Epidural anesthesia has been accepted as the preferred anesthetic technique for cesarean delivery in severely preeclamptic patients among both anesthesiologists (1) and obstetricians (2). Spinal anesthesia can be performed faster, has fewer complications, and is more cost-effective for uncomplicated cesarean delivery (3). However, concern has been raised that spinal anesthesia might be unsuitable for cesarean delivery in severely preeclamptic patients, as the potential for rapid and profound hypotension might further compromise an already potentially compromised newborn (4).
Several studies have investigated the hemodynamic effects of spinal anesthesia in severely preeclamptic patients. These studies include a descriptive study (5), comparative studies with spinal anesthesia in normal pregnancy (6,7), and a comparative study with general anesthesia (8). On the basis of these studies, it has been suggested that spinal anesthesia may be used in these patients to avoid the risks associated with general anesthesia in emergency situations and those arising from the use of larger epidural needles in patients with thrombocytopenia (9,10). Furthermore, minimal hemodynamic effects from spinal anesthesia in severely preeclamptic patients have been demonstrated when using a small spinal dose in a combined spinal epidural (CSE) technique (11).
Although a large retrospective study by Hood and Curry (12) and two small prospective studies by Wallace et al. (13) and Sharwood-Smith et al. (14) have shown that the hemodynamic effects of spinal anesthesia were similar to those seen with epidural anesthesia in severely preeclamptic patients, a large prospective analysis has not been reported. Therefore, a large prospective randomized multicenter study in severely preeclamptic patients was undertaken to evaluate the difference in hemodynamic changes between spinal and epidural anesthesia in severely preeclamptic patients.
After approval by the IRB of 5 participating tertiary care centers, 120 severely preeclamptic patients scheduled to have elective or urgent cesarean delivery under regional anesthesia were studied during the 2-yr period from November, 2000 to December, 2002. Severe preeclampsia was defined as a systolic arterial blood pressure (SAP) of 160 mm Hg or more or a diastolic arterial blood pressure (DAP) of 110 mm Hg or more and proteinuria of 100 mg/dL or more. Patients with coagulopathy, placental abruption, severe “fetal distress,” or a history of allergy to local anesthetics were excluded from the study. Intravenous magnesium sulfate (MgSO4) 4 g was given initially, followed by 1 g/h for seizure prophylaxis. Intravenous hydralazine 5 mg was given at 20-min intervals to decrease the DAP to approximately 90 mm Hg. Patients giving written informed consent to participate in this study were randomly assigned according to a random number table for block randomization (15), using numbered sealed envelopes to receive either epidural or spinal anesthesia. Preoperative fluid administration, which was limited to lactated Ringer’s solution 100 mL/h, was given on arrival in the labor room. In the operating room all patients received sodium citrate 30 mL orally and 5 L/min of 100% oxygen by facemask. The patients were monitored with standard monitoring devices including automated blood pressure cuff, electrocardiogram, and pulse oximetry.
Before administering regional anesthesia, 500 mL of colloid solution (6% hydroxyethyl starch in balanced salt solution) was given over 20 min (16,17), followed by lactated Ringer’s solution 100 mL/h. In the epidural group, a 17-gauge epidural needle was inserted at the L3-4 interspace with the patient in the lateral decubitus position, and an 18 gauge epidural catheter was inserted 3–4 cm into the epidural space. The patient was then placed supine with left uterine displacement and 5 L/min of oxygen was given via a facemask. Three mL of 2% lidocaine with epinephrine 1: 200,000 was given as a test dose, followed by 5 mL of 2% lidocaine with epinephrine 1:400,000 plus 50 μg of fentanyl. Incremental doses of 5 mL of 2% lidocaine with epinephrine 1:400,000 were given (total dose of 18–23 mL) until loss of sensation to pinprick to at least the T6 level was achieved. Mean arterial blood pressure (MAP) was measured every minute for the first 20 min, then every 2 min for the next 10 min, and every 5 min thereafter until the end of the surgery. Close monitoring of the MAP every minute during the first 20 min would allow us to detect the severity and duration of hypotension during this period.
