Cesarean Delivery: A Randomized Trial of Epidural versus Patient-controlled Meperidine Analgesia during Labor

The study protocol was developed by investigators from the Departments of Anesthesiology and Obstetrics and Gynecology and was approved by the Institutional Review Board of the University of Texas Southwestern Medical Center at Dallas. Healthy parturients with a singleton cephalic gestation at full term in spontaneous active labor were offered the chance to participate in this randomized investigation. Labor was diagnosed by the presence of regular uterine contractions associated with cervical effacement and cervical dilatation of 3-5 cm. All pregnancies were managed by certified nurse midwives under the direct supervision of obstetrical faculty and house officers, according to protocols established by our medical staff.

The study commenced on June 1, 1995 and ended on February 28, 1996. Participation was offered to eligible women by our nurse practitioner staff in the triage area adjacent to the labor and delivery areas. In obtaining informed consent, patients were informed of the random allocation of either epidural analgesia or PCIA. Although not encouraged, patients were aware that should the allocated analgesia fail to provide adequate pain relief, switching to an alternative type of analgesia would be allowed. Women giving written consent were then randomly assigned by the nurse practitioner staff in the triage area to receive either epidural analgesia or PCIA with meperidine during labor using numbered, sealed opaque envelopes. The randomization sequence was computer derived in blocks of 20 patients. After allocation, preprinted admission orders accompanied the patients to the labor and delivery unit. These orders specified either an anesthesia consultation for epidural analgesia or meperidine PCIA at the first report of pain. To minimize the potential for post-randomization selection bias, the anesthesiologist offered epidural analgesia to all women randomized to that group and did not consult with the PCIA group.

All staff followed procedures recorded in a written manual that prescribed the intrapartum management of nulliparous and multiparous women. Routine intrapartum management of all women included intravenous fluid administration and periodic auscultation with Doppler or continuous electronic fetal heart rate surveillance. Internal electronic fetal heart rate monitoring was used in those women with meconium-stained amnionic fluid, known fetal heart rate decelerations, or inadequate progress of labor. Our labor management approach encourages amniotomy in active labor when the fetal head is applied to the cervix. Pelvic examinations were performed approximately every 2 h to evaluate the progress of labor.

Cervical change of less than 1 cm/h coincidental with a hypotonic contraction pattern measured using intrauterine pressure transducers resulted in oxytocin augmentation of labor. Oxytocin was administered per written protocol, which was described previously. [6]Briefly, oxytocin starting at 6 mU/min was increased by 6 mU/min at 40-min intervals, to a maximum of 42 mU/min. Uterine activity of 200-250 Montevideo units for 2-4 h was considered adequate. Dystocia was diagnosed when adequate uterine activity did not result in progressive cervical dilation or descent of the fetus's head. Indications for the use of low forceps were limited to inadequate voluntary pushing or fetal heart rate abnormalities. Inadequate voluntary pushing was determined at the bedside, and if good descent was observed when effort was made (and, conversely, if no descent was observed when little effort was made) then lack of descent due to inadequate maternal expulsive efforts was diagnosed. Umbilical artery blood was obtained at all births from a doubly clamped cord segment for the analysis of blood gases.

Women randomized to epidural analgesia received an intravenous bolus of 500 ml Ringers lactate, after which epidural analgesia was initiated according to a written protocol using an indwelling catheter inserted into the lumbar epidural space using a 17-gauge Tuohy needle. Analgesia was achieved with 3-ml increments of 0.25% bupivacaine to a bilateral T-10 sensory level after a negative test dose with 0.25% bupivacaine. A continuous epidural infusion of 0.125% bupivacaine with 2 micro gram/ml fentanyl followed at an initial infusion rate of 8-10 ml/h, which was thereafter titrated to maintain a T-10 sensory level. This was maintained throughout the first stage of labor. If progress during the second stage of labor was inadequate after 1 h, the infusion was halved or discontinued to restore maternal expulsive efforts. Additional boluses of fentanyl, bupivacaine, or both were injected to overcome inadequate analgesia. If required, the epidural catheter was replaced. Left uterine displacement was maintained to avoid aortocaval compression.

