Secondary Logo

Journal Logo

Predictors of successful neuraxial block: a prospective study

de Oliveira Filho, G. R.*; Gomes, H. P.; da Fonseca, M. H. Z.; Hoffman, J. C.; Pederneiras, S. G.*; Garcia, J. H. S.*

European Journal of Anaesthesiology: June 2002 - Volume 19 - Issue 6 - p 447-451
Original Article

Background and objective: The epidural and subarachnoid spaces should be identified at the first attempt, since multiple punctures increase the risk of postdural puncture headache, epidural haematoma and neural trauma. The study aimed to describe the predictors of successful neuraxial blocks at the first attempt.

Methods: After institutional Review Board approval, 1481 patients undergoing spinal or epidural anaesthesia were prospectively enrolled. For each block we recorded: gender, age, height, weight, body habitus, anatomical landmarks (palpability of the spinous processes), spinal anatomy, patient positioning, premedication, needle type and gauge, approach, spinal level of the block, and the provider's level of experience. Retrieval of cerebrospinal fluid or loss of resistance to saline or air identified the subarachnoid and epidural spaces, respectively. The outcome variable was the first attempt success or failure (whether or not the needle was correctly located with one skin puncture and produced adequate surgical anaesthesia). Backward stepwise logistic regression tested its association with the other variables.

Results: The first-attempt rate of success was 61.51%. Independent predictors (Odds ratio, 95% confidence limits) were the quality of anatomical landmarks (1.92 (1.57; 2.35)), the provider's level of experience (1.24 (1.15; 1.33)) and the adequacy of patient positioning (3.84 (2.84; 5.19)).

Conclusions: The successful location of the subarachnoid or the epidural space at the first attempt is influenced by the quality of patients' anatomical landmarks, the adequacy of patient positioning and the provider's level of experience.

Hospital Governador Celso Ramos, *Department of Anaesthesiology, Florianópolis, SC, Brazil and Hospital Regional Homero de Miranda Gomes,Department of Anaesthesiology, São José, SC, Brazil

Correspondence to: Getúlio R. de Oliveira Filho, Rua Luiz Delfino 111/902, 88015-360, Florianópolis, SC, Brazil. E-mail:; Tel: +55 48 997 20210; Fax: +55 48 324 01052

Accepted for publication July 2001 EJA 598

Back to Top | Article Outline


The anaesthesiologist's objective when placing a neuraxial block is to produce adequate surgical anaesthesia after locating the appropriate space at the first attempt, i.e. when the needle is first advanced. This is desirable because multiple attempts are associated with complications such as postdural puncture headache [1], trauma to neural structures [2-4] and spinal haematoma [5]. Other important aspects relate to the patient's comfort and to the time taken to place the block. It is desirable to identify, preoperatively, those patients who may have a reduced chance of a successful block. These patients could have their blocks planned in advance, thus avoiding disturbances to the surgical schedule. Little is known about the factors that predict a successful neuraxial block. Except for one comprehensive study on this subject [6], there is a paucity of data about these factors. Predictors of success may depend on the patient, on the provider or on technical aspects [7-9]. Another important issue concerns the predictors of complications occurring during the placement of the block, e.g. inadvertent dural taps, subdural or intravascular injection of local anaesthetic, and paraesthesiae.

This study aimed to identify factors that independently predict a successful neuraxial block at a single attempt and also to identify factors associated with complications related to the needle advancement into the epidural or subarachnoid spaces.

Back to Top | Article Outline


The study was carried out at two teaching institutions, after approval by the respective Institutional Ethics Committees. No patient consent was required, given the non-interventional nature of the study. The sample was collected prospectively during 10 months and included 1481 patients (254 obstetric), male/female ratio: 658/823, 13-97 yr of age, who were undergoing spinal (844 patients) or epidural (637) anaesthesia. The type of block, the level, the approach, the needle and the local anaesthetic were at the provider's discretion. Identification of the subarachnoid space required the free flow of cerebrospinal fluid. The epidural space was identified by the loss of resistance to saline or air technique.

Twenty-two anaesthesiologists and 10 residents in anaesthesiology provided the blocks and completed a data collection form for each anaesthetic. The following variables about patients' characteristics were recorded before placing the block: gender, age, height and weight. Body mass index (BMI) was calculated (kg m−2). Biotype was classified as brevilineal if the xiphocostal angle was >90°, normolineal if the xiphocostal angle = 90°, or longilineal if the xiphocostal angle was <90°. Body habitus was subjectively classified as thin, normal, muscular, obese or pregnant. The spinous processes of the lumbar or thoracic vertebrae were taken as the anatomical landmarks, the quality of which was assessed by palpation, and classified as: none, if the spinous processes were non-palpable; poor, if hardly palpable; or good, if easily palpable. The apparent spinal anatomy was classified by inspection and rated as normal or abnormal, according to the absence or the presence of visible deviations.

