The choice of intrapartum pain relief in women with severe preeclampsia is controversial.1,2 Although retrospective studies have recommended both epidural analgesia and systemic opioids, to date, we are unaware of any published comparative trials in laboring women with severe preeclampsia. Epidural analgesia is considered by most obstetric anesthesiologists to be the preferred method of intrapartum pain relief in women with severe preeclampsia. In a retrospective study of 327 women with severe hypertensive disease who had labor, epidural analgesia did not increase the cesarean delivery rate.3 However, epidural analgesia can cause a maternal sympathetic blockade, resulting in hypotension, decreased uteroplacental perfusion, and nonreassuring fetal heart rate patterns.4 Hypotension may be treated by crystalloid volume expansion, but this can predispose the parturient with severe preeclampsia to pulmonary edema and hypoxemia.1 Conversely, epidural analgesia may lower maternal plasma concentrations of catecholamines,5 and thereby improve uterine intervillous blood flow.6,7
Systemic opioids are relatively easy to administer and are unlikely to cause maternal hypotension and nonreassuring fetal heart rate patterns. However, these agents readily cross the placenta and may cause neonatal respiratory depression. Systemic opioids may also exacerbate preexisting maternal respiratory depression from parenteral magnesium sulfate therapy.
To investigate the relationship between intrapartum analgesia and the cesarean delivery rate, we performed a randomized trial of two methods of analgesia in women with severe preeclampsia who requested pain relief in labor.
MATERIALS AND METHODS
This study was approved by the Institutional Review Board at the University of Alabama at Birmingham. Laboring women with a singleton pregnancy and vertex presentation, and who were at least 24 weeks' gestation and had a cervical dilatation of no more than 5 cm were eligible if they had a clinical diagnosis of severe preeclampsia. The clinical diagnosis of severe preeclampsia required at least one of the following four criteria: 1) systolic blood pressure greater than or equal to 160 mmHg or diastolic blood pressure greater than or equal to 110 mmHg, with proteinuria of either 2+ on dipstick or a protein/creatinine ratio of 1 or more; 2) systolic blood pressure greater than or equal to 140 mmHg or diastolic blood pressure greater than or equal to 90 mmHg with severe proteinuria (3+ or more on dipstick or 3.5 g or more per 24-hour collection) 3) systolic blood pressure greater than or equal to 140 mmHg or diastolic blood pressure greater than or equal to 90 mmHg with proteinuria of either 2+ on dipstick or a protein/creatinine ratio of 1 or more with an aspartate aminotransferase greater than 75 units/L and a platelet count of less than 100 × 109/L; or 4) eclampsia, without evidence of increased intracranial pressure or focal neurologic deficit. Women with a platelet count of less than 80 × 109/L or evidence of pulmonary edema were ineligible. In addition, women with a nonreassuring fetal heart rate tracing requiring imminent delivery or an abnormal airway examination that might predict an increased risk of difficult intubation were also excluded.
Eligible women were invited to participate, and written informed consent was obtained either before the onset of induced labor or during early labor, before the patient's request for pain relief. A screening log was not maintained. A computer-generated block randomization schedule was stratified according to gestational age less than 35 weeks' versus 35 weeks' or longer gestation, and group assignments were sealed in consecutively numbered, opaque envelopes. Consenting women were not assigned to their randomization group until they requested pain relief and the envelope was opened.
All women received intravenous crystalloid at a rate of 100 mL per hour and intravenous magnesium sulfate for seizure prophylaxis from the diagnosis of severe preeclampsia until 24 hours postpartum. Our standard magnesium sulfate regimen is a 4-g intravenous bolus given over 20 minutes, followed by a continuous infusion of 2 g per hour. The volume status of each woman was frequently assessed with attention to total intravenous fluids administered and urinary output. All women were monitored during the magnesium sulfate infusion with continuous pulse oximetry.