Three milligrams of ephedrine was administered IV if the SAP decreased to 120 mm Hg but was more than 100 mm Hg, and 6 mg of ephedrine was given if the SAP decreased to or less than 100 mm Hg. To control the unblinded ephedrine administration, ephedrine was strictly given according to the requirements related to a change in SAP and was recorded along with the printed recorded SAP. The recorded numbers for ephedrine and SAP could be used as a cross-check for verification of the necessity of ephedrine administration. Immediately after delivery, blood from the umbilical artery was drawn from an isolated segment of the umbilical cord and immediately transported to the laboratory. After delivery, 3 mg of preservative-free morphine was administered via the epidural catheter. The patients were closely observed for all complications of severe preeclampsia and epidural morphine within the 24-h postoperative period.
In the spinal group, after fluid administration similar to that in the epidural group, a 27-gauge spinal needle was placed at the L3-4 interspace with the patient in the lateral decubitus position. After observing the flow of cerebrospinal fluid, 2.2 mL of 0.5% hyperbaric bupivacaine with 0.2 mg of preservative-free morphine was injected into the subarachnoid space. The patient was then turned supine with left uterine displacement. Measurement of the MAP in the operating room and postoperative observation was done in the same way as it was in the epidural group.
Demographic data, the highest MAP recorded before any medication administered in the labor room, the baseline MAP (the mean of two consecutive measurements taken 2 min apart) measured on arrival in the operating room, the volume of IV fluid administered and estimated blood loss were recorded. The pediatricians who examined the newborns and gave Apgar scores were unaware of the anesthetic technique used in the mothers. The laboratory technicians who analyzed the umbilical cord blood were also unaware of the anesthetic technique used in the patients.
A pilot study was performed in 10 severely preeclamptic patients undergoing cesarean delivery under epidural or spinal anesthesia. The purpose was to determine the necessary descriptive statistics for sample size estimation using the lowest MAP (primary outcome variable), which was unavailable from previous studies. These 10 patients were not included in the main study.
We hypothesized that the lowest MAP would have to be at least 10 mm Hg less in the spinal group than in the epidural group to be clinically significant and result in adverse neonatal effect. The statistical null hypothesis of no difference in the lowest MAP between the 2 groups was tested against the alternative hypothesis of 10 mm Hg difference. Using a 2-sided type I error of 0.05, 80% power, and the standard deviation of the lowest MAP of 17.5 (from the pilot study), the required sample size was 51 per group. We decided to study 120 patients (24 patients in each of 5 centers). By using a block randomization (15), we acquired 10 sets of 12 numbers for patients to be equally enrolled in the two groups of each set
Data were presented as mean (sd), median or number (percentage) as appropriate. χ2 test or Fisher’s exact test was used to test the difference in qualitative variables between the two groups. An unpaired Student’s t-test was used for normally distributed data whereas the Mann-Whitney U-test was used for non-normally distributed data. To test the difference between the mean SAP and DAP over time (0 to 30 min) between the two groups, an unpaired Student’s t-test was used . Type I error was set at 0.05 without adjustment for multiple hypothesis testing because these comparisons over time were only for the purpose of exploratory analysis. All statistical data analysis was performed using SPSS version 11.5 (SPSS, Chicago, IL).
Five patients in the epidural group were excluded from the study as a result of inadequate anesthesia and the need to proceed to general anesthesia. Eight of 55 patients in the epidural group and seven of 60 patients in the spinal group were in labor; the remaining 47 and 53 patients, respectively, were not in labor. Because laboring patients may have less hypotension during regional anesthesia as a result of periodic augmentation of circulating blood volume during uterine contractions (18), statistical analysis was performed only in nonlaboring patients.