Technical factors were recorded after the placement of the block. Patient positioning was rated as good or poor according to the ability to flex the spine adequately. Needle type and gauge, approach (median or paramedian), the level of the block (lumbar or thoracic), and the patient's position during the placement of the block (sitting or lateral decubitus) were also recorded. Each attempt corresponded to one skin puncture. Needle redirections were not considered as new attempts. The number of skin punctures and interspaces used until the local anaesthetic had been injected were calculated. The resulting anaesthesia was rated as either successful, if the surgery could proceed without any opioid or general anaesthetic supplementation, or failed, if otherwise.

The provider's experience was classified as: a resident with ≤6 months' training; a resident with >6 months' training; a second year resident; an anaesthesiologist with ≤5 yr of clinical experience; or an anaesthesiologist with >5 yr of clinical experience. Complications occurring during the placement of the block were recorded and compiled.

The overall success rate was calculated independently from the number of attempts made, i.e. the number of blocks that produced adequate surgical anaesthesia divided by the total number of blocks. The outcome was classified as a success, if the identification of the subarachnoid or epidural space was possible with one single attempt and the subsequent anaesthesia was classified as successful, or failure, if otherwise. Univariate comparisons between groups were made by unpaired t-tests, for continuous variables, and by Yates's corrected χ2-test, for categorical variables. Variables differing significantly between groups were subjected to backward stepwise logistic regression to test their association with the outcome variable. Logistic regression was also used to identify anthropometric factors associated with poor or no anatomical landmarks and poor patient positioning. Complications occurring during placement of the block were pooled on a dichotomous variable. Univariate analysis was performed as described above. This variable entered a logistic regression model having all other significant variables as independent. Fisher's exact tests were used for comparisons of success rates between the level of experience categories and the occurrence of paraesthesiae between the median and paramedian approaches. The number of attempts at each technique, and between patients presenting complications, or not, were compared using the U-test. The level of significance was set at P < 0.05.

Back to Top | Article Outline


Blocks were placed at the first chosen interspace in 1351 patients (91.22%); in 130 (8.77%) patients, the block was placed at a second interspace. The total number of attempts varied between 1 and 18. For both spinal and epidural anaesthesia, the median (lower, upper quartiles) were 1 (1, 2) attempts (P = 0.3). In 979 patients (66.10%), the block was placed with one attempt; in 268 (18.09%), with two attempts; in 126 (8.5%), with three attempts; the remaining 108 patients (7.29%) needed four or more attempts for block placement. Successful anaesthesia following one or more attempts at the first or another interspace was attained in 90.34% of patients, while the success rate at the first attempt was 61.51%.

First attempt success rates of first year residents with ≤6 months of training, first year residents with >6 months of training, second year residents, anaesthesiologists with ≤5 yr of experience and anaesthesiologists with >5 yr of experience were 51.39, 63.89, 61.33, 60.56 and 71.39%, respectively (χ2, 4 d.f. = 43.11; P = 0). Anaesthesiologists with >5 yr experience had a success rate significantly higher that the other categories, while first year residents had a significantly lower success rate than second year residents and staff anaesthesiologists with ≤5 yr experience, but only during the first 6 months of training.

Tables 1 and 2 show the results of the univariate analysis comparing success and failure. Table 3 shows the logistic regression model that best fitted the data. Analysis of the predictive capability of the model was performed by reapplying the model to the data. Using the 0.5 cut-off point for the logit, the model correctly classified 87.59% of successes and 38.59% of failures. This was the best performance of the model, correctly classifying 68.73% of the sample. Anthropometric factors that were significant independent predictors of no or poor anatomical landmarks (Odds ratios, 95% confidence limits) were: age > 40 yr (1.82 (1.38; 2.41)), BMI >25 kg m−2 (8.65 (6.20; 12.05)), abnormal spinal anatomy (10.75 (6.4; 18.07)) and brevilineal biotype (3.96 (2.88; 5.45)). Anthropometric factors significantly associated with poor patient positioning (Odds ratios, 95% confidence limits) were: age >65 yr (2.86 (1.98; 4.15)), BMI >30 kg m−2 (1.80 (1.27; 2.56)), abnormal spinal anatomy (3.02 (2.04; 4.48)) and brevilineal biotype (1.61 (1.15; 2.25)).

