Low back pain is common among women in developed countries, with population-based prevalence estimates ranging from 19% to 33% [1-4]. The prevalence of back pain is higher in women than in men, which suggests that pregnancy may influence the development (or course) of low back pain [5-7]. Studies of clinical outcomes in the postpartum period have shown that low back pain is a common complaint [8-10]. What is controversial, however, is the relationship between epidural anesthesia for labor and delivery and the subsequent development of low back pain. Several retrospective studies have demonstrated an association [11-13]. However, other prospective studies have shown no increased risk of acute low back pain after epidural use [14-16]. In a previous cohort study conducted in an urban tertiary care labor and delivery unit in Montreal, Quebec, Canada, we examined the relationship between epidural anesthesia and postpartum low back pain . All women who delivered a live infant were eligible for the study, except for those with a history of significant back pain preceding pregnancy and those undergoing elective cesarean section (because spinal anesthesia was the anesthetic of choice for elective sections). Patients were identified by a research nurse on Day 1 postpartum and were approached for consent (only one patient refused to participate). In total, 329 women were followed prospectively from delivery to the end of the postpartum period (six weeks). Of these, 164 women had received epidural anesthesia for labor and delivery, whereas 165 had not. The epidural catheter was placed by one of five obstetric anesthetists, and analgesia was initiated with a dose of 10-12 mL of bupivicaine 0.25%, followed by intermittent boluses of 8 mL of bupivicaine 0.25% as required and 8 mL of lidocaine 2% or chlorprocaine 3% for second-stage analgesia. Epidural fentanyl boluses of 50-100 micro g were given during labor at the discretion of the attending anesthetist. A significant difference in the prevalence of back pain between the two groups (epidural and nonepidural) was noted only on Day 1 after delivery. Using the same study population, the objective of the current study was to determine the prevalence of chronic low back pain (one year after delivery) in women who had received epidural anesthesia during labor and delivery compared with those who had not.
Written, informed consent was obtained from all subjects, and the study was approved by the hospital research ethics board. The sample size for the study (n = 329) was based on having 80% power (alpha error of 5%) to detect a twofold increased risk of postpartum back pain in women who had received epidural analgesia (given an estimated 12% prevalence of low back pain in the nonepidural group). Using the home telephone number provided on the medical chart, all 329 subjects were called 1 yr (+/- 1 mo) after delivery. To maximize the response rate, subjects were called on at least five separate occasions (including evening hours). If these attempts were unsuccessful, the office of the delivering physician was contacted to confirm the accuracy of the telephone number. Back pain was evaluated using the same standardized questionnaire used in the acute postpartum period study. The questionnaire was administered by a research nurse who was blinded to patient grouping. Women were asked to report the presence of back pain (yes/no), to quantify the pain using a numeric rating score  (with 0 representing no pain and 10 representing the worst pain imaginable), and to describe the degree to which back pain interfered with their daily activities (e.g., rest, requiring analgesia, physician consult).
The prevalence of low back pain in the epidural group was divided by the prevalence in the nonepidural group to calculate the crude relative risk of low back pain at 1 yr (epidural versus nonepidural). A 95% confidence interval around the point estimate was derived using Miettinen's test-based method . Multivariate logistic regression was used to calculate an adjusted relative risk, i.e., by taking into account the differences between the two groups on potential confounding variables . (Although logistic regression analysis provides odds ratio estimates, because the prevalence of back pain was low, odds ratios were taken to approximate relative risks).
Differences between the two groups (epidural and nonepidural) on the numeric rating scale pain scores were tested using the nonparametric Mann-Whitney U-test, whereas differences in the proportion of women in each group with functional impairment because of low back pain were tested using the chi squared test. P values of less than 0.05 were considered significant.
Of the 329 women enrolled in our original study, 244 (74%) were contacted by telephone 1 yr after delivery. The demographic and clinical characteristics of the responders (epidural versus nonepidural) are shown in Table 1. The epidural and nonepidural groups were significantly different by parity, method of delivery, ethnicity, and weight gain during pregnancy. As shown in Table 2, responders and nonresponders were similar with regard to demographic and clinical factors. Table 3 describes the frequency of back pain in the two groups (epidural and nonepidural) over time. The overall frequency of low back pain at 1 yr was 12% (29 of 244), with no significant differences between the two groups (epidural and nonepidural) in the prevalence of low back pain, functional impairment, or numeric rating scores at 1 yr. At 1 yr, 18 women complained of functional impairment because of the low back pain, comprising those with interference with normal activities (n = 6) and those with a need for analgesia and/or physician consultation (n = 12).
