ED⁵⁰and ED⁹⁵of Intrathecal Bupivacaine in Morbidly Obese Patients Undergoing Cesarean Delivery

• Whether obese parturients require less spinal local anesthetic for anesthesia for cesarean delivery is unknown

• Using a randomized dose response of spinal bupivacaine with fentanyl and morphine, the median dose of bupivacaine for successful anesthesia in morbidly obese parturients was 9.8 mg, similar to a previous study in nonobese parturients

CESAREAN delivery is most commonly performed under spinal anesthesia using hyperbaric bupivacaine.1Bupivacaine provides an appropriate duration of anesthesia to perform cesarean delivery. Hyperbaric and hypobaric bupivacaine both provide effective spinal anesthesia for cesarean delivery,2–4although with smaller intrathecal doses, hyperbaric solutions may ensure a more predictable block.5Many investigators recommend a reduced intrathecal dose in morbid obesity due to an observed lower neuraxial local anesthetic requirement.6–8This is supported by Hogan et al .,9who reported lower cerebrospinal fluid volumes in patients with a high body mass index (BMI). Cerebrospinal fluid volumes can be reduced by several mechanisms that exert external pressure on the dural sac: increased intraabdominal pressure from abdominal fat; epidural venous plexus engorgement secondary to compression of the inferior vena cava and diversion of venous return; and the inward movement of soft tissue (especially fat) in the intervertebral foramen displacing cerebrospinal fluid.9,10 

Although successful spinal anesthesia has been reported with a low dose of 5 mg hyperbaric bupivacaine in a morbidly obese patient,11the limited studies that have investigated the relationship between weight and the spread of intrathecal local anesthetics have demonstrated conflicting results.12–19In women undergoing cesarean delivery, Norris et al .15,16showed no correlation between the height of the sensory block and the patient's BMI. However, morbidly obese patients were not specifically studied, and fixed doses of hyperbaric bupivacaine were used.

The optimal dose of intrathecal bupivacaine in the morbidly obese population undergoing cesarean delivery with spinal anesthesia is unknown. Previous studies have determined the optimal dose of hyperbaric bupivacaine for cesarean delivery in patients of normal body habitus.20,21The aim of this study was to use logistic regression to determine the ED⁵⁰and ED⁹⁵values of intrathecal hyperbaric bupivacaine when coadministered with intrathecal fentanyl (10 μg) and morphine (200 μg) in morbidly obese patients (BMI ≥40). We hypothesized that the dose requirements would be lower than obtained for hyperbaric bupivacaine in nonobese patients when using the same logistic regression methodology under similar study conditions.20 

We used a prospective, randomized, double-blinded, dose-ranging study to determine the ED⁵⁰and ED⁹⁵values of intrathecal hyperbaric bupivacaine for cesarean delivery. Forty-two healthy term parturients presenting for elective cesarean delivery were enrolled after written informed consent into this Stanford University Institutional Review Board–approved study. The study was conducted at Lucile Packard Children's Hospital, Stanford University Medical Center (Stanford, CA). Parturients were enrolled over a 24-month period.

Inclusion criteria included the following: American Society of Anesthesiologists physical status class II patients, aged between 18 and 40 yr, current (at time of enrollment) BMI greater than or equal to 40, height more than 150 cm, singleton pregnancy, and gestational age of more than 37 completed weeks undergoing elective cesarean delivery. Exclusion criteria included labor, ruptured membranes, more than three previous cesarean deliveries, significant medical or obstetric morbidity, intrauterine growth retardation, placenta previa, and congenital anomaly.

