Studies suggest that administration of maintenance epidural solutions as programmed or automated intermittent boluses, rather than continuous infusions, results in lower bupivacaine consumption, decreased need for manual “top-up” boluses by the anesthesiologist, and higher patient satisfaction.1–6 In this technique, the epidural maintenance dose is administered as a bolus by the infusion pump at regular intervals instead of as a continuous infusion. However, the optimal combination of bolus volume and dosing interval has not been determined. At one end of the spectrum, a small volume and short bolus dose interval will likely behave similar to a continuous infusion. At the other end of the spectrum, a large volume and long bolus dose interval may lead to an increased incidence of breakthrough pain. The purpose of this randomized, double-blind trial was to determine how manipulation of the programmed intermittent time interval and volume influences total drug use, quality of analgesia, and patient satisfaction during maintenance of labor analgesia. We hypothesized that manipulation of the programmed intermittent bolus time interval and volume during the maintenance of epidural labor analgesia influences total drug use (primary outcome), quality of analgesia, and patient satisfaction.
The study was approved by the Northwestern University IRB. The study was registered (identifier: NCT00417027) at ClinicalTrials.gov on December 27, 2006. Healthy, term (≥37 weeks' gestation), nulliparous women in spontaneous labor or with spontaneous rupture of membranes who planned neuraxial labor analgesia were eligible to participate in the study. Exclusion criteria included systemic disease (e.g., diabetes mellitus, hypertension, preeclampsia), use of chronic analgesic medications, systemic opioid labor analgesia before the request for neuraxial labor analgesia, or cervical dilation <2 cm or >5 cm at time of initiation of combined spinal-epidural (CSE) analgesia. Subjects were excluded from data analysis after randomization if they delivered within 90 minutes of the intrathecal injection (assumption that analgesia for the first 90 minutes was provided by the intrathecal injection and epidural analgesia had a minimal role).
Eligible women were asked to participate shortly after admission to the Labor & Delivery Unit at Prentice Women's Hospital and informed, written consent was obtained. At the time of request for labor analgesia, the cervix was examined and a baseline visual analog scale (VAS) score for pain (100-mm unmarked line with the end points labeled “no pain” and “worst pain imaginable”) was determined. Labor analgesia was initiated using a CSE technique in the sitting position at the L3–4 or L2–3 interspace using a loss-of-resistance technique to identify the epidural space and a 27-gauge pencil-point spinal needle to puncture the dura-arachnoid and administer the spinal dose. After injection of intrathecal bupivacaine 1.25 mg and fentanyl 15 μg, a single-orifice epidural catheter was secured 4 to 5 cm in the epidural space. A test dose of lidocaine 45 mg with epinephrine 15 μg was administered through the epidural catheter, and the parturient was placed in the lateral position.
The VAS score for pain was determined 10 minutes after the intrathecal injection. If the VAS score was ≤10 mm, the parturient was randomized (by a computer-generated random number table) to 1 of 3 programmed intermittent epidural bolus analgesia maintenance techniques: 2.5 mL every 15 minutes (2.5/15), 5 mL every 30 minutes (5/30), or 10 mL every 60 minutes (10/60) (Table 1). Group assignments were sealed in sequentially numbered opaque envelopes that were opened by an anesthesia research nurse after successful initiation of CSE analgesia. The maintenance epidural solution for all 3 groups consisted of bupivacaine 0.625 mg/mL with fentanyl 1.95 μg/mL; the total programmed maintenance dose was 10 mL/h. The initial programmed intermittent bolus dose was initiated 30 minutes after the intrathecal injection in all groups (time zero).