Group assignments were made when consenting women requested pain relief at greater than or equal to 2-cm and less than or equal to 6-cm cervical dilatation. Women requesting pain relief at less than 2-cm dilatation received either intravenous or intramuscular opioids until their labors had progressed. Women whose labors had progressed beyond 6-cm cervical dilatation when they requested pain relief did not receive their assigned treatment, but were analyzed in their assigned group. Intramuscular analgesia included meperidine hydrochloride 75–100 mg as needed every 3 hours. Alternatively, intravenous medications included meperidine hydrochloride 50–75 mg as needed every 1 to 2 hours or butorphanol tartrate, 1–2 mg as needed every 1 to 2 hours.
Women assigned to receive epidural analgesia (epidural group) were managed according to a standardized protocol. Each woman received 250–500 mL of lactated Ringer's solution over 20 minutes, before the administration of epidural analgesia. Using a sterile technique, an epidural catheter was placed at the L3-L4 interspace. A test dose of 3 mL of 0.25% bupivacaine was administered, and epidural analgesia was established with incremental bolus injections of 3–5 mL of 0.25% bupivacaine. The goal was to obtain a T-10 sensory level. After satisfactory analgesia was established, analgesia was maintained by a continuous epidural infusion of 0.125% bupivacaine with fentanyl, 2 μg/mL, at an initial rate of 10 mL per hour. Hypotension was treated by the anesthesiologist with either additional intravenous crystalloid or intravenous bolus doses (eg, 2.5–5.0 mg) of ephedrine. Reductions in blood pressure of less than 25% in the presence of a reassuring fetal heart rate tracing (as determined by the managing obstetrician) were not treated.
Women in the opioid group received intravenous meperidine hydrochloride via a patient-controlled analgesia device. The self-administered dose was 10 mg, with a lock-out interval of 10 minutes. A maximum dose of 240 mg could be administered in any 4-hour period. Intravenous promethazine hydrochloride, 25 mg every 6 hours, was also administered by the nursing staff as needed for control of nausea.
A visual analogue pain score8 was obtained and recorded hourly by the labor nurse. Scores could range from 0 to 10, and the arithmetic mean of all intrapartum scores was determined for each patient. Within 48 hours of delivery, a research nurse performed a verbal survey of the woman's overall assessment of pain relief during labor and delivery with four response levels (poor = 1, fair = 2, good = 3, excellent = 4).
After delivery, an obstetric research nurse recorded maternal demographic information, laboratory measurements, delivery data, and maternal complications. Neonatal data included birth weight, Apgar scores at 1 and 5 minutes, umbilical cord arterial blood gas results, and selected morbidities and mortality.
The primary outcome was the cesarean delivery rate for maternal or fetal indications. Maternal secondary outcomes included intrapartum pain scores, postpartum patient satisfaction, requirement for pharmacologic interventions to treat hypotension, and the incidence of serious complications including pulmonary edema, eclampsia, and acute renal dysfunction. Neonatal secondary outcomes included naloxone treatment, Apgar scores, umbilical cord arterial blood pH, neonatal intensive care unit admissions, and selected morbidities and mortality.
We estimated that in our population of laboring women with severe preeclampsia, the cesarean delivery rate would be 50%. To have 80% power to detect at least a 50% difference in the intergroup cesarean delivery rates (ie, 50% versus 25%), a sample of 116 women would be required (α = 0.05, two-tailed χ2 test). Statistical analyses were performed using SAS (Cary, NC). Proportional data were compared using χ2 or Fisher exact test, as determined by the expected cell size. Continuous data were compared with either the Student t test or Wilcoxon rank sum test (nonparametric) as determined by the Kolmogorov-Smirnov statistic, which establishes whether the data fit a normal (parametric) distribution. All intergroup comparisons were performed by intent-to-treat, and P < .05 was considered to represent statistical significance.