All patients received 1,000 ml lactated Ringer's solution plus 500 ml hetastarch via  peripheral intravenous access. Fluid infusion and a premedication (intravenous metoclopramide 10 mg and ranitidine 50 mg) were administered approximately 30 min before spinal anesthesia. After enrollment, patients were randomly assigned using batched computer-generated random allocations, and blinding was maintained using opaque envelopes containing dose assignments. Patients were allocated to one of seven possible groups to receive 0.75% hyperbaric bupivacaine (0.75% in 8.25% dextrose; Abbott Laboratories, North Chicago, IL) in doses of 5, 6, 7, 8, 9, 10, or 11 mg. Fentanyl 10 μg (0.2 ml) and morphine 200 μg (0.4 ml) were added to each bupivacaine injection, with 10% dextrose added (0–0.8 ml) to make the total volume 2.1 ml in all cases. A combined spinal epidural (CSE) was performed at the L2/L3 or L3/L4 interspace using a loss of resistance technique with the patient in the sitting position. The spinal component was performed with a needle-through-needle technique using a 26-gauge pencil-point needle. After aspiration of cerebrospinal fluid, the intrathecal dose was injected over 5–10 s. A multiple orifice epidural catheter was threaded 5 cm into the epidural space. No drug was injected into the epidural catheter at this time. The patient was immediately laid on her right side, and the epidural catheter was taped into place. The patient was then rapidly transferred to the supine position, with a right pelvic wedge placed to facilitate left uterine displacement.

The success or failure of the intrathecal block was the primary data endpoint. Success (induction) was defined as a bilateral T6 sensory level to pinprick within 10 min of intrathecal drug administration. A failure (induction) was recorded when a T6 sensory level was not obtained within 10 min after drug administration. If a failure (induction) was recorded, an epidural supplementation of 2% lidocaine (with sodium bicarbonate 1 mEq per 10 ml and epinephrine 1:200,000) in 5-ml increments was administered to attain a T6 level. A success (operation) was defined as a successful initial sensory level, with no additional epidural anesthetic required during surgery. A failure (operation) was recorded when, despite an adequate T6 sensory level, supplemental epidural analgesia was required to complete surgery because of either a visual analog pain scale (0–100 mm; 0 = no pain and 100 = worst pain imaginable) more than 20 mm, or patient's request for additional analgesia. In cases of failure (operation), 2% lidocaine (with bicarbonate and 1:200,000 epinephrine) was administered in 5-ml bolus injections and repeated as required. Hypotension was defined as mean arterial pressure (MAP) less than 90% baseline MAP. Baseline MAP was taken as the average of three readings at admission. Phenylephrine (50–100 μg) boluses were used to treat hypotension, keeping MAP more than 90% of baseline. Intraoperative nausea and vomiting were treated with intravenous ondansetron (4 mg).

Demographic variables recorded included age, height, weight, parity, number of previous cesarean deliveries, and gestational age. Neonatal weight and Apgar scores were recorded after delivery. MAP was determined by noninvasive blood pressure measurements made at baseline (averaged over three measurements), at 2-min intervals after drug injection for the first 10 min and at 5-min intervals until the end of surgery. The lowest MAP (absolute and change from baseline) and the total dose of phenylephrine administered were all recorded. The sensory level was determined bilaterally by pinprick at 2, 4, 6, 8, and 10 min after drug administration. Subjective pain scores were determined with the use of a visual analog pain scale (0–100 mm; 0 = no pain and 100 = worst pain imaginable) at the following intervals: skin incision, delivery, uterine exteriorization, and skin closure. The incidence (i.e ., presence or absence) of nausea and vomiting was assessed at 15-min intervals from intrathecal drug administration until the end of surgery. Maternal satisfaction with anesthesia (0–100%) was assessed at end of surgery. The times at which patients met recovery discharge criteria (e.g ., hemodynamic stability, sensory, and motor block receding) were also recorded.

Demographic data are presented as mean (± SD) or median (interquartile range), where appropriate. Analysis was performed with use of the SPSS 17.0 for Windows statistical package (SPSS, Inc., Chicago, IL). Data were assessed for normal distribution of variance. Normally distributed data were assessed by one-way ANOVA, and nonnormally distributed data were assessed by the Kruskal-Wallis test. Incidence data were analyzed by the Fisher exact test. Statistical significance was defined as P < 0.05. Correlations were assessed with the use of linear regression, unless otherwise indicated.

