STUDY DESIGN: A review was made of the epidemiological literature on the association between smoking and low back pain (LBP).

OBJECTIVES: The first objectives was to identify studies that challenged their preliminary results with additional test factors and to see what effect this had on the outcomes. The second objective was to identify test factors that were unique to studies in which the original association disappeared after multivariate analysis.

SUMMARY OF BACKGROUND DATA: The literature on this subject is confusing and no systemic investigation seems to have been made on the possibility of spurious data interpretation.

METHODS: Twenty-four articles reporting on 126 epidemiological studies (in 24 reports) on the association between smoking and LBP were systematically reviewed by the authors, independently and jointly.

RESULTS: In eight of the thirteen studies that tested a preliminary positive association between smoking and LBP, this association remained after multivariate analysis, whereas it disappeared in after analysis in five. Two variables, marital status and occupation, were uniquely present in some of the studies in which the initial positive association was lost. However, their study samples were probably nonrepresentative of the general population.

CONCLUSIONS: A more strategic approach is needed in the study of the role of extraneous factors. In particular, marital status and occupation should be further investigated.

This abstract is reproduced with the permission of the publisher. Article only available in print.