Demographic data, the volume of IV fluid administered, the proportion of patients receiving MgSO4 or hydralazine therapy, and the volume of estimated blood loss were similar in both groups. Anesthesia duration was longer in the epidural group than in the spinal group (P < 0.001), although the surgical duration time, skin incision to delivery time, and the uterine incision to delivery time were similar in both groups. Median sensory blocked levels at the time of incision were also similar, with one level difference (T5 and T4 in epidural and spinal groups, respectively) (Table 1).
As illustrated in Table 2, the mean highest SAP, DAP, and MAP values measured during the preoperative period were similar in both groups. The mean lowest SAP, DAP, and MAP measured during the induction to delivery period were consistently lower in the spinal group than in the epidural group (mean difference, 14, 9, and 10; 95% confidence intervals, 6–21, 2–15, and 4–17 mm Hg for SAP, DAP, and MAP, respectively). The mean lowest SAP, DAP, and MAP measured after delivery to the end of the operation were similar in both groups. Although the incidence of significant hypotension (SAP ≤100 mm Hg) was about 2 times (51% versus 23%) more frequent in the spinal group than in the epidural group, the duration of hypotension (based on blood pressure measurement every 1 min) was short (median, 1 versus 0 min) in both spinal and epidural groups, respectively. One patient in each group had SAP ≤80 mm Hg for only 1 min. Three newborns from mothers with the maximum duration of hypotension at each level (≤80, 80–100, and 101–120 mm Hg) (Table 2) in the spinal group had 5-min Apgar scores of 8, 10, and 10 as opposed to 10, 10, and 10 in the epidural group and umbilical arterial blood pH of 7.26, 7.20, and 7.27 versus 7.38, 7.38, and 7.13 in the epidural group, respectively.
As illustrated in (Figure 1, there were significant differences in SAP at 1 to 15 min (P < 0.0001) and at 16 to 20 min (P < 0.005) and DAP at 1 to 15 min (P < 0.0001) and at 16 to 20 min (P < 0.01) between the 2 groups. There were no significant differences in SAP and DAP at 22 to 30 min between groups
Ephedrine was used to treat hypotension before delivery more frequently in the spinal group than in the epidural group (72% versus 45%, P = 0.006). The amount of predelivery ephedrine and total ephedrine were also larger in the spinal group than in the epidural group (Table 2).
During the preoperative period, patients who needed hydralazine therapy had a higher MAP than did the patients who did not need hydralazine therapy (P = 0.0014 and 0.0002, respectively). After the regional block, the lowest MAP during the induction to delivery period was significantly less in the spinal group than in the epidural group, regardless of hydralazine therapy (Table 3).
There were 4 sets of twins and one fetal death with no neonatal data (hydrop fetalis) in the epidural group and 2 sets of twins and 1 set of triplets in the spinal group. To evaluate the effects of regional anesthesia on neonatal outcomes we analyzed only the outcomes of the first baby of the twins or triplets. Therefore, only 46 and 53 neonates in the epidural and spinal groups, respectively, were included in the analysis.
Newborn weight, Apgar scores, and neonatal intensive care unit admission were similar in both groups (Table 4). The proportion of newborns with 5-min Apgar scores ≤7 was 7% in the epidural group as compared with 2% in the spinal group (95% confidence interval, −3,13). The umbilical arterial blood pH, Pco2, Po2, base excess, and HCO3 were similar in both groups (Table 5). Approximately 18% of newborns in the spinal group had an umbilical arterial blood pH <7.20, as compared with 13% in the epidural group (95% confidence interval, −11, 21). Only 2 premature newborns with pH <7.20 (7.08 and 6.92) had 5-min Apgar scores <7; they were in the epidural group (patients number 3 and 4 in (Table 6).
This study shows that spinal anesthesia for cesarean delivery in severely preeclamptic patients causes slightly more hypotension than does epidural anesthesia during the induction to delivery period. The duration of hypotension, however, was short and there was no difference in neonatal status.
