3-8 Additionally, the presence of a preexisting epidural catheter is an important factor in reducing maternal mortality during cesarean section.3-9 It would be a shame if all these benefits were to be secondary to on an effect of the Editor. In the of Lieberman et that in their study population maternal intrapartum fevers, neonatal sepsis evaluations, and neonatal antibiotic are by epidural analgesia during labor. The also that these of epidural analgesia should be discussed with patients and their care when the about the method of pain management during have that epidural analgesia during labor is associated with an increase in maternal temperature approximately to In no study were women randomized to an group group. to this it be maternal temperature are by epidural analgesia or if women with associated with a rise in maternal temperature are more likely to epidural analgesia. As an women with are at risk for of intrapartum but may have more labors and subsequently may be more likely to epidural analgesia. who epidural analgesia may have a different course of labor. Dr study was not the and groups may of women with different labor The epidural group may have had a different of patients with a from rupture of to fetal or other factors that were not controlled study not were used for epidural analgesia. Epidural analgesia is not a but a of including and of as as different and of all or some of the of epidural analgesia are only in the aspect that they are all given via the epidural only the of administration of epidural may have different maternal and fetal than a epidural. to study the of epidural analgesia, of the and is of authors also from their data that epidural analgesia increases the risk for sepsis evaluations and antibiotic therapy in There is no of criteria the need for a neonatal sepsis other than a maternal temperature of because less than one third of for sepsis had mothers with a temperature criteria other than elevated maternal temperature for the of these notable there is no of length of of ruptured or of delivery between the epidural and factors that can increase the rate of neonatal sepsis Although 25% of all the in the study were for the incidence of sepsis was to no noted between the epidural group and the group). our a maternal temperature of not a neonatal for sepsis. Perhaps the to from the study is that a maternal temperature of is a of neonatal sepsis and should not be a criterion for a neonatal sepsis that some of epidural analgesia in labor may increase maternal temperature, the relationship between epidural use and maternal fever concluded by the authors is not by the data related in this It can only be from the presented that some of epidural analgesia in labor may be a risk for increased maternal temperature. Although some of epidural analgesia may maternal temperature and other factors that would neonatal sepsis workups, a randomized study would need to be performed to show such an We the for their regarding our and we the to to their We and for their experience related to fever associated with the use of epidural analgesia. the rate of fever to be 2.5% without epidural compared with 7.5% with epidural. The increase in risk they note may be attributable to the fact that they all women our study population was to women who tend to have As indicate that maternal temperature to in the presence of labors would be associated with an increased risk of number of the et et and the of our results because our study was not Although randomized are an important of can and results when are of the with were randomized these results are as agree that between women who and not epidural analgesia and these in Table 1 of our In our study, we standard techniques to control for those potential were from our study In our we controlled for rupture of the and use of active management of labor. The remained even after these et al also concern that the delivery method might the of sepsis at our method of delivery is not one of the criteria for of a neonatal sepsis and also note the in epidural The protocol for epidural use in this population has been in and comment that our study claims that length of labor is not an important risk factor for maternal We made no such Figure 1 in our article clearly that length of labor is associated with the rate of fever in the epidural group but that the rate of fever in the group low of length of labor. It is also clear from Figure 1 that women receiving an epidural had a rate of intrapartum fever of length of labor. The we did not control for length of labor the of length of ruptured in our was that we believe there is to that epidural analgesia may be for the labors in the epidural group. remained associated with epidural use even after for length of labor and or for length of ruptured noted by et al and and some of the women in our study had their labor managed under a protocol of active management of labor. concern is by et al that the use of active management be associated with an increased risk of sepsis evaluations of maternal In women in the active management of labor group who received an epidural had a lower rate of intrapartum fever for the active management, in the and a lower rate of sepsis active management, The association of epidural use with fever, sepsis and antibiotic treatment was present in both but greater with also a number of factors as for the we and that in our study the temperature elevation associated with epidural use appears to be of earlier onset than in prospective studies, and with other et and that the women epidural may have already had elevated we did not report data related to the of fever onset in our We also did not report data related to temperature at because those data were not We have temperature for the women with the onset of labor and found no in baseline temperature for women who received an epidural compared with women who did not an epidural In addition, for women who received an epidural, the temperature before epidural administration was also blood were not for women who received an et al and and that an increased number of vaginal examinations may be an important factor in the rate of fever we report among women receiving epidural. As women with epidural analgesia had it would be that they would on more vaginal examinations than women who did not epidural analgesia. when the number of vaginal examinations is to the epidural use associated with intrapartum fever, sepsis and antibiotic use. The most mechanism by which an increased number of vaginal examinations would cause fever would be an increase in As noted in our article, it is difficult during labor to intrapartum fever is of infectious because are not Although there was no of an increase in sepsis among the of women receiving epidural, our study out the that women receiving epidural did more during labor related to their and that an increase in neonatal sepsis was by treatment of the mother with antibiotics during et al that may also impair thermoregulation. We found no for this in our data. receiving had the risk of fever as women not receiving such in both the epidural and epidural also that our be related to diurnal in temperature. To evaluate these we performed a for of and the use of analgesia to factors controlled in the original The odds for the association of epidural with fever, sepsis and antibiotic use were from those we of the of environmental temperature was also raised and et the prospective in the anesthesia literature a rise in temperature with epidural we not believe that in environmental temperature the temperature increase we One of those prospective was at the as the current study and the environmental in labor and delivery as labor at and are without or to the and approximately and room temperature is at to other factors noted to be in our the use of and tepid sponging et are not potential of the temperature increase but to the temperature As noted in our article, we also believe that it would be to study these as of temperature for our raised by the to the more raised by et ie, is it likely that the we observed can be to the use of epidural As noted above, a number of prospective have an increase in maternal temperature with epidural and in our study the association remained even after for multiple We of no other factor that be for the of the association we We believe that the association of epidural with fever we observed is most likely As this study in the literature to an increase in we that the most likely explanation is thermoregulatory epidural also that a large number of sepsis evaluations were performed in the infants of women whose temperature did not et and In the article, we discussed reasons for the increased rate of sepsis in that group. As we it is likely that these women had other factors as a fever not and rupture of that the criteria used by at our to which infants are for sepsis. As in the article, more were not presented as we did not have temperature at that as we have all of the records of study women who had the onset of labor, we be to more the reasons for that increase in potential benefits of epidural use were and note the potential increase in noninfectious morbidity in the specifically neonatal secondary to the use of In our study population, naloxone use was one in the epidural group). of infants whose mothers had an epidural compared with of infants of mothers without an epidural required and or more agree that the criteria for neonatal sepsis should be a we made in the We also agree that epidural analgesia is the most effective of pain relief currently and that women should be to epidural analgesia for pain relief during labor. We also that women should be fully informed about the risks and benefits of epidural analgesia before their The informed should include about the association of epidural use with rise in maternal temperature has been in the anesthesia as as related to the need for general anesthesia in the of an delivery by et al to be an of epidural Although in our population general anesthesia was used somewhat less often in the presence of an epidural, use was with or without an epidural in without epidural, with rate believe that women should be to epidural analgesia for pain relief during labor. Because epidural is an the about use for when women are fully informed about the can they the potential risks and benefits to the that is for