Secondary Logo

Journal Logo

Obstetric Anesthesiology: Research Reports

The Effect of Obesity on Neuraxial Technique Difficulty in Pregnant Patients: A Prospective, Observational Study

Ellinas, Elizabeth H., MD*†; Eastwood, Daniel C., MS; Patel, Smita N., MD*†; Maitra-D’Cruze, Anita M., MD*†; Ebert, Thomas J., MD, PhD

Author Information
doi: 10.1213/ANE.0b013e3181b5a1d2
  • Free


Obstetric anesthesia practice has experienced two major changes over the last several decades: increased use of neuraxial anesthesia and analgesia and ever increasing patient obesity.1

Neuraxial anesthesia and analgesia techniques are desirable for most obese parturients for whom the increased risk of cesarean delivery leads to concerns regarding adequate anesthesia and the potential for a difficult airway.2 Although some authors found higher failure rates of epidural anesthesia/analgesia in obese pregnant patients,3,4 there is much concern, but little data, regarding factors predicting difficulty in cannulating the neuraxial canal with an epidural or spinal needle in this population or in pregnant patients in general.

Investigations in nonobstetric and mixed populations have indicated that palpation of bony landmarks is the key factor in predicting difficult neuraxial techniques.5–7 We hypothesized that obesity, as defined by current body mass index (BMI), would be the best predictor of neuraxial technique difficulty. We designed this study to determine which factors, if any, predicted technically difficult neuraxial anesthetics in pregnant patients.


After obtaining IRB approval, with a waiver of consent for this observational study, investigators recorded factors potentially affecting the initiation of 427 neuraxial anesthetics for obstetric procedures. All adult pregnant patients who were undergoing neuraxial anesthetic procedures were eligible for the study.

To maintain consistency in data collection, four investigators (all attending anesthesiologists) recorded data, which were collected whenever one of the four investigators was able to be present for the study portion of a patient’s procedure. Investigators met before and periodically during the study period to ensure as much uniformity in recording as possible. Investigators retained complete control over the level of practitioner, type of neuraxial anesthetic used, needles and drugs used, and all other aspects of care.

The investigator recorded practitioner characteristics, including whether an attending, resident, or both performed the procedure. Resident experience was recorded as both years in anesthesiology residency and days on the obstetric anesthesiology service (where our residents perform the vast majority of their neuraxial blocks). In addition to practitioner characteristics, investigators recorded other potential patient predictors of neuraxial technique difficulty, including self-declared height and weight, age, ethnicity, position (sitting versus lateral), spinal deformity (by report or observation), and procedure type (spinal, epidural, or combined spinal-epidural).

Investigators assessed two physical characteristics of the patient’s back. First, “Back Flexion” was observed by standing at the side of the patient and was recorded as either convex, straight, or concave. This characteristic was judged by the curve of the skin or flesh but not by an estimate of bony spine position. Second, “Palpation” of the patient’s bony landmarks was performed by the practitioner, whether the resident, attending, or both. This assessment was reported to and recorded by the investigator as one of four levels: visible or very easily palpable, not visible but palpable, vaguely palpable, and not palpable.

During the procedure, investigators recorded neuraxial technique difficulty in two ways: the number of needle passes and the neuraxial anesthetic placement time. A “Pass” was defined as a single attempt in which the needle moved beyond the tip of the spinous process and either succeeded in reaching the desired neuraxial space or was retracted for any reason (such as patient discomfort or contact with bony structures). Reaching the epidural or spinal space in a single try was therefore counted as “one pass.” If the first pass failed, then any second forward movement was counted as pass number 2, and so on. “Time” was recorded to the nearest minute and was defined as the interval from skin infiltration to either spinal drug injection (spinal or combined spinal-epidural) or threading of the epidural catheter (epidural). The recorded time did not include patient positioning or administration of drugs. The anesthetic was considered successful when bilateral analgesia/anesthesia of the desired height and density was achieved. Only data from eventually successful procedures were retained, because neither “passes” nor “time” was obtainable in unsuccessful procedures.

Statistical Analysis

A sample size calculation was completed for two groups of patients (high and low BMI) and an estimated odds ratio of difficulty between groups of 2.5. Assuming a Type 1 error of 0.05, groups of 400 and 450 patients gave 76% and 80% power to detect a difference between groups.