Programmed intermittent epidural doses were administered via a Hospira Gemstar infusion pump (Hospira, Inc., Lake Forest, IL). The patient, anesthesiologist, and research nurse recording data were blinded to group assignment. The unblinded anesthesia research nurse programmed the pumps and initiated maintenance epidural analgesia according to group assignment. Because there is no commercial pump that can be programmed to administer intermittent boluses and patient-controlled boluses, 2 pumps were prepared for each subject with the same epidural solution. One pump was programmed to administer the programmed intermittent bolus at a rate of 300 mL/h at regular intervals. The second pump was programmed to administer patient-controlled epidural analgesia (PCEA) (bolus dose 5 mL delivered at 150 mL/h, lockout interval 10 minutes, basal infusion rate 0 mL/h, maximum dose 15 mL/h). A 3-way stopcock was connected to the epidural catheter connector, and the infusion tubing from both pumps was connected to the stopcock. The programmed intermittent bolus pump was covered with an opaque bag to maintain blinding.
The parturient was instructed to push the PCEA demand button whenever she felt uncomfortable. Therefore, the maximum volume per hour administered by the epidural pumps was 25 mL in all groups. If the patient continued to have breakthrough pain after 2 PCEA boluses, the anesthesiologist was called to evaluate the patient, and if appropriate, administer bupivacaine 1.25 mg/mL in 5-mL aliquots (total 5–15 mL) to reestablish analgesia (VAS score ≤10 mm). VAS scores for pain were determined every 120 minutes until delivery beginning 60 minutes after the intrathecal injection. A modified Bromage score was determined every 120 minutes during the first stage of labor (0 = no motor paralysis; 1 = inability to raise extended leg, but able to move knee and foot; 2 = inability to raise extended leg and to move knee, but able to move foot; 3 = inability to raise extended leg or to move knee and foot). Bilateral dermatomal sensory level to ice was determined at 15 minutes and 3 hours after the intrathecal injection. Sensory threshold to a rigid von Frey filament was determined bilaterally at baseline (immediately before the initiation of CSE analgesia), and 3 hours after the intrathecal injection at the T12, T10, T7, and T4 dermatomes (IITC Life Science, Inc., Woodland Hills, CA). The system uses a calibrated internal load cell to measure and display force (in grams) exerted against the tip of the probe. Pressure was applied with the rigid tip until the subject reported feeling the pressure. Epidural analgesia was discontinued shortly after delivery. Before discharge from the Labor & Delivery Unit, the parturient was asked to mark her overall satisfaction with labor analgesia using a 100-mm unmarked line with the left end labeled “not satisfied at all” and the right end labeled “extremely satisfied.”
Additional recorded data included maternal age, height, weight, cervical dilation at initiation of CSE analgesia, duration of labor (initiation of analgesia to delivery), maximum oxytocin infusion rate during labor, and mode of delivery. Analgesia data included maintenance epidural solution volume administered from the programmed intermittent bolus pump, time of first PCEA bolus request, number of PCEA bolus requests and delivered doses, PCEA bupivacaine dose, time to first request for manual rebolus, number of manual bolus doses, manual bolus bupivacaine dose, total bupivacaine dose, and total fentanyl dose.
The primary outcome was total bupivacaine consumption per hour of labor analgesia. In our previous study, the median bupivacaine consumption was 12.3 mg/h (95% confidence interval [CI], 10.5–14.0 mg/h) in the control group (continuous infusion with PCEA) and 10.5 mg/h (95% CI, 9.5–11.8 mg/h) in the programmed intermittent epidural bolus group.1 A 1-way design with 3 groups with 62 subjects per group achieves a power of 0.81 using the Kruskal-Wallis test with a target significance level of 0.05, assuming the null hypothesis and the alternative hypothesis that the step increase in programmed intermittent interval would result in an incremental decrease of 1.8 mg bupivacaine per hour of analgesia. The estimated within-group standard deviation was 6 mg, and the groups were assumed to be evenly spaced. The estimate was based on simulation using 2000 Monte Carlo samples from the null and alternative distributions. To account for dropouts, 192 subjects were randomized into 3 groups. The sample-size calculation was made using PASS 2008 version 8.0.13, release date January 14, 2010 (NCSS, LLC, Kaysville, UT).