Table 1

Table 1

Table 2

Table 2

Table 3

Table 3

Complications occurred in 25 patients (1.68%), as follows: four (0.27%) inadvertent dural puncture; 13 (0.87%) epidural vascular puncture being six during spinal and seven during epidural block; one (0.06%) epidural intravascular injection of local anaesthetic; one (0.06%) subdural injection of local anaesthetic; and six (0.4%) paraesthesiae during spinal needle advancement. Paraesthesiae occurred only during the paramedian approach (P = 0). The median (lower, upper quartiles) of the number of attempts among patients presenting complications and no complications were 2 (2, 3) and 1 (1, 2), respectively (P = 0). Spinal abnormalities were present in 8 (32%) patients presenting complications, and in 133 (9.13%) of those presenting no complications (χ2, 1 d.f. = 12.38; P = 0). Complications occurred in 8.46% of patients in whom a second interspace was used for block placement (8.46%) and in 1.04% of those in which the block was placed on the first interspace (χ2, 1 d.f. = 39.39; P = 0). All other variables did not reach significant probability at the corresponding univariate analysis. The independent predictors of the occurrence of complications (Odds ratios, 95% confidence limits) were the number of attempts (2.01 (1.44, 2.80)) and abnormal spine anatomy (3.22 (1.32, 7.86)). However, the logistic model was unable to classify correctly any of the 25 cases of complication, when applied back to the sample.

Back to Top | Article Outline


The study identified the quality of anatomical landmarks, the adequacy of patient positioning during the block placement and the provider's level of experience as independent predictors of success at placing surgical spinal or epidural anaesthesia with one, single needle advancement. The criterion of success used was quite stringent, but it represented the expected anaesthesiologist's performance, minimizing patient discomfort, reducing the risk of complications and assuring agility to the surgical schedule. It also implied that judgement errors about the use of neuraxial anaesthesia have not occurred. As has been previously described, inappropriate selection of the local anaesthetic agent, its dose, volume and baricity, wrong selection of the position and interspace for block placement, and inappropriate choice of the anaesthetic technique are the most important determiners of failed spinal anaesthesia [9].

The overall success rate in this series (90.34%) is in accordance with reported success rates at other teaching institutions [7,9]. Our success rate for correct identification of the subarachnoid or epidural space at the first attempt (66.10%) was comparable with others previously reported (68.8 and 64%) [4,6].

The quality of anatomical landmarks is strongly associated with successful anaesthesia at the first attempt. In another study, the quality of anatomical landmarks was an independent predictor of successful spinal or epidural needle placement at the firstly chosen interspace [6]. Age, weight, body mass index, biotype, body habitus and spinal anatomy significantly differed between failures and successes in univariate analysis. However, our findings support the interpretation that they are minor predictors of success or failure, since they are hierarchically subordinates of the quality of anatomical landmarks. Independent predictors of poor anatomical landmarks were age > 40 yr, BMI > 25 kg m−2, abnormal spinal anatomy and brevilineal biotype.

The study highlights the role of adequate patient positioning during block placement as an independent predictor of success. Factors associated with poor positioning were age > 65 yr, BMI > 30 kg m−2, abnormal spinal anatomy and brevilineal biotype. Elderly patients and those presenting spinal deformities usually have more difficulty in adequately flexing the lumbar spine [10,11]. Moreover, patients having fractured bones on the lower extremities may pose a difficulty in being positioned properly.

Other technical factors, e.g. use of the median or paramedian approach and the gauge of the spinal needle, differed significantly between failures and successes at the univariate analysis. However, these variables were discarded during the logistic regression analysis because of both their insignificant P and inability to increase the predictive capability of the model.

The provider's level of experience was another independent predictor of success. Not surprisingly, the more experienced the provider, the greater the chances of success. This study has included first year residents with <6 months of training, working under strict staff supervision. Even so, their failure rate was significantly higher that those in the other categories. In contrast, staff anaesthesiologists with >5 yr of clinical experience had a success rate significantly greater than all other categories. This agrees with other studies in which the training level was significantly associated with success at neuraxial blocks [12,13] and disagrees with those which did not find such an association [6,9]. Methodological factors, especially regarding the definition of success and the inclusion of first year residents with <6 months of training in this study, may have accounted for these discrepancies. Moreover, even among experienced individuals, huge differences in the number of attempts to reach the subarachnoid space have been reported, reflecting the reported wide interindividual variability in manual dexterity [10,14].

