BackgroundWide international variation in the prevalence of disabling low back pain (LBP) among working populations is not explained by known risk factors. It would be useful to know whether the drivers of this variation are specific to the spine or factors that predispose to musculoskeletal pain more generally.MethodsBaseline information about musculoskeletal pain and risk factors was elicited from 11 710 participants aged 20–59 years, who were sampled from 45 occupational groups in 18 countries. Wider propensity to pain was characterized by the number of anatomical sites outside the low back that had been painful in the 12 months before baseline (‘pain propensity index’). After a mean interval of 14 months, 9055 participants (77.3%) provided follow‐up data on disabling LBP in the past month. Baseline risk factors for disabling LBP at follow‐up were assessed by random intercept Poisson regression.ResultsAfter allowance for other known and suspected risk factors, pain propensity showed the strongest association with disabling LBP (prevalence rate ratios up to 2.6, 95% CI: 2.2–3.1; population attributable fraction 39.8%). Across the 45 occupational groups, the prevalence of disabling LBP varied sevenfold (much more than within‐country differences between nurses and office workers), and correlated with mean pain propensity index (r = 0.58).ConclusionsWithin our study, major international variation in the prevalence of disabling LBP appeared to be driven largely by factors predisposing to musculoskeletal pain at multiple anatomical sites rather than by risk factors specific to the spine.SignificanceOur findings indicate that differences in general propensity to musculoskeletal pain are a major driver of large international variation in the prevalence of disabling low back pain among people of working age.