In the previously published prospective study by Wallace et al. (13) comparing general (n = 26), epidural (n = 27), and CSE (n = 27) anesthesia for cesarean delivery in severely preeclamptic patients, the mean lowest SAP and DAP values after CSE technique were similar, approximately 110 and 60 mm Hg, respectively, similar to the lowest SAP and DAP values in the spinal group (113 and 63 mm Hg) in our study. However, the mean lowest SAP and DAP in the epidural group of our study were higher than in the epidural group of their study (126 and 72 mm Hg versus 110 and 59 mm Hg, respectively). This may be explained by the finding that the sensory level achieved in our study tended to be higher in the spinal group (T4) than in the epidural group (T5). In our study, we used 2.2 mL of 0.5% hyperbaric bupivacaine (11 mg) for spinal anesthesia, which was within the dose range of 8–12 mg used in other studies (7,8). Using a lower spinal dose or using a CSE technique with small-dose bupivacaine and supplementing with local anesthetics via the epidural catheter or using more aggressive MAP management may have reduced the difference in hypotension.
A retrospective study by Hood and Curry (12) compared 103 severely preeclamptic patients having spinal anesthesia with 35 patients having epidural anesthesia for cesarean delivery. There was only a 13% decrease in the mean lowest MAP from the baseline MAP in both epidural and spinal groups compared with a 25% decrease in both groups in the Wallace et al. study and with a 23% (epidural) and 31% (spinal) decrease in each group in our study. This difference is probably attributable to differences in study design among the various studies.
Ephedrine was administered more often in the spinal group (72%) than in the epidural group (45%) in our study. This is in contradistinction to previous studies. Ephedrine use was similar in the Hood and Curry study (23 and 26%) and Wallace et al. study (22 and 30%) in the spinal and epidural groups, respectively. We treated hypotension as soon as the SAP decreased to 120 by administering 3 mg of ephedrine or giving 6 mg if it decreased to or less than 100 mm Hg. In contrast, Wallace et al. (13) administered ephedrine 5 mg IV when SAP reached 100 mm Hg. In a severely preeclamptic patient with some degree of compromised placental function, we felt that a SAP of 100 mm Hg might compromise the fetus. We did not find any massive hypertension in any patients given early treatment with ephedrine in this study.
By subgroup analysis, we found that patients who needed hydralazine had higher MAP than did patients who did not. However, the presence or absence of hydralazine therapy did not influence the subsequent decrease in MAP.
The failure rate of epidural block in our study (8%) was more frequent than the 4.7% rate in a previous report (19). However, this study had enough power (86% by post hoc analysis) to analyze 47 nonlaboring patients undergoing epidural anesthesia, as compared with 53 nonlaboring patients undergoing spinal anesthesia.
Neonatal outcomes assessed by Apgar scores and the umbilical artery blood gas analysis were similar for both groups. Because the duration time of SAP ≤100 mm Hg was short, the uteroplacental blood flow might not have been impaired in either group (5). However, with a wide range of 95% confidence interval differences in the incidence of newborns with 5-min Apgar score ≤7 (−3,13%) and in umbilical arterial pH <7.20 (−11,21%), a larger study with adequate power or a systematic review is needed to evaluate differences in neonatal outcomes.
In summary, although the incidence of hypotension and ephedrine requirement was slightly more frequent in the spinal group than in the epidural group, we found evidence that supports the use of spinal anesthesia in severely preeclamptic patients. First, the difference in mean lowest MAP (mean difference, 10 mm Hg; 95% confidence interval, 4–17 mm Hg) did not appear to be clinically significant. Second, the hypotension was easily treated and there was only a brief period of significant hypotension in either group. Third, the neonatal outcomes assessed by the Apgar score and the umbilical arterial blood gas analysis were similar in both groups. Fourth, all 6 newborns who were born with the maximum duration times of hypotension in both groups had 5-min Apgar scores and umbilical arterial blood pH within normal ranges.
