Smoking and Low Back Pain
The purpose of this article is to provide an update on the evidence, or the lack of, that exists between low back pain and cigarette smoking. This relationship has been researched and studied extensively within the past 25 years. Several and extensive reviews on this subject, many of which are described in this article, proved inconclusive. Despite these studies, there seems to be an overwhelming agreement that all researchers, as well and many clinicians, feel that there is an association between cigarette smoking and back pain. Proving this link, however, is not as simple as making that assumption.
Review of Pertinent Articles
In order to assess the association between cigarette smoking and the development of nonspecific back pain and related outcomes, Goldberg et al. reviewed 38 epidemiological studies published between 1976 and 1997 in peer-reviewed journals. This review included twelve population-based cross-sectional studies, four occupational cross-sectional studies, five prospective- general population and work place studies, three population studies, seven occupational studies combining cross-sectional and prospective follow-up studies, three case control studies, two retrospective cohort studies in the general population, one workplace study, and one retrospective cohort and cross-sectional general population study. They found positive associations between current smoking and nonspecific back pain in 18 of 28 studies in men, and 18 of 20 studies in women. For sciatica and herniated discs, there were four of eight studies for men and one of five studies for women where positive associations were found.
However, a majority of these studies were cross-sectional (18 in men and 16 in women) with only a few prospective studies (five of nine studies on men and five of six studies on women) making causal inferences problematic. The authors concluded that the research studies that were reviewed through 1997 supported the notion that smoking is associated with the incidence and prevalence of nonspecific back pain, but the evidence was too scarce to reach any conclusion regarding smoking and back pain with specific diagnoses such as sciatica and herniated discs. Furthermore, the possibility that this perceived association is an artifact arising from either selection bias or confounding factors cannot be ruled out due to the cross-sectional nature of the research.
A separate systematic and independent review by Leboeuf-Yde2 included forty-one original research reports that included 47 studies, published between 1974 and 1996. This review was undertaken to assess whether smoking causes low back pain and whether cessation of smoking reduces the incidence and/or prevalence of low back pain. The findings were disappointing in that there was no consistency of statistically significant positive associations found between smoking and low back pain.
The author pointed out that, when present, the association was usually weak and was found only in studies using very large samples. When restricting the analysis to studies with large sample size, no consistent findings were found in relation to dose-response, temporality, or reversibility. Evidence of possible causality was present only in the study with the largest sample size (N> 30,000). The author concluded that smoking should be considered a weak risk factor and not a cause of low back pain.
“...there is the hint of a link between cigarette smoking and back pain. While the link may be causal, it is safer to state that abstinence from smoking may be a useful means of primary prevention of certain types of low back pain.”15
In a separate study, Leboeuf-Yde and Yashin3 identified research that challenged their initial findings with additional “test factors” using multivariate analysis. Twenty-four publications, reporting on 126 epidemiological studies, were located and systematically reviewed by the two authors independently and jointly. The authors concluded that, of the thirteen studies that had found a preliminary positive association, the association remained in only eight of them after multivariate analysis was performed. The authors suggested that, due to their unique presence in the studies where the initial positive association was lost, marital status and occupation should be further investigated.
One study did investigate the possible confounding effect of occupation on the relationship between smoking and low back pain.4 Thirteen occupations were studied in regard to this relationship. The results indicated that smoking and low back pain were significantly correlated only in occupations that require physical exertion. Interestingly, pain in the extremities was more clearly related to smoking than low back pain. The study concluded that smoking may have a general influence on pain but a clear association between smoking and low back pain was not evident. Furthermore, the study indicated that other extraneous factors such as occupation could confound the relationship between smoking and low back pain.
In order to study causal link between smoking and back pain, one has to identify and rule out the effects of extraneous variables that could influence this relationship. This is extremely difficult to do since multiple factors could influence the experience of pain and smoking behavior. One way to get around this is to utilize, as subjects, monozygotic (identical) twins. They are genetically identical and usually have identical upbringing. One such study involved the Danish Twin Register that contains 20,888 twin pairs.5 Results indicated an association between smoking and low back pain that was found to be stronger for low back pain of longer duration. However, there was no dose-response correlation and smokers of smaller stature were not more susceptible to back pain. Furthermore, the study found no evidence linking smoking cessation with lower prevalence of this pain. The authors concluded that smoking is associated with the report of nonspecific low back pain and, in particular, recurrent low back pain or low back pain of long duration but that the relationship is unlikely to be causal.
A major criticism of the literature on smoking and spinal pain has been the cross-sectional design used in a majority of these studies. A recent prospective study has been completed in Canada in which 502 high school students were followed for one and one half years and assessed at six month intervals.6 The primary outcome for low back pain was defined as “low back pain occurrence at a frequency of at least once a week within the past six months.” Secondary outcomes included medication use for pain and disability. The results indicated that low back pain is prevalent among adolescents, with a cumulative annual incidence of 17%. Risk factors associated with the development of this pain in adolescents were a high growth spurt, poor quadriceps and hamstrings, working during school year, and smoking. The authors concluded that smokers were more likely to develop low back pain, at least among the adolescents studied.