Continuous external electronic fetal heart rate monitoring was used for 30 min after the test dose and then discontinued if the fetal heart rate pattern was normal. Maternal blood pressure was recorded every 5 min for 30 min and then every 30 min until delivery. Intravenous fluids were given to treat hypotension defined as a systolic blood pressure < 25% of the baseline or a systolic blood pressure < 100 mmHg. Persistent hypo-tension was treated with 5 mg ephedrine given intravenously as needed.

Women randomized to PCIA received 50 mg meperidine with 25 mg promethazine hydrochloride intravenously as an initial bolus, after which a Abbott-Lifecare 4100 patient-controlled-pump (Abbott Laboratories, North Chicago, IL) was set up to deliver 10 mg meperidine every 10 min as needed during the first hour and 15 mg every 10 min thereafter as needed until delivery. Additional 25-mg doses of meperidine were given on request not to exceed 100 mg in 2 h. In the event that pain relief was inadequate despite these measures, epidural analgesia was administered on patient request. Maternal blood pressure was recorded as already described in women who received epidural analgesia.

Pain was assessed with a linear 10-cm visual analog scale [7](0 = no pain, 10 = worst possible pain) before the initiation of analgesia, at complete cervical dilatation, and during the second stage of labor. In addition, the quality of pain relief during the first and second stage of labor was reassessed within 24 h after delivery using a five-point descriptive scale of excellent, very good, good, fair, or poor. The women were also asked if they would choose the same analgesia for a future labor and delivery. Nausea/vomiting (0 = no nausea/vomiting; 1 = nausea/vomiting not requiring treatment; 2 = nausea/vomiting requiring treatment) and the degree of sedation (0 = awake and alert; 1 = awake but drowsy; 2 = asleep but easily arousable; 3 = asleep but difficult to arouse; 4 = unresponsive) were also noted with each type of analgesia studied.

Obstetric data and other additional information were abstracted from maternal and neonatal charts and assessed for completeness and consistency before electronic storage in an on-line perinatal data system.

We performed an ad hoc power analysis based on our previous investigation, [5]in which the rate of cesarean delivery was increased from 4-9% when epidural analgesia was used. For 80% power with a one-tailed significance level of 0.05, 290 patients needed to be enrolled in each arm of this study to determine the same difference.

All other tests of significance were performed using two-tailed tests. The data were analyzed using SAS statistical software (SAS Institute, Cary, NC). Statistical significance (P < 0.05) was determined using an unpaired Student's t test, Fisher's exact test, and the Mann-Whitney U test as indicated. Data were analyzed according to the group assignment of randomization regardless of the eventual analgesia obtained (intention to treat). Secondarily, those patients compliant with their randomization assignment were compared ignoring those patients not receiving the prescribed analgesia of randomization (protocol compliant).

A total of 715 women were randomized in this investigation (Figure 1). Of 358 women who were allocated to receive epidural analgesia, 243 women completed the study as allocated and 115 (32%) women did not comply with the epidural analgesia protocol. Of these 115 women, 78 progressed rapidly to delivery before epidural analgesia could be initiated, and 37 women refused epidural analgesia during anesthesia consultation. Of these 37 women, 29 received no analgesia, and 8 received intravenous meperidine boluses (Figure 1). Of 357 women who were allocated to receive PCIA, 259 completed the study as allocated, and 98 (28%) women did not comply with the PCIA protocol. Of these 98 women, 73 progressed rapidly to delivery and did not receive any analgesia, 20 refused any analgesia, and 5 women who received meperidine PCIA as randomized later crossed over to epidural analgesia due to inadequate pain relief (Figure 1).

There were no significant differences in demographic characteristics between the groups (Table 1). Amniotomy was performed during labor in approximately 70% of women in each of the groups.