Categorical data were compared among groups with the χ2 test. Ordinal and continuous data were compared among groups using the Kruskal-Wallis test. Post hoc comparisons were made with the Mann-Whitney U test. To account for multiple comparisons, a P value <0.025 was used to reject the null hypothesis. A linear mixed-effects model using the restricted maximum likelihood method to account for individual analgesic requirements related to labor characteristics was constructed to model each patient's overall bupivacaine consumption per hour. The area under the VAS pain score versus time curve (AUC) was calculated using the trapezoidal method to assess the impact of pain burden on bupivacaine consumption (GraphPad Prism version 5.03; GraphPad Software, Inc., La Jolla, CA). Cervical dilation, maximum oxytocin rate, labor duration, delivery mode, and VAS · time AUC were tested for their association with bupivacaine consumption using analysis of covariance. In the final model, the fixed effect was basal bupivacaine administration rate (6.25 mg/h) and the random effect was the VAS · time AUC. The effect of group was assessed using the F test and pairwise comparisons were made using the estimated marginal means and 95% confidence limits. Data were analyzed using PASW version 18.0.2, release date April 2, 2010 (SPSS, Inc., Chicago, IL) and NCSS version 7.1.18, release date June 29, 2009 (NCSS2007; NCSS, LLC).
Three hundred sixty-eight subjects were approached to participate in the study; 190 women consented to participate and were enrolled in the study (Fig. 1). There were no differences among the groups in subject characteristics, cervical dilation at initiation of analgesia, or baseline VAS pain scores (Table 2). All women had successful initiation of CSE analgesia and were randomized to one of the study groups. There were no differences in the duration of labor, maximum oxytocin dose, and mode of delivery among groups (Table 3).
Patients randomized to group 10/60 consumed less total bupivacaine as well as less bupivacaine adjusted for the basal bupivacaine administration rate and pain burden than women in groups 2.5/15 and 5/30 (Table 4). Median total bupivacaine per hour consumption was 0.9 mg (95% CI, 0.2–2.3 mg) less in group 10/60 compared with group 2.5/15 (P = 0.008) and 0.8 mg (−0.3 to 1.9 mg) less than group 5/30 (P = 0.18). The mean difference in bupivacaine consumption adjusting for basal administration and pain burden in group 10/60 was −1.5 mg (95% CI, −0.6 to −2.5 mg) (P = 0.002) and −1.2 mg (95% CI, −0.2 to −2.1 mg) (P = 0.02) compared with group 2.5/15 and group 5/30, respectively. The amount of fentanyl per hour of labor consumed was not different among groups. The VAS · time AUC, and the VAS score at delivery in subjects with vaginal delivery were not different among groups (Table 3); however, the AUC correlated with bupivacaine consumption per hour (ρ = 0.33, P < 0.005). There were no differences among groups in PCEA requests or administrations, number of manual rebolus doses, or time to first PCEA dose or manual bolus dose (Table 4).
The cephalad extent of sensory analgesia to ice did not differ among groups at 15 minutes or 3 hours, nor was there a difference in the pressure sensory threshold to the von Frey filament at the T12, T10, T7, and T4 dermatomes (Table 5).
The main finding of this study was that extending the programmed intermittent epidural bolus interval and volume from 15 minutes to 60 minutes, and 2.5 mL to 10 mL, respectively, decreased bupivacaine consumption without decreasing patient comfort or satisfaction. However, secondary outcome measures of analgesia quality, including PCEA requests, number of manual bolus doses, cumulative fentanyl dose, and time to first manual bolus were not significantly different among groups. The finding of decreased bupivacaine consumption with increasing programmed intermittent bolus interval was achieved as a result of the combined need for decreased PCEA and manual reboluses to treat breakthrough pain with the longer programmed bolus interval.
By study design, group 10/60 was either even or ahead of the other 2 groups in terms of delivered maintenance dose at all times (the first programmed bolus doses were given at the beginning of the hour). Therefore, the finding of less drug use in this group is not an artifact of the timing of the last bolus dose relative to delivery. Indeed, this protocol biases the groups with the longer bolus intervals compared with shorter intervals in terms of drug consumption. Visual inspection of the data suggests trends of decreased requirements for PCEA and manual rescue analgesia as the bolus time interval/volume increased, as well as a longer interval for request for manual rebolus, suggesting that the study is underpowered to find significant differences in these secondary outcomes.