The success or failure (binary option) and corresponding spinal bupivacaine dose were fitted to the following version of the Hill equation: Probability of successful block = dose^γ/(dose^γ50+ dose^γ), where dose is the spinal bupivacaine dose in milligrams, dose⁵⁰is the dose of bupivacaine at which there is a 50% probability of success of the spinal block, and γ is the slope of the response curve and describes the shape of the data distribution. The binary endpoints used for the logistic regression included success (operation) compared with failure (operation), and success (induction) compared with failure (induction). A naïve pooled analysis was performed, with each subject providing one data point for the fit. ED⁵⁰and ED⁹⁵values were estimated using NONMEM® version V (GloboMax™; Hanover, MD). The quality of the fit was considered based on improvement in the log likelihood value of NONMEM (an improvement of 4 of the log likelihood value consistent with P < 0.05 was considered significant) and visual assessment of the fit. This same methodology was used to determine whether the parameter estimates were different from previously published results. Comparisons between the populations were made of the equations as a whole. The equation was structured on the ED⁵⁰value and thus better predicts the ED⁵⁰value^.The ED⁹⁵value was an extrapolation of the curve, and the equation was also parameterized for the ED⁹⁵value to allow ED⁹⁵value estimates.

All 42 patients enrolled and randomly assigned completed the study according to the protocol and were included in the analysis. There were 6 parturients randomly allocated to each of the 5-, 6-, 7-, 9-, 10-, and 11-mg bupivacaine groups, 5 patients to the 8-mg bupivacaine group, and 7 patients to the 11-mg group. The unequal numbers in the groups were due to a randomization allocation error. Demographic and baseline obstetric characteristics were similar among treatment groups (table 1). The mean duration of surgery was 61 (56–75) min, with no differences among the groups.

A logistic regression plot was drawn for success (operation) (fig. 1). The 0.5 and 0.95 y -intercepts show the ED⁵⁰and ED⁹⁵values, respectively. The slope of the curve (γ) for success (operation) was 7.6 (SE, 1.6). Overall anesthetic success (success [operation]) ED⁵⁰and ED⁹⁵values (with 95% confidence interval; CI) in this morbidly obese study population were 9.8 (8.6–11.0) and 15 mg (10.0–20.0), respectively (fig. 1). These values compare with calculated ED⁵⁰and ED⁹⁵values in nonobese (mean BMI, 29) patients of 7.6 (6.7–8.5) and 11.0 (9.1–12.9), respectively, previously reported by our group using similar methodology and study conditions (fig. 1).20 

We were unable to measure the success (induction) ED⁵⁰and ED⁹⁵values due to too many initial successes and too few failures (spinal block < T6), even in the low dose range. The data resulted in an indeterminate transition point, and we were unable to construct a logistic regression curve for success (induction). Success (induction) did not reliably predict overall intraoperative success (success [operation]). The success (induction) and success (operation) at each of intrathecal bupivacaine doses are shown in figure 2. The majority of the failures occurred in the lower dose range (fig. 2). The mean time to failure in cases of failure (operation) was 47 (20–62) min (P = 0.07 among groups).

The mean time to T6 sensory block onset to pinprick among the success (induction) group was 8 ± 2 min. We found no correlation between bupivacaine dose and the speed of T6 block onset to pinprick (R =−0.03, P = 0.856). There were no differences in pain scores among the groups at skin incision, uterine externalization, delivery, and skin closure. There was an inverse correlation between the dose of bupivacaine and pain on uterine exteriorization (R =−0.44, P = 0.016). Overall median (interquartile range) maternal satisfaction with anesthesia was 100 (96–100), with no differences among the doses studied (P = 0.913). The mean postanesthetic care unit stay was 74 ± 20 min, with no differences detected among the different bupivacaine dose groups (P = 0.329).

The changes in MAP after spinal anesthesia, the dose of phenylephrine required to maintain the MAP during surgery, and the occurrence of nausea or vomiting during surgery are summarized in table 2. Eleven of the 42 patients studied (26%) experienced either nausea or vomiting intraoperatively, but no statistically significant dose-dependent differences with incidence of nausea or vomiting were found (table 2; P > 0.05).

This dose-ranging study suggests that obese and nonobese patients do not respond differently to modest doses of intrathecal bupivacaine. The study quantified the ED⁵⁰and ED⁹⁵values for intrathecal hyperbaric bupivacaine (coadministered with intrathecal fentanyl and morphine) in morbidly obese patients undergoing spinal anesthesia for cesarean delivery. A secondary analysis, comparing the ED⁵⁰and ED⁹⁵values of this study of morbidly obese patients to a nonobese population previously reported by our group using similar methodology,20suggests that intrathecal bupivacaine dose requirements for obese and nonobese women undergoing cesarean delivery are not significantly different. However, the use of a historical control for comparison (i.e ., the previous methodologically similar study by our group in nonobese patients) is associated with many known problems and biases. In addition, this secondary analysis only assessed the ED⁵⁰and ED⁹⁵values and not the entire dose response, and the power of the study would probably only have detected large differences between the populations. Although these results need to be confirmed in a prospective, controlled study, findings suggests that morbidly obese patients undergoing cesarean delivery with spinal anesthesia with modest doses of bupivacaine have similar, rather than smaller, dose requirements, compared with nonobese patients.