A significance level of P < 0.01 was used throughout the analyses. For all primary analyses, missing values were coded as a separate group (1 degree of freedom) within each factor and retained in the analyses; subjects with missing data were excluded from some secondary analyses. Statistical power was improved by consolidating some groups to produce more equal numbers. The ability to palpate spinous processes, or “Palpation,” was reduced from four groups to two: “Easy” (visible or palpable but not visible) and “Difficult” (vaguely palpable or not palpable). “Back Flexion” data were reduced from three groups to two: “Convex” and “Straight or Concave.” BMI was analyzed both as a continuous covariate and a categorical variable, divided into two groups.

For all models, univariate analysis was performed on all predictors, including age, height, weight, BMI, ethnicity, patient position, spinal deformity, palpation of spinous processes, back flexion, experience of the practitioner, and procedure type. Univariate analysis was performed to identify a smaller set of statistically significant factors to include in the multivariate model. Once a multivariate model was established, factors initially excluded were reexamined for associations that might have been previously missed. All analyses were performed using SAS version 9.1 (The SAS Institute, Cary, NC).

We considered several types of models to test the association between predictors of anesthetic technique difficulty and “Passes,” including negative binomial and Poisson error models. Examination of the cumulative residuals suggested that the negative binomial model was the most appropriate fit to the data. The scaled deviance statistic for the negative binomial model was 1.02, indicating no lack of fit. Lack of fit might be expected if there was a significant dependence between the individual passes. In the end, no statistical method produced contradictory results. We present the linear model with negative binomial error. This model predicts the probability of successful needle placement at each pass and presents results as the total number of passes, including the successful pass.

A linear model was used to assess the relationship between the difficulty predictors and log of placement time. The model of log time expressed results in relative time rather than absolute time; therefore, a procedure lasting 40 min was “twice as long” as one lasting 20 min, rather than “20 min longer.” Relative time provided a better fit to our model. This model excluded all cases for which an attending physician intervened with a resident’s attempts to place the anesthetic, because the resulting total time was not representative of either resident or attending efforts.

A survival model (a semiparametric Cox proportional hazards model that is based on a nonparametric baseline hazard, no censored observations) was created to gauge the effect of attending intervention on passes.

Binary logistic regression was used to test the association between BMI and the significant predictors of neuraxial technique difficulty.


Observations occurred over an 8-mo period at a teaching hospital with approximately 4000 deliveries per year. Data sheets were completed for 427 pregnant patients. Of those, 14 were rejected either because the recorded data were inconsistent with the anesthetic record or because the patient did not meet inclusion criteria, leaving 413 patient encounters for data analysis. For three patients (BMI 26, 43, and 46), the first anesthetic failed, requiring a second procedure. In those patients, time and passes were added cumulatively.

Twenty-three percent (n = 93) of patients had visible or very easily palpable spinous processes, 41% (n = 169) had palpable processes, 33% (n = 137) were found to be either vaguely or not palpable, and 3% (n = 14) had missing palpation data. Maximum back flexion was judged to be convex in 43% (n = 177) of patients, straight in 43% (n = 177), and concave in 12% (n = 50) and was not recorded in 2% (n = 9) of patients.

BMI distribution is shown in Figure 1. The mean BMI was 33 kg/m2, and 74 patients had a BMI of 40 kg/m2 or more. BMI was used as a continuous covariate for all analyses. In addition, BMI was reduced to categorical data (BMI <35 kg/m2 and BMI ≥35 kg/m2). There appeared to be a natural break in the data at 35 kg/m2, and examination of log-likelihood model fit produced no better break in the data. Patient characteristics segregated by BMI are shown in Figure 2.

Figure 1.
Figure 1.:
Distribution of body mass index (BMI). Histogram of BMI.
Figure 2.
Figure 2.:
Predictor and outcome data per body mass index (BMI) group. Percent of patients in each BMI group (BMI <35 kg/m2, BMI ≥35 kg/m2) displaying the characteristic listed. Back Flexion: Maximum ability to flex the back, judged from the patient’s side and based on the curve of the patient’s skin/flesh. Convex = curved toward practitioner; Concave = curved away from the practitioner. Palpation: Ability to palpate spinous processes. Easy = visible or easily palpable; Palpable = not visible, but reasonably palpable; Vague = vaguely palpable; Not = nonpalpable spinous processes. Total Passes: Number of needle passes needed to reach desired neuraxial space. Neuraxial Time: Time from skin infiltration to epidural catheter threading or injection of local anesthetic into cerebrospinal fluid. Attending Rescue: Number of times an attending physician took over the case from the resident physician. Staff only = attending physician did the procedure by themselves.