Previous studies of programmed intermittent bolus technique compared continuous infusions (with and without patient-controlled boluses) with programmed intermittent epidural bolus alone (Table 6). Programmed intermittent epidural bolus volume and interval varied among studies; however, all studies except one7 concluded that this technique was superior to continuous infusion regimens. We have previously demonstrated that programmed intermittent epidural bolus (6 mL administered at 30-minute intervals) decreased bupivacaine consumption and need for manual reboluses, and improved patient satisfaction compared with continuous epidural infusion (12 mL/h).1 Other investigators found a decrease in hourly bupivacaine consumption,3,4,6 an increase in the time to first rescue bolus,2–4,6 lower number of manual rebolus doses,3,5 lower pain scores,2 reduced PCEA demands,4,6 and higher patient satisfaction scores5,6 in the programmed intermittent epidural bolus group compared with the continuous epidural infusion group. The single study in which the programmed intermittent bolus technique was not superior to continuous infusion, published in 2010, used an intermittent bolus volume/time interval of 2.5 mL every 15 minutes.7 To our knowledge, the programmed intermittent bolus technique has not been compared with a pure PCEA technique.
The results of these studies of the programmed intermittent bolus technique are similar to those of PCEA studies in which the patient-controlled bolus volume and interval were manipulated. In a systematic review, Halpern and Carvalho8 concluded that there is no ideal bolus dose or lockout interval setting for labor PCEA, although larger bolus doses of dilute anesthetic may provide superior analgesia and maternal satisfaction.
Although the programmed intermittent epidural bolus technique clearly provides equal or better analgesia with less drug, the mechanism of this finding has not been fully elucidated. Solutions injected into the epidural space may spread more evenly when injected as a bolus, as compared with a continuous infusion.9 Body position has very little influence on the spread of sensory analgesia after an epidural bolus injection,10 whereas prolonged maintenance of one position during a continuous epidural infusion (which is often observed in women with neuraxial labor analgesia) may influence the distribution of sensory blockade. In a study of thoracic epidural analgesia in postoperative gynecologic patients, patients randomized to the programmed intermittent technique had a greater number of blocked dermatomal segments than women who received the same hourly dose of ropivacaine via a continuous epidural infusion.11 These findings suggest that more extensive neuroblockade may explain the improved analgesia with the intermittent bolus technique.
The sensory blockade data in studies in laboring patients are inconsistent. Chua and Sia2 found a higher maximum cephalad sensory level to cold in the programmed intermittent bolus group, whereas later studies from the same institution were unable to reproduce this finding.4,6 Density of motor block was reported to be not different between the programmed intermittent bolus and continuous infusion groups in any study.1–6 Similarly, in the current study, we did not detect a difference in the extent and density of neuroblockade among study groups. Therefore, current clinical data do not support or refute greater spread of anesthetic solution by intermittent bolus as the mechanism of reduced analgesia consumption.
Another possible explanation for our findings is that the dura-arachnoid puncture made by the spinal needle as part of the CSE technique affected the translocation of anesthetic solution from the epidural to subarachnoid space to a greater degree in larger versus smaller bolus volumes. Additionally, the administration of anesthetic solution into the epidural space 30 minutes after the intrathecal injection may have resulted in an epidural volume extension effect (compression of the dural sac by the epidural injection resulting in extension of neuroblockade). However, the finding in our previous study of a more pronounced difference in bupivacaine consumption after longer compared with shorter labors,1 and the finding by others of improved analgesia with lower drug doses using a programmed intermittent bolus compared with continuous infusion technique after initiation of epidural analgesia,3 argue against these mechanisms contributing substantially to our findings. Additionally, an epidural volume extension effect is unlikely 30 minutes after the intrathecal injection.12
Studies of epidural opioid analgesia suggest that epidural bolus administration of lipid soluble opioids results in segmental spinal opioid analgesia, whereas continuous epidural opioid infusion primarily results in systemic opioid analgesia.13 The authors of one such study suggested that a fentanyl bolus results in a significantly larger amount of fentanyl in the epidural space compared with that which occurs at any single time point during an infusion.13 Thus, even though only a small fraction of the administered fentanyl is able to distribute to the spinal cord opioid receptors, in the case of epidural bolus administration, the fraction is sufficient to produce a spinally mediated analgesic effect. This phenomenon may contribute to the dose-sparing effect of the bupivacaine and fentanyl epidural solution that we observed in patients who received programmed intermittent boluses compared with continuous infusion maintenance labor analgesia.