This large local anesthetic dose requirement in a morbidly obese population was an unexpected finding. Possible reasons for this larger than expected dose requirement include more stimulation from the forces necessary to retract adipose tissue and pannus from the surgical field, slower transfer to a supine position after block placement in the sitting position (due to body habitus), which may have affected the cephalad spread of the hyperbaric solutions, longer surgical preparation time, and greater surgical duration.

In contrast to our findings, many reviews suggest a local anesthetic dose reduction for spinal anesthesia in morbidly obese patients due to a potential lower neuraxial local anesthetic requirement.6–8Studies that have investigated the relationship between weight and the spread of intrathecal local anesthetics have demonstrated conflicting results.12–19Our study does not support an intrathecal dose reduction in morbidly obese patients. Norris et al . and Hartwell et al . found no effect of weight or BMI on the spread of sensory blockade in women undergoing cesarean delivery.15,16,22However, morbidly obese patients were not specifically studied, and fixed doses of hyperbaric bupivacaine were used. A recently published study by Lee et al .23using a modified up-down sequential allocation method, estimated similar dosing requirements (ED⁹⁵) for obese (BMI more than 30) and normal-weight patients undergoing cesarean delivery (11.9 mg [95% CI: 11.3–15.6]vs . 12.8 mg [95% CI: 10.8 to + infinity], respectively).

Minimizing the dose of intrathecal bupivacaine may reduce the incidence of maternal hypotension, decrease vasopressor requirements, decrease nausea, reduce time to discharge from the postanesthesia care unit, and improve maternal satisfaction.5,24–26However, lower intrathecal doses of local anesthetics are also associated with more reports of intraoperative pain, nausea, and late spinal anesthetic failures that sometimes require conversion to general anesthesia.5,24–26In the same studies that report these potential advantages outlined above, the incidence of visceral pain and discomfort using intrathecal bupivacaine doses of 5 and 8 mg was 5024and 35%,25respectively.

Morbid obesity confers a higher incidence of difficulty with airway management and oxygenation. If neuraxial anesthesia fails, conversion to general anesthesia may be more problematic, compared with a leaner population. The current study suggests caution when using doses less than 10 mg intrathecal bupivacaine. These results are consistent with those previously reported for hyperbaric and hypobaric bupivacaine in nonobese patients using a similar methodology,20,27where doses of bupivacaine greater than 10 mg were required for reliable anesthesia. Similarly, Petersen et al .28found that increasing the intrathecal dose of bupivacaine from 7.5–10 to 10–12.5 mg decreased the incidence of pain associated with visceral traction from 71 to 32%, emphasizing the relationship between larger doses and greater patient comfort.

There were few initial failures (spinal block < T6) in the lower dose range, and an initial T6 block to a pinprick sensation did not reliably predict overall success. This is consistent with our group's previous hyperbaric and hypobaric bupivacaine ED⁹⁵studies performed under the same conditions as described in this current study.20,27These studies highlight that T6 block to pinprick may not be adequate to predict intraoperative anesthetic success, and that the touch modality may be preferable, as suggested by Russell et al ., especially when low intrathecal doses are used.29,30CSE techniques may cause a higher sensory block of the intrathecal component, possibly as a consequence of opening the epidural catheter to atmospheric pressure.31Therefore, the ED⁵⁰and ED⁹⁵values may be slightly higher if a single-shot spinal technique is used.