Intention-to-treat analysis based on the original randomization showed no significant difference between the two groups in the incidence of cesarean deliveries (epidural analgesia, 4% [95% CI: 1.9-6.2%] versus PCIA 5% [95% CI: 2.6-7.2%]), either in nulliparous (epidural 5% [9 of 197] versus PCIA 6% [11 of 189]) or parous (epidural 3% [4 of 161] versus PCIA 3% [5/168]) women (Table 2). In addition, the administration of epidural analgesia did not increase the incidence of forceps deliveries (epidural, 7% [26 of 358] versus PCIA, 4% [15 of 357]).

As shown in Table 3, infant outcomes showed no evidence that the type of analgesia had any deleterious effect on neonatal Apgar scores or acid-base condition. Significantly more neonates required naloxone for depressed respiration with PCIA than with epidural analgesia (3.4% versus 0.8%, P < 0.05). Two neonates in the epidural analgesia group and three in the PCIA group had to be transferred to the neonatal intensive care unit. There were no neonatal deaths and no infants suffered seizures within 24 h of birth.

Because one third of the randomized patients did not comply with the analgesia protocol, we separately compared those patients who did comply with their allocated analgesia protocol. There were no significant differences in demographic characteristics between the protocol-compliant groups (Table 4). The progress of labor and the method of delivery in the protocol-compliant study groups are summarized in Table 5and Table 6, respectively. Administration of epidural analgesia prolonged the first stage of labor and increased the incidence of oxytocin augmentation. In addition, fever developed in more women during epidural analgesia. The rate of spontaneous delivery was similar in both groups. Similar to what was observed with the intention-to-treat analysis, the analysis of the protocol-compliant groups also showed no significant difference between the two groups in the rate of cesarean deliveries (epidural analgesia, 5% [95% CI: 2.6-8.5%] versus PCIA, 6% [95% CI: 3-8.9%]), either in nulliparous (epidural, 5% [8 of 154] versus PCIA, 7% [11 of 148]) or parous (epidural, 5% [4 of 89] versus PCIA, 3% [3 of 111]) women. None of the five women initially given PCIA who crossed over to epidural analgesia had a cesarean delivery. Low forceps deliveries were more frequent in women who received epidural analgesia (epidural analgesia, 7% versus PCIA, 3%; P < 0.05). However, the incidence of outlet forceps was similar in both groups. The results of the comparison of neonatal outcome in the protocol-compliant groups were similar to the results of neonatal outcome in the intention-to-treat groups (Table 7).

Women who received epidural analgesia had a significantly higher incidence of hypotension compared with women who received PCIA (31% versus 0; P < 0.0001). There was no significant difference in the incidence of nausea and vomiting between the two groups (epidural analgesia, 7% versus PCIA, 4%). Sedation scores were significantly higher with PCIA compared with epidural analgesia (P < 0.0001). Specifically, the median sedation score with epidural analgesia was 0, (0, 0; first and third quartiles, respectively). The median sedation score with PCIA was 1 (0, 2; first and third quartiles, respectively).

As Figure 2shows, although the preanalgesic visual analog pain scale scores were similar for the two groups, women who received epidural analgesia reported lower pain scores during the first and second stage of labor compared with women who received PCIA. When parturients were queried within 24 h of delivery, 90% of women who received epidural analgesia rated their satisfaction as excellent, very good, or good compared with 65% of women who received PCIA (P < 0.001). In the epidural analgesia group, 95% of women, and in the meperidine PCIA group, 70% of women (P < 0.001), expressed their desire for the same type of analgesia during a future delivery. With epidural analgesia, the mean bupivacaine and fentanyl doses were 82 +/- 47 mg (SD) and 180 +/- 103 micro gram, respectively. With meperidine PCIA, the mean and maximum meperidine doses were 139 +/- 100 g (SD) and 500 mg, respectively. Sixty-one women (24%) received more than 200 mg meperidine in the PCIA group.