Neuraxial Anesthesia “Passes”

The most common number of passes was “1 pass” (n = 184), and the median (interquartile range [IQR]) number of passes was 2 (1, 3). The highest number of passes was 28.

Based on our negative binomial model, the only significant predictors of increased passes were palpation (P = 0.002) and back flexion (P < 0.001). After controlling for palpation and flexion, all other potential predictors were nonsignificant, including days the resident spent on obstetric anesthesiology (P = 0.47), resident training year in anesthesiology (P = 0.26), type of anesthesia (P = 0.17), spinal deformity (P = 0.81), ethnicity (P = 0.46), and starting position (sitting versus lateral, P = 0.21). BMI was not a predictor of difficulty, whether considered a continuous (P = 0.50) or categorical variable (P = 0.19). This model could not adequately address differences between attending and resident physicians because of attendings’ intervention in some resident procedures; a survival model (see below) addresses this issue.

The negative binomial model predicted the passes required for successful neuraxial procedures in pregnant patients with varying palpation and flexion (Table 1). It also estimated the relative increase in the number of passes required for successful placement based on palpation and flexion scores (Table 2). Palpation and back flexion were independent predictors, and for practical interpretation, the effects multiply. Therefore, a patient with both difficult palpation and reduced flexion will require approximately 3.2 times the number of passes required for the well-flexed parturient with visible spinous processes.

Table 1
Table 1:
Palpation and Back Flexion Score and Mean Number of Passes
Table 2
Table 2:
Palpation and Flexion Scores and Relative Increase in the Number of Passes

Neuraxial Anesthesia “Time”

The most frequent time recorded was 3 min; the mean time was 7 min, and the median (IQR) time was 5 min (3, 8). The shortest time recorded was 1 min; the longest was 76 min.

Based on the linear model of log time, both palpation (P = 0.001) and back flexion (P = 0.001) were significant predictors of increased time. In addition, practitioner characteristics also affected time. This was true both for attending-only (versus resident) cases (P < 0.001) and for residents stratified by days spent on the obstetric anesthesia service (P < 0.001). After controlling for palpation and flexion, all other potential predictors were nonsignificant, including BMI considered as a continuous (P = 0.29) or categorical (P = 0.53) variable.

The derived model was used to estimate the placement time based on palpability, flexion scores, and practitioner. Results are presented as the relative increase (or decrease if ratio <1) in time required for a given predictor (Table 3).

Table 3
Table 3:
Relative Change in Neuraxial Procedure Time for a Given Set of Predictors

Survival Analysis for Attending “Rescues”

In 52 of our cases, attending physicians intervened in resident physicians’ anesthetic placement, essentially “taking over” the procedure. This could produce a bias in the data, making the residents appear falsely proficient; that is, we never see how many passes they would have needed if staff had not intervened. To account for this, we performed a discrete time survival analysis on our passes data, with a time-dependent factor representing whether a resident or attending was performing each pass. (“Time” here refers to survival time in the model but not to our other outcome, neuraxial time.) The model predicted the hazard of failure to reach the desired space at each pass and yielded similar results to the other models (Table 4). Palpation and back flexion were still the only two significant predictors of difficulty, whereas attending intervention had the opposite effect, significantly decreasing the risk of failing to reach the epidural or spinal space.

Table 4
Table 4:
Survival Model—Attending Physician Intervention in Resident Procedures

Body Mass Index

After controlling for the effects of back flexion and palpation, BMI was not a significant predictor of neuraxial technique difficulty in our models for passes or time. However, in a binary linear regression model that considered only BMI, BMI strongly predicted both difficult palpation and reduced flexion (Table 5). The model predicted that a relatively small weight change would significantly impact palpation difficulty. For example, a parturient with a BMI of 35 kg/m2 would have approximately 4.6 greater odds of having a difficult-to-palpate back when compared with a patient with a BMI of 30 kg/m2.

Table 5
Table 5:
Odds of Difficult Palpation and Straight/Concave Back Flexion Per Unit Increase in Body Mass Index (BMI)

Selection of Anesthesia Provider

With the initial presumption that a higher BMI implied greater technical difficulty, we considered the possibility that more experienced personnel might be assigned to the more obese patients. We found, however, no association between BMI and total days the residents spent on the obstetric anesthesiology service (Kendall Tau-b = 0.04, P = 0.22). In addition, the Mann–Whitney (Wilcoxan) test showed no difference in the BMI of patients whose neuraxial blocks were placed by residents versus attending physicians (P = 0.84). We intend to present this and other data relating to resident education in a future publication.