During the 42-month study period, 116 women were enrolled. Fifty-six were assigned to receive epidural analgesia, and 60 were assigned to receive intravenous opioid analgesia. Ten women did not receive the assigned treatment, three in the epidural group and seven in the opioid group (P = .23). Rapid labor was the most common event that precluded the assigned treatment (epidural, n = 3 versus opioid, n = 5). One woman assigned to the opioid group received epidural analgesia at the discretion of the attending anesthesiologist. Another woman who was assigned to opioids received epidural analgesia after experiencing severe nausea and vomiting after the opioids were administered and which persisted in spite of antiemetic therapy.
Selected maternal demographics were similar in the two groups (Table 1). Women with chronic hypertension (epidural, n = 9 versus opioid, n = 7, P = .51), pre-existing insulin-dependent diabetes mellitus (epidural, n = 1 versus opioid, n = 2, P = 1.0), and a self-reported history of severe preeclampsia (epidural, n = 5 versus opioid, n = 1, P = .11) were similarly distributed between the groups. The frequency of the syndrome of hemolysis, elevated liver enzymes, and low platelets was similar between the groups (epidural 14% versus opioid 22%, P = .30). Clinical diagnostic criteria and baseline laboratory measurements were similar between the two groups (Tables 2 and 3).
Before randomization, a similar number of women received parenteral opioids (epidural, n = 42 versus opioid, n = 41, P = .43) and hydralazine (epidural, n = 25 versus opioid, n = 19, P = .13). The majority of women (75%) enrolled in the study received either laminaria or extra-amniotic saline infusion for cervical ripening.9 The type of cervical ripening agents was evenly distributed between the groups (laminaria: epidural, n = 15 versus opioid, n = 12 and extra-amniotic saline infusion: epidural, n = 25 versus opioid, n = 35, P = .34).
The cesarean delivery rate in the entire study population was 15%. There was no difference in the overall cesarean delivery rate between the two groups (epidural 18% versus opioid 12%, P = .35). There was no significant difference in the indications for cesarean delivery between the two groups (P = .07, Table 4).
Five women (9%) in the epidural group received intravenous ephedrine for treatment of hypotension, whereas no woman in the opioid group required ephedrine (P = .02). Women in the opioid group were more likely to receive a local anesthetic at the time of vaginal delivery (epidural 4% versus opioid 27%, P = .001). Of those women who were delivered by cesarean, five received general endotracheal anesthesia (epidural, n = 2 versus opioid, n = 3), 11 received epidural anesthesia (epidural, n = 8 versus opioid, n = 3), and one patient in the opioid group received spinal anesthesia (P = .19).
Women assigned to the opioid group received a higher median dosage of intrapartum opioids than women in the epidural group (epidural 50 mg, range 0–100 mg versus opioid 160 mg, range 0–860 mg, P < .001). Mean intrapartum pain scores were lower in women in the epidural group (epidural 4 ± 3 versus opioid 7 ± 3, P < .001). In addition, women in the epidural group were more satisfied with their analgesia than women in the opioid group according to their median postpartum satisfaction scores (epidural 3, range 1–4 versus opioid 2, range 1–4, P = .002).
There were no maternal deaths, and maternal complications were rare. Pulmonary edema occurred in one woman in each group (P = 1.0), postpartum hemorrhage in four women in the epidural group and two women in the opioid group (P = .43), and endometritis in five women in each group (P = 1.0). We observed no cases of postpartum pneumonia, renal failure, or urinary tract infection.
Neonatal outcomes were similar between the groups (Table 5), and 1-minute Apgar scores were also similar (P = .6). Neonates of women who were assigned to the opioid group received naloxone more often at the time of delivery (P < .001). Three infants born to women randomized to the epidural group died, and no deaths occurred in the opioid group. One 480-g infant was delivered at 26 weeks' gestation by cesarean because of a nonreassuring fetal heart rate tracing. The umbilical cord arterial pH was 7.24, and the 5-minute Apgar score was 6. This neonate died on day of life 14 from necrotizing enterocolitis and a bowel perforation. The second neonatal death was a 599-g infant who was delivered by cesarean at 25 weeks' gestation because of a nonreassuring fetal heart rate tracing and died on day of life 10 from complications related to extreme prematurity: respiratory distress syndrome and pulmonary interstitial emphysema. The umbilical cord arterial pH and 5-minute Apgar score were 7.31 and 7, respectively. The third neonatal death occurred on day of life 10 in a 798-g, 28-week infant delivered by cesarean because of failed induction. The umbilical cord arterial pH and 5-minute Apgar score were also normal (7.25 and 7, respectively). On the day of death, the infant was noted to have severe abdominal distension and a probable ileus; however, no perforation was apparent on abdominal radiograph. The family declined autopsy, and the death was attributed clinically to bowel perforation.