The primary finding in this investigation, in which crossovers to epidural analgesia were negligible (2%), was that cesarean deliveries were not increased as a result of epidural analgesia during labor. However, the first stage of labor was prolonged (approximately 1 h), and there was a twofold increase in oxytocin augmentation and low forceps delivery. The prolonged first stage of labor with epidural analgesia might have occurred due to decreased myometrial contractility secondary to neural blockade, or crystalloid infusion during labor. [8]Maternal fever in labor occurred more frequently in women who received epidural analgesia, presumably as a result of epidural analgesia, prolonged labor, or chorioamnionitis. [9]Clearly epidural analgesia has effects on the progress of labor, but this does not necessarily increase cesarean deliveries.

Few controlled trials have been designed specifically to determine whether epidural analgesia affects the incidence of cesarean deliveries, [10]and some of these investigations have methodologic flaws. Noble et al. [11]assessed obstetric outcomes in 245 women randomized to receive either epidural analgesia or intermittent boluses of a narcotic. This randomized study was inconclusive because many patients allocated to receive narcotics experienced such severe labor pain that they switched over to epidural analgesia. This was also a major flaw in our previous investigation that made it impossible for us to determine the effects of epidural analgesia without labor pain selection bias. [5]Philipsen and Jensen [12]randomized 111 women to receive either epidural analgesia or intramuscular meperidine during labor. The incidence of cesarean deliveries was not increased in women given epidural analgesia, although the number of women studied was too small to reach significance. The only other study of the effects of epidural analgesia on cesarean births was by Thorp et al. [2]and included only 93 patients, where epidural analgesia was reported to increase the cesarean rate from 2-25%. However, this unprecedented result has been challenged intensely. [4]Their study was also criticized for investigator bias, lack of guidelines for the management of labor, and inadequate information regarding anesthetic management.

Patient-controlled intravenous analgesia is widely used in the United States to manage postoperative pain, although use during childbirth has been limited. One concern is newborn respiratory depression from increased narcotic administration to the mother. Many women in our study used more than 200 mg of meperidine during the course of their labor, but only 5% of infants were given naloxone to reverse respiratory depression. The mothers were visibly sedated but were invariably arousable and none experienced respiratory depression. Importantly, nearly 65% of women reported good to excellent pain relief with meperidine PCIA, and 70% said they would use this form of analgesia during a future delivery.

Meperidine PCIA, which does not require the involvement of an anesthesiologist, is likely to be less expensive than epidural analgesia for obstetric patients. Such a difference in patient cost for pain relief during labor can have significant implications when extrapolated to the nation's obstetric population. Potential economic implications will need careful scrutiny in the prevailing managed care environment to ensure that the quality of analgesia is considered as well the cost. We found meperidine PCIA to be a safe and effective method of pain relief during labor. The economic advantage of such analgesia should not, however, be used to preempt a patients' decision to elect epidural analgesia when the latter is clearly a superior method of pain relief.

In this study, 290 patients needed to be enrolled in each arm to determine significant difference in the cesarean delivery rate. Approximately 350 women were randomized to each arm and the cesarean rate was unaffected when analyzed using the intention-to-treat approach and actual analgesia received. The limitation to our investigation, however, was the time that elapsed between randomization and initiation of analgesia, which resulted in deviation from the protocol in approximately one third of women giving consent who either progressed rapidly to delivery or declined the offer of labor analgesia. The randomization of women to either analgesia group at the time of initiation of analgesia would have reduced the rate of protocol failure in our study. However, the major advantage of our investigation in assessing the effects of epidural analgesia on cesarean rates was the elimination of crossovers after initiating the allocated analgesia, which allowed us to compare the effect of two unconfounded and randomized analgesic methods on labor outcome.

In conclusion, this study shows that labor epidural analgesia in women at full term with uncomplicated pregnancies and in spontaneous active labor is not associated with increased numbers of cesarean delivery. In addition, intravenous meperidine using a patient-controlled device is a suitable method of pain relief during labor.

The authors thank Jackie Wiley, R.N., for assistance in data collection, Donald McIntire, Ph.D., for assistance in statistical analysis, and all the nurses whose cooperation made this study possible.