Abstract
STUDY OBJECTIVE: To explore the previously stated hypothesis that risk factors for atherothrombotic disease are associated with back pain. DESIGN: Prospective (mean of four years of follow up) and retrospective analyses using two main outcome measures: (a) short (< or = 7 days) and long (> 7 days) spells of sickness absence because of back pain reported separately in men and women; (b) consistency of effect across the resulting four duration of spell and sex cells. SETTING: 14 civil service departments in London. PARTICIPANTS: 3506 male and 1380 female white office-based civil servants, aged 35-55 years at baseline. MAIN RESULTS: In age adjusted models, low apo AI was associated with back pain across all four duration-sex cells and smoking was associated across three cells. Six factors were associated with back pain in two cells: low exercise and high BMI, waist-hip ratio, triglycerides, insulin and Lp(a). On full adjustment (for age, BMI, employment grade and back pain at baseline), each of these factors retained a statistically significant effect in at least one duration-sex cell. Triglycerides were associated with short and long spells of sickness absence because of back pain in men in fully adjusted models with rate ratios (95% confidence intervals) of 1.53 (1.1, 2.1) and 1.75 (1.0, 3.2) respectively. There was little or no evidence of association in age adjusted