The major finding of this prospective study was that the best predictors of neuraxial technique difficulty in pregnant patients were back flexion and ease of palpation of bony landmarks. Obesity per se, as measured by current BMI, did not directly predict difficulty but did predict both difficult palpation and reduced flexion.

Neuraxial anesthesia is especially desirable in obese patients. Prepregnancy BMI of >30 kg/m2 has been linked to nearly every adverse pregnancy outcome,8,9 including both cesarean and emergency cesarean delivery,3,4,10–12 with the accompanying threats of failed neuraxial anesthesia and difficult airway.3,13 Approximately one-third of United States adults were overweight in 2006 (with variation by age and ethnic group),14,15 and pregnant patients followed that trend, with prepregnancy obesity rates ranging from 20% to 25%.16,17 Studies corroborate the general conviction that obese pregnant patients have “difficult backs,” with several studies linking obesity to increased risk of epidural catheter failure.3,4,18 However, it is not clear from these studies whether failed epidural analgesia includes failure because of difficulty in cannulating the neuraxial canal.

We chose to use BMI as an obesity indicator. Although a better indicator than weight alone, BMI is an imperfect measure of obesity and does not describe many factors that might contribute to difficulty, such as adipose distribution. Nonetheless, BMI gives a general impression of the degree of overweight (used by the World Health Organization (WHO)19 and the Centers for Disease Control14 for this purpose) and may influence the practitioner’s presupposition of difficulty. The American College of Obstetricians and Gynecologists recommends measuring BMI on the first prenatal visit,20 but these data were not available to us. Because prepregnancy BMI does not predict the pattern of weight gain during pregnancy, term-gestation BMI is probably at least as valid (or invalid) a measure as prepregnancy BMI. It is likely that the limitations of BMI in describing/defining obesity contributed to its lack of predictive value.

There is no current agreement on appropriate BMI ranges for pregnant women. The Centers for Disease Control and WHO criteria in the nonpregnant patient list “obese” as a BMI ≥30 kg/m2, and a BMI of 35 kg/m2 corresponds to WHO “obesity class II,”19 but neither of these definitions considers lean gestational weight gain of pregnancy. In addition, only 14 study patients had a BMI >50 kg/m2. Therefore, conclusions regarding extremes of obesity cannot be made from our data.

In nonobstetric patients, authors have examined patient and practitioner characteristics that may predispose patients to placement difficulty. These authors found that difficulty was most affected by two things: obvious spinal deformity and the ability to adequately palpate bony spinal landmarks.5–7 These studies in nonpregnant and mixed populations used end points such as the number of skin punctures and attempts at different interspinous spaces. Our results further delineated the definition of difficulty and evaluated the influence of obesity on difficulty. Rather than recording interspace attempts, we chose needle “passes” as one determinate of difficulty, because we do not often attempt neuraxial procedures at multiple interspinous spaces or make second skin punctures. Instead, we change the needle angle through a single skin puncture to make subsequent passes. We used placement “time” as an additional measure of difficulty. It should be noted that this interval represented only working time and did not include preparation or completion time, both of which could be influenced by body habitus. Our working time, therefore, cannot be used to infer the economic costs of anesthetic time in pregnant patients.

There are several limitations to our study design. Our predictors are subjective: for example, a “palpable” spinous process to one practitioner might be “vaguely palpable” to another. Investigators met before beginning the study to attempt a consensus on these issues, but some variation may have occurred. Although designating a single practitioner to separately collect data may have mitigated this potential confounder, the 24-h nature of obstetric anesthesiology practice made this a practical impossibility.

In addition, the same practitioner collected data and performed the procedure, potentially leading to difficulty as a self-fulfilling prophecy.21 Mitigating this potential tendency is the actuality that the study results defied our a priori hypothesis; it was not obesity that best predicted difficulty with neuraxial block placement but the ability to palpate landmarks and flex the back. Obesity effects on difficulty appear to be mediated through increasing tendencies toward poorly palpable spinous processes and inadequate back flexion. Although unexpected, these conclusions are nonetheless consistent with the clinical observation that some obese parturients have surprisingly easy neuraxial needle placements.