In the current study, we used logistic regression to describe the dose-response curve from a linear distribution of seven doses of hyperbaric bupivacaine. The logistic regression technique uses the binary endpoint of success versus  failure and has been validated elsewhere in the anesthetic literature.32,33This methodology was previously used by our group to determine ED⁵⁰and ED⁹⁵values for nonobese patients for cesarean delivery.20,27Logistic regression, as used in this study, does, however, have potential weaknesses. The ED⁵⁰value is generally determined with greater confidence than the ED⁹⁵because it is measured from the rapidly increasing portion of the dose-response curve, whereas the ED⁹⁵value is extrapolated from the plateau portion of the curve. The success (operation) ED⁹⁵value of 15 mg is higher than the upper limit study dose (11 mg) due to mathematical extrapolation of the logistic regression dose curve to the ED⁹⁵point (fig. 1). It must be emphasized that we do not  recommend using this extrapolated ED⁹⁵value as a guide for clinical dosing in morbidly obese patients. We did not administer or evaluate doses more than 11 mg and therefore cannot comment on the safety of this mathematically derived ED⁹⁵dose. The likelihood of high spinal or severe cardiovascular side effects at this calculated ED⁹⁵dose are unknown. Although we do not recommend using doses as large as 15 mg in the morbidly obese patients, we would caution against doses less than 10 mg (i.e ., doses less than the ED⁹⁵95% CI value), if using a single-shot spinal technique.

The ideal dose of intrathecal local anesthetic for cesarean delivery balances the provision of adequate patient comfort while avoiding adverse maternal effects (particularly hypotension, nausea, and respiratory compromise). Increasing the dose of local anesthetic has been found to increase maternal hypotension24,26,34and nausea,24,34with a resultant reduction in maternal satisfaction.24We were unable to demonstrate that increasing doses of bupivacaine resulted in a greater incidence or severity of hypotension or significant differences in the incidence of nausea or vomiting. However, this study was not sufficiently powered to detect small changes in these variables. Spreading the patient sample between large numbers of different study groups, as we have done in this study, is useful for determining dose (ED⁵⁰and ED⁹⁵), but markedly reduces the power for detecting differences in continuous or discrete data.

This study highlights the difficulty of hyperbaric bupivacaine dose prediction in the morbidly obese population undergoing cesarean delivery. Morbidly obese patients appear to have much more variable response to intrathecal dosing, compared with leaner patients. This is reflected by the narrower confidence intervals around the ED⁹⁵estimates in our previous study using hyperbaric bupivacaine in a leaner patient population.20In addition, there were no failures at doses greater than or equal to 10 mg in the study of leaner patients, whereas in the current study, there were intraoperative failures at every dose studied. Population-based ED⁵⁰and ED⁹⁵values may not be helpful in individual dose determination, and determining the optimal dose for every patient is impossible due to the large variations in individual response to intrathecal local anesthetics. A CSE, continuous spinal, or epidural technique for morbidly obese patients undergoing cesarean delivery is therefore advantageous. The CSE neuraxial technique allows the adequate surgical anesthesia obtained initially with the spinal technique to be extended (with additional boluses of local anesthetic via  the epidural catheter in situ ). Butwick et al .35recently reported that a catheter-based neuraxial technique may be optimal in obese patients undergoing cesarean delivery, based on longer intraoperative times and greater requirements for intraoperative analgesia. This study similarly highlights that morbidly obese patient undergoing cesarean delivery are ill-suited for a single-shot technique, and the intrathecal dose should be used as part of a catheter-based technique.

In conclusion, this study suggests that morbidly obese and nonobese patients do not respond differently to modest doses of intrathecal bupivacaine. Findings from this dose-ranging study show that doses less than 10 mg are not to be recommended when using a single-shot spinal technique in morbidly obese patients undergoing cesarean delivery, and that intrathecal bupivacaine dose reduction is not necessary or prudent. We strongly emphasize that we are not recommending that doses as large as 15 mg are administered in morbidly obese patients undergoing cesarean delivery. This dose was a mathematically extrapolated ED⁹⁵value and was not administered or evaluated for safety in this study. Obese patients appear to have a more variable response to intrathecal dosing than leaner patients and therefore may be better suited to a CSE, epidural, or continuous spinal anesthetic technique. Even if morbidly obese patients need less intrathecal bupivacaine to achieve an initial dermatomal level, their needs for anesthesia may be greater (perhaps because of greater surgical stimulation from retraction of adipose tissue from the surgical field or due to longer surgical preparation and duration). This study highlights that adequate initial sensory levels to pinprick obtained with low doses do not guarantee adequate intraoperative anesthesia, and future studies should evaluate this important outcome (intraoperative anesthetic adequacy) in parallel to the initial sensory level achieved.