The crude relative risk of low back pain at 1 yr (epidural versus nonepidural) was 0.71 (10% vs 14%). Logistic regression analysis (including parity, method of delivery, ethnicity, and weight gain during pregnancy in the model) provided an adjusted relative risk (epidural versus nonepidural) of 0.63 (95% confidence interval 0.25, 1.56).
This prospective follow-up study showed no difference in the prevalence of low back pain at one year after delivery between women who had received epidural anesthesia during labor and delivery and those who had not. There were also no differences between the two groups with respect to back pain scores or level of interference with daily activities. Sampling bias was considered unlikely to have given the high response rate (74%) and the similarity between responders and nonresponders on demographic and clinical factors. Information bias was minimized by using a standardized questionnaire and by blinding the research nurse to the study hypothesis. Finally, logistic regression analysis was used to reduce or eliminate the effects of confounding variables.
An important issue in any null study is the statistical power of the study to detect a clinically important difference. A priori, we considered a twofold increase in the prevalence of low back pain in the epidural group (compared with the nonepidural group) to represent a clinically important difference. Based on the prevalence of low back pain at one year, this study had 77% power to detect a twofold difference in back pain frequency between the epidural and nonepidural groups. Previous retrospective studies, however, have suggested that the association between epidural analgesia and chronic back pain is of a smaller magnitude, i.e., these studies estimated relative risks (epidural versus nonepidural) of 1.52 and 1.80 [11,12]. The sample size in our study provided 30% and 60% power to detect relative risks of 1.5 and 1.8, respectively. Both retrospective studies, however, were compromised by poor response rates (39% and 63%), and both showed demographic differences between responders and nonresponders, which raises the possibility that response bias or recall bias may have inflated their relative risk estimates.
The objective of our original study was to evaluate the association between epidural anesthesia and back pain during the postpartum period (i.e., up to six weeks after delivery). A follow-up evaluation at one year was considered appropriate, however, given the findings of the retrospective studies, which had shown an association between epidural use and chronic back pain.
A MEDLINE search (1966-1996) identified 13 English-language publications on postpartum low back pain. Only six of these studies, however, directly compared the frequency of postpartum low back pain between women who had received epidural anesthesia during labor and delivery with those who had not. The main features and results of these six studies are presented in Table 4. Of the six studies, three were retrospective studies conducted in the United Kingdom [11-13], whereas three were prospective studies performed in North America [14-16]. A statistically significant association between epidural anesthesia and subsequent low back pain was noted only in the retrospective studies. The prospective studies, however, had sufficient power to detect similar effect sizes. There is also no evidence that there were differences between the British and North American centers in their delivery of epidural anesthesia, e.g., in terms of technique, drugs, or doses, at the time the studies were conducted.
Observational study designs (case-control or cohort) are useful and appropriate methods to identify and quantify the etiologic nature of suspected risk factors . A statistically significant association between a risk factor and an outcome, however, does not confirm causation. Generally accepted criteria to define causality include strength of the association, consistency, biologic plausibility, specificity, dose-response gradient, and temporality . Using these guidelines, epidural anesthesia as a cause of postpartum back pain has yet to be confirmed. For example, studies to date have demonstrated only a small relative risk (or strength of association), and this has been found only in retrospective studies, which raises the issues of response bias or recall bias. In addition, this association has not been replicated in prospective studies (i.e., no consistency). It has been argued that the association between epidural anesthesia and subsequent low back pain may be explained by poor posture during the delivery period because of muscular relaxation and immobility . However, equally plausible biologic mechanisms for postpartum back pain include loss of abdominal muscle support, a changing center of gravity, and repetitive lifting tasks after delivery . Biologic plausibility is therefore equivocal, and the association seems to be nonspecific. To our knowledge, no studies have demonstrated a dose-response gradient between epidural anesthesia and back pain, i.e., whether lower local anesthetic concentrations and subsequent lesser motor block reduces the likelihood of back pain. In this study, patients received a greater concentration of local anesthetic (with subsequent motor dysfunction) than is currently practiced. Patients were unable to ambulate, and some had difficulty moving independently in bed. Therefore, the dose-response gradient issue could possibly be evaluated by assessing the frequency of back pain in current patients with "walking epidurals," compared with "historical" controls. The criterion of temporality (i.e., cause precedes effect) for epidural anesthesia and back pain seems to have been met. However, given the other (equally plausible) reasons for postpartum back pain associated with pregnancy, epidural anesthesia may represent a risk marker rather than a risk factor. Other reported risk factors for postpartum chronic back pain include back pain before pregnancy, back pain during pregnancy requiring sick leave from work, physically heavy work, antenatal complaints of headaches or abdominal pain, young maternal age, single marital status, primiparity, and acute postpartum back pain [11,12,23].