Poor spinal anatomy and increasing number of attempts were independent predictors of complications occurring during neuraxial block placement. This finding agrees with others who found an increased incidence of postdural puncture headache, trauma to neural structures or spinal haematoma in patients presenting either one or both factors [1-5]. Paraesthesiae during needle advancement have been associated with persistent paraesthesiae and dysesthesiae [2,3], although such association has been disputed [4]. In this series, paraesthesiae occurred only when the paramedian approach was used. No such finding has been reported in other studies [2-5] and our finding disagrees with the results of another study in which a higher incidence of paraesthesiae was found when introducing an epidural catheter by the median as compared with the paramedian approach [15].

In spite of disclosing a significant association between the independent predictors of success at the first attempt, the logistic model derived in this study, when applied back to the sample, failed in classifying around 61% of failures as well as 12% of successes. This finding suggests the existence of other factors that may contribute to success, which have not been included in this study.

Our results suggest that greater chance of first-attempt success and, consequently, lesser chances of complications would be more likely in patients having easily palpable spinous processes and being capable of fully flexing the spine. Also, more experienced anaesthesiologists would have greater chances of reaching the epidural or subarachnoid space at the first attempt. Additionally, since each predictor is independent of the others, using their best possible combination might attain optimization. As the quality of anatomical landmarks cannot be altered, special attention should be given to proper patient positioning and the provider's level of experience in the management of patients presenting 'difficult backs'.

Back to Top | Article Outline


1. Harrison DA, Langham BT. Spinal anaesthesia for urological surgery. A survey of failure rate, postdural puncture headache and patient satisfaction. Anaesthesia 1992; 47: 902-903.
2. Horlocker T, McGregor D, Matsushige DK, et al. A retrospective review of 4767 consecutive spinal anesthetics: central nervous system complications. Anesth Analg 1997; 84: 578-584.
3. Auroy Y, Narchi P, Messiah A, et al. Serious complications related to regional anesthesia: results of a prospective survey in France. Anesthesiology 1997; 87: 479-486.
4. Poulakka R, Haasio J, Pirkänen MT, Kallio M, Rosenberg PH. Technical aspects and postoperative sequelae of spinal and epidural anesthesia. A prospective study of 3230 orthopedic patients. Reg Med Pain Med 2000; 25: 488-497.
5. Wulf H. Epidural anaesthesia and spinal haematoma. Can J Anaesth 1996; 43: 1260-1271.
6. Sprung J, Bourke DL, Grass J, et al. Predicting the difficult neuraxial block: a prospective study. Anesth Analg 1999; 89: 384-389.
7. Levy JH, Islas JA, Ghia JN, Turnbull C. A retrospective study of the incidence and causes of failed spinal anesthetics in a university hospital. Anesth Analg 1985; 64: 705-710.
8. Manchikanti L, Hadley C, Markwell SJ, Colliver JA. A retrospective analysis of failed spinal anesthetic attempts in a community hospital. Anesth Analg 1987; 66: 363-366.
9. Munhall RJ, Sukhani R, Winnie AP. Incidence and etiology of failed spinal anesthetics in a university hospital: a prospective study. Anesth Analg 1988; 67: 843-848.
10. Tessler MJ, Kardash K, Wahba RM, et al. The performance of spinal anesthesia is marginally more difficult in the elderly. Reg Anesth Pain Med 1999; 24: 126-130.
11. Schelew B, Vaghadia H. Ankylosing spondylitis and neuraxial anaesthesia - a 10 year review. Can J Anaesth 1996; 43: 65-68.
12. Kestin IG. A statistical approach to measuring the competence of anaesthetic trainees at practical procedures. Br J Anaesth 1995; 75: 805-809.
13. Kopacz DJ, Neal JM, Pollock JE. The regional anesthesia 'learning curve'. What is the minimum number of epidural and spinal blocks to reach consistency? Reg Anesth 1996; 21: 182-190.
14. Dashfields AK, Coghill JC, Langton JA. Correlating obstetric epidural anaesthesia performance and psychomotor aptitude. Anaesthesia 2000; 55: 744-749.
15. Jaucot J. Paramedian approach of the peridural space in obstetrics. Acta Anaesthesiol Belg 1986; 37: 187-192.


© 2002 European Academy of Anaesthesiology