Background: The natural history of low back pain remains ill defined. This study has provided a unique opportunity to follow a community based cohort of people over a seven year period from 1994 to 2000.

Methodology: The respondents to the 1994 survey were the panel of adults used in this survey. This sample of 3,184 was validated against the current electoral register resulting in a sample size of 1,918. Valid responses totalled 957. Questions were designed to elicit data in an identical manner to the previous surveys.

Results: Periodicity data indicated that the majority of people with chronic persistent pain had recurrent acute episodes rather than one continuous episode. Significant predictors of persistent low back pain included higher scores on the Zung Depression Index, the Roland and Morris Disability Questionnaire, visual analogue pain scales and the Back Beliefs Questionnaire. People who developed low back pain in 2000 but didn’t report low back pain in 1994 or 1997, were distinguishable from those who did not develop low back pain by the number of disabilities reported in 1994.

Conclusions: The natural history of low back pain prevalence in this cohort of people does not entirely support the view that back pain starts as a series of acute episodes when younger which gradually merge to become chronic as the person becomes older. Many people will continue to experience intermittent acute episodes of low back pain well in to their sixties. Pre-morbid disability may result from pain in areas other than the low back, probably reflecting an increased reporting of pain generally. Some of this group will ultimately include people with chronic widespread pain. Adverse coping strategies may influence the persistence of pain in this group.

The loss of cervical lordosis in radiographs of patients presenting with neck pain is sometimes ascribed to muscle spasm. We performed a cross-sectional study of the prevalence of 'straight' cervical spines in three populations: 83 patients presenting to an accident department with acute neck pain, 83 referred to a radiology department with chronic neck problems, and 80 radiographs from a normal population survey carried out in 1958. Curvature was assessed on lateral radiographs both subjectively and by measurement. The prevalence of 'straight' cervical spines was 19% in the acute cases and 26% in the chronic cases. The 95% confidence interval for the difference was -6.4% to +19.3%. In the normal population 42% showed a straight spine, but a further third of these films had been taken in a position of cervical kyphosis; this probably reflects a difference in positioning technique. Women were more likely than men to have a straight cervical spine, with an odds ratio of 2.81 (95% CI 1.23 to 6.44). Our results fail to support the hypothesis that loss of cervical lordosis reflects muscle spasm caused by pain in the neck.