Although we have shown that a larger volume administered at a longer interval results in less bupivacaine consumption than smaller volumes at shorter intervals, the difference in bupivacaine consumption is unlikely to be clinically relevant when using low-concentration solutions that result in minimal or no motor blockade. We did not observe a difference among groups in the incidence of motor blockade or rate of instrumental vaginal delivery. The optimal time interval/bolus dose regimen may depend on several factors including the duration of labor, the concentration and specific components of the epidural solution, and the rate of administration of the programmed bolus dose.
Rates of programmed bolus dose administration in previous studies varied between 75 and 400 mL/h.2–7 Epidural catheter tip design may also influence outcome because the infusion/bolus rate influences whether the epidural solution exits one or more orifices of a multiorifice catheter.14 In an in vitro study, the solution exited only the most proximal orifice when the infusion/bolus rate was ≤120 mL/h.14
Safety concerns may limit the bolus dose. There is inherent safety in administering potentially toxic drugs as an infusion rather than a bolus. The largest dose we chose to administer as a bolus was bupivacaine 6.25 mg because we believed it unwise to exceed this dose in patients who are intermittently monitored and might also self-administer a PCEA dose immediately before or after the programmed dose.
There are several limitations to our study design and conclusions. We did not include a group that received a continuous infusion or PCEA without a background infusion. The study was underpowered to identify a small difference in bupivacaine consumption between the 2.5/15 and 5/30 groups if there is such a difference, and to identify differences in the incidence of breakthrough pain. The methods of sensory and motor block assessment may not have been sensitive enough to detect a subtle difference in the extent or density of neuroblockade. Our conclusions are limited to our patient population, drug concentrations, and mode of initiation of labor analgesia. The programmed maintenance dose was low (the average requirement for labor analgesia when bupivacaine is administered with fentanyl is bupivacaine 10–12 mg1) and this resulted in a high rate of breakthrough pain requiring intervention by an anesthesiologist. We intentionally chose a low dose so that the rate of breakthrough pain would be measurable, and most women required only one manual bolus. However, given the high rate of breakthrough pain, this dose may be too low for routine clinical practice. Alternatively, a larger PCEA dose than that used in the current study may be necessary. Finally, current epidural pump technology supports continuous epidural infusion, PCEA without a background infusion, and PCEA with a background infusion, but not programmed intermittent epidural bolus regimens, with or without supplemental PCEA.
Previous studies have shown that the programmed intermittent bolus technique results in less bupivacaine consumption, less breakthrough pain, and improved patient satisfaction than continuous infusions or PCEA with a basal infusion.1–6 Our study confirms that analgesic consumption may depend on the timing and volume of intermittent boluses. Further study is required to determine whether this technique has a role in labor analgesia or postoperative epidural pain control, and to elucidate whether manipulation of the bolus volume and interval affect clinically important outcomes.
Name: Cynthia A. Wong, MD.
Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.
Attestation: Cynthia A. Wong has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.
Name: Robert J. McCarthy, PharmD.
Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.
Attestation: Robert J. McCarthy has seen the original study data, reviewed the analysis of the data, approved the final manuscript, and is the author responsible for archiving the study files.
Name: Bradley Hewlett, MD.
Contribution: This author helped design the study and conduct the study.
Attestation: Bradley Hewlett reviewed the analysis of the data and approved the final manuscript.
Cynthia A. Wong is Section Editor of Obstetric Anesthesiology for the Journal. This manuscript was handled by Steve Shafer, Editor-in-Chief, and Dr. Wong was not involved in any way with the editorial process or decision.