We therefore suggest that anesthesiologists facing any pregnant patient, but especially an obese one, perform a careful examination of the patient’s back at the initial patient encounter. Based on our study data, it is this examination, not the degree of obesity, which predicts neuraxial technique difficulty.


1.Saravanakumar K, Rao SG, Cooper GM. Obesity and obstetric anaesthesia. Anaesthesia 2006;61:36–48
2.Lyons G. Failed intubation: six years’ experience in a teaching maternity unit. Anaesthesia 1985;40:759–62
3.Hood DD, Dewan DM. Anesthetic and obstetric outcome in morbidly obese parturients. Anesthesiology 1993;79:1210–8
4.Perlow JH, Morgan MA. Massive maternal obesity and perioperative cesarean morbidity. Am J Obstet Gynecol 1994;170:560–5
5.Sprung J, Bourke DL, Grass J, Hammel J, Mascha E, Thomas P, Tubin I. Predicting the difficult neuraxial block: a prospective study. Anesth Analg 1999;89:384–9 Oliveira Filho GR, Gomes HP, da Fonseca MH, Hoffman JC, Pederneiras SG, Garcia JH. Predictors of successful neuraxial block: a prospective study. Eur J Anaesth 2002;19:447–51
7.Atallah MM, Demian AD, Shorrab AA. Development of a difficulty score for spinal anaesthesia. Br J Anaesth 2004; 92:354–60
8.Weiss JL, Malone FD, Emig D, Ball RH, Nyberg DA, Comstock CH, Saade G, Eddleman K, Carter SM, Craigo SD, Carr SR, D’Alton ME; FASTER Research Consortium. Obesity, obstetric complications and cesarean delivery rate—a population-based screening study. Am J Obstet Gynecol 2004;190:1091–7
9.Sebire NJ, Jolly M, Harris JP, Wadsworth J, Joffe M, Beard RW, Regan L, Robinson S. Maternal obesity and pregnancy outcome: a study of 287,213 pregnancies in London. Int J Obes Relat Metab Disord 2001;25:1175–82
10.Joseph KS, Young DC, Dodds L, O’Connell CM, Allen VM, Chandra S, Allen AC. Changes in maternal characteristics and obstetric practice and recent increases in primary cesarean delivery. Obstet Gynecol 2003;102:791–800
11.Lynch CM, Sexton DJ, Hession M, Morrison JJ. Obesity and mode of delivery in primigravid and multigravid women. Am J Perinatol 2008;25:163–7
12.Poobalan AS, Aucott LS, Gurung T, Smith WC, Bhattacharya S. Obesity as an independent risk factor for elective and emergency ceasarean delivery in nulliparous women—systematic review and meta-analysis of cohort studies. Obes Rev 2009; 10:28–35
13.Rocke DA, Murray WB, Rout CC, Gouws E. Relative risk analysis of factors associated with difficult intubation in obstetric anesthesia. Anesthesiology 1992;77:67–73
14.Department of Health and Human Services. Centers for Disease Control and Prevention. Obesity trends among US Adults between 1985 and 2006. Available at: Accessed April 2009
15.Ogden CL, Carroll MD, McDowell MA, Flegal KM. Obesity among adults in the United States—no change since 2003–2004. NCHS data brief no 1. Hyattsville, MD: National Center for Health Statistics, 2007
16.Yeh J, Shelton JA. Increasing prepregnancy body mass index: analysis of trends and contributing variables. Am J Obstet Gynecol 2005;193:1994–8
17.Chu SY, Kim SY, Bish CL. Prepregnancy obesity prevalence in the United States, 2004–2005. Matern Child Health J 2008;1–7
18.Orbach-Zinger S, Friedman L, Avramovich A, Ilgiaeve N, Orvieto R, Sulkes J, Eidelman LA. Risk factors for failure to extend labor epidural analgesia to epidural anesthesia for Cesarean section. Acta Anaesthesiol Scand 2006;50:1014–8
19.World Health Organization. Obesity: preventing and managing the global epidemic. Report on a WHO consultation. World Health Organ Tech Rep Ser 2000;894:9
20.American College of Obstetricians and Gynecologists. ACOG Committee Opinion No. 315, September 2005. Obesity in pregnancy. Obstet Gynecol 2005;106:671–5
21.Kettler RE. The difficult neuraxial block: predictors or self-fulfilling prophecy? Anesth Analg 2000;90:765–71
© 2009 International Anesthesia Research Society