To our knowledge, augmented by a MEDLINE search of the years 1966–2001 using the key words “preeclampsia,” “epidural analgesia,” and “clinical trials,” the trial reported herein is the first randomized comparison of epidural and intravenous opioid analgesia in a selected population of women with severe preeclampsia. Lucas et al10 reported a randomized trial of 738 women with pregnancy-induced hypertension who received epidural analgesia with bupivacaine versus intravenous meperidine. Although there was no difference in the cesarean delivery rate, the majority of their population did not have severe hypertensive disease. Similar to our findings, these investigators noted significantly greater requirements for ephedrine in women with severe preeclampsia who received epidural analgesia and reported no attributable adverse effects. Wallace et al evaluated the administration of epidural analgesia in women with severe preeclampsia who underwent cesarean delivery, and reported that clinically significant hypotension requiring intravenous ephedrine occurred in 30% of women who received epidural analgesia.11
A retrospective study of 200 women with pregnancy-associated hypertension demonstrated significantly higher rates of ominous fetal heart rate tracings in the hypertensive cohort as compared with a cohort of 2023 normotensive gravidas.4 This effect was more pronounced in the hypertensive group who received epidural analgesia. Compared with normotensive women with epidural analgesia, women with hypertension and epidural analgesia had a three-fold increased risk of ominous fetal heart rate tracings. In the cohort that did not receive epidural analgesia, there was a two-fold increased risk between the hypertensive and normotensive women. In the current study, epidural analgesia appeared to be a safe method of pain relief with low rates of maternal hypotension and maternal morbidities; however, our sample size was not large enough to confirm the relative safety of these two methods with regard to uncommon complications. The apparent safety of epidural analgesia in our population is likely a function of both our intrapartum management of women with severe preeclampsia as well as our standardized protocol for administration of epidural analgesia. In agreement with previous studies,12 epidural analgesia provided superior pain relief and had the additional advantage of facilitating surgical anesthesia in women who underwent a cesarean delivery.13
All three of the observed neonatal deaths occurred in infants born to women who received epidural analgesia; however, these infants were all extremely preterm (24, 26, and 28 weeks' gestation), and death occurred remote from delivery. Umbilical cord gas sampling at the time of delivery was performed in these three infants, and there was no evidence of intrapartum fetal compromise. This nonsignificant difference in neonatal deaths is likely due to chance and cannot be attributed to the method of analgesia.
The results of this study suggest that epidural analgesia administered to women with severe preeclampsia according to a standardized protocol at a tertiary care center is safe and effective. In women with severe preeclampsia, as defined in our investigation, intent to achieve vaginal delivery also appeared to be safe for both mother and fetus and resulted in an overall cesarean delivery rate of only 15% at a mean gestational age of 33 weeks. Because our pretrial estimates for cesarean delivery (50%) were higher than we observed, we cannot exclude the possibility of smaller, but still clinically significant incremental intergroup differences in the cesarean delivery rates. However, considering a baseline cesarean delivery rate of 12% (as observed in the opioid group), approximately 650 women would have to be studied to have 80% power to detect at least a 50% change in the cesarean rate. Such a study, although possible, would take many years to complete at a single center and, based on our findings, we believe that it would be unlikely to detect clinically significant differences in the rates of serious maternal or neonatal complications. Therefore, we now endorse the use of epidural analgesia administered according to a standardized protocol for the relief of labor pain in women with severe preeclampsia at our institution.