The authors would like to thank Dr. Chulaluk Komoltri, DrPH, Division of Clinical Epidemiology, Faculty of Medicine, Siriraj Hospital, for being the consultant for statistics in this study and the Mahidol University Research Fund for providing support for this study.
1. Newsome LR, Bramwell RS, Curling PE. Severe preeclampsia: hemodynamic effects of lumbar epidural anesthesia. Anesth Analg 1986;65:31–6.
2. Hypertensive disorders in pregnancy. In: Cunningham FG, Gant NF, Leveno KJ, eds. Williams obstetrics, 21st ed. New York: McGraw-Hill, 2001:607–8.
3. Riley ET, Cohen SE, Macario A, et al. Spinal versus epidural anesthesia for cesarean section: a comparison of time efficiency, costs, charges and complications. Anesth Analg 1995;80:709–12.
4. Malinow AM. Spinal anesthesia in preeclamptic patients: “supportive” evidence. Anesthesiology 2000;92:622–3.
5. Karinen J, Rasanen J, Alahuhta S, et al. Maternal and uteroplacental haemodynamic state in pre-eclamptic patients during spinal anesthesia for caesarean section. Br J Anaesth 1996;76:616–20.
6. Van Bogaert LJ. Spinal block caesarean section in parturients with pregnancy-induced hypertension. East Afr Med J 1998;75:227–31.
7. Aya AG, Mangin R, Vialles N, et al. Patients with severe preeclampsia experience less hypotension during spinal anesthesia for elective cesarean delivery than healthy parturients: a prospective cohort comparison. Anesth Analg 2003;97:867–72.
8. Dyer RA, Els I, Farbas J, et al. Prospective, randomized trial comparing general with spinal anesthesia for cesarean delivery in preeclamptic patients with a nonreassuring fetal heart trace. Anesthesiology 2003;99:561–9.
9. Santos AC, Birnbach DJ. Spinal anesthesia in the parturient with severe preeclampsia: time for reconsideration. Anesth Analg 2003;97:621–2.
10. Santos AC. Spinal anesthesia in severely preeclamptic women: when is it safe? Anesthesiology 1999;90:1252–4.
11. Ramanathan J, Vaddadi AK, Arheart KL. Combined spinal and epidural anesthesia with low doses of intrathecal bupivacaine in women with severe preeclampsia: a preliminary report. Reg Anesth Pain Med 2001;26:46–51.
12. Hood DD, Curry R. Spinal versus epidural anesthesia for cesarean section in severely preeclamptic patients. Anesthesiology 1999;90:1276–82.
13. Wallace DH, Leveno KJ, Cunningham FG, et al. Randomized comparison of general and regional anesthesia for cesarean delivery in pregnancies complicated by severe preeclampsia. Obstet Gynecol 1995;86:193–9.
14. Sharwood-Smith G, Clark V, Watson E. Regional anaesthesia for caesarean section in severe preeclampsia: spinal anaesthesia is the preferred choice. Int J Obstet Anesth 1999;8:85–9.
15. Pocock SJ. Method of randomization. In: Pocock SJ, ed. Clinical trials: a practical approach. New York: John Wiley & Sons, 1994:66–89.
16. Wasserstrum N, Kirshon B, Willis RS, et al. Quantitative hemodynamic effects of acute volume expansion in severe preeclampsia. Obstet Gynaecol 1989;73:546–50.
17. Belfort M, Uys P, Dommisse J, Davey DA. Haemodynamic changes in gestational proteinuric hypertension: the effects of rapid volume expansion and vasodilator therapy. Br J Obstet Gynaecol 1989;96:634–41.
18. Brizgys RV, Dailey PA, Shnider SM, et al. The incidence and neonatal effects of maternal hypotension during epidural anesthesia for cesarean section. Anesthesiology 1987;67:782–6.
19. Gerig HJ, Kern F. Success and failure rate in peridural anesthesia: a 1 year study. Reg Anaesth 1985;8:25–32.