In summary, we found no difference in the frequency of low back pain one year after delivery between women who had received epidural anesthesia during labor and delivery and those who had not. Our study excluded women with a history of significant back pain preceding pregnancy; therefore, whether our results are generalizable to this subgroup is unknown. A review of the comparative literature demonstrated that the association between epidural anesthesia and postpartum back pain was of low magnitude, inconsistent, nonspecific, and of equivocal biologic plausibility.
1. Croft PR, Rigby AS. Socioeconomic influences on back problems in the community in Britain. J Epidemiol Commun Health 1994;48:166-70.
2. Hildebrandt VH. Back pain in the working population: prevalence rates in Dutch trades and professions. Ergonomics 1995;38:1283-9.
3. Skovron ML, Szpalski M, Nordin M, et al. Sociocultural factors and back pain. Spine 1994;19:129-37.
4. Biering-Sorense F, Thomsen C. Medical, social and occupational history as risk indicators for low-back trouble in a general population. Spine 1986;11:720-5.
5. Svensson H, Andersson GBJ, Hagstad A, Jansson P. The relationship of low-back pain to pregnancy and gynecologic factors. Spine 1990;15:317-5.
6. Ostgaard HC, Andersson GBJ. Previous back pain and risk of developing back pain in a future pregnancy. Spine 1991;16:432-6.
7. Orvieto R, Achiron A, Ben-Rafael Z, et al. Low-back pain of pregnancy. Acta Obstet Gynecol Scand 1994;73:209-14.
8. Crawford JS. Lumbar epidural block in labour: a clinical analysis. Br J Anaesth 1972;44:66-74.
9. Grove LH. Backache, headache and bladder dysfunction after delivery. Br J Anaesth 1973;45:1147-9.
10. Ong B, Cohen MM, Cumming M, Palahniuk RJ. Obstetrical anaesthesia at Winnipeg Women's Hospital 1975-83: anaesthetic techniques and complications. Can J Anaesth 1987;34:294-9.
11. MacArthur C, Lewis M, Knox EG, Crawford JS. Epidural anaesthesia and long term backache after childbirth. BMJ 1990;301:9-12.
12. Russell R, Groves P, Taub N, et al. Assessing long term backache after childbirth. BMJ 1993;306:1299-303.
13. MacLeod J, Macintyre C, McClure JH, Whitfield A. Backache and epidural analgesia. Int J Obstet Anaesth 1995;4:21-5.
14. Groves PA, Breen TW, Ransil BJ, Oriol NE. Incidence of long-term post partum back pain and its relationship with epidural anesthesia [abstract]. Richmond, VA: Society of Obstetric Anesthetists and Perinatologists (SOAP), 1995:29.
15. Breen TW, Ransil BJ, Groves PA, Oriol NE. Factors associated with back pain after childbirth. Anesthesiology 1994;81:29-34.
16. Macarthur A, Macarthur C, Weeks S. Epidural anaesthesia and low back pain after delivery: a prospective cohort study. BMJ 1995;311:1336-9.
17. Jensen MP, Karoly P, Braver S. The measurement of clinical pain intensity: a comparison of six methods. Pain 1986;27:117-26.
18. Miettinen OS. Estimability and estimation in case-referent studies. Am J Epidemiol 1976;103:226-35.
19. Hosmer DW, Lemeshow S. Applied logistic regression. New York: John Wiley and Sons, 1989.
20. Bombardier C, Kerr MS, Shannon HS, Frank JW. A guide to interpreting epidemiologic studies on the etiology of back pain. Spine 1994;19:2047S-56S.
21. Bradford-Hill A. The environment and disease: association or causation? In: Proceedings of the annual meeting of the Royal Society of Medicine. London: Royal Society of Medicine, 1965;58:295-300.
22. Ostgaard HC, Andersson GBJ. Postpartum low-back pain. Spine 1992;17:53-5.
23. Green LW. Manual for scoring socioeconomic status for research on health behavior. Public Health Rep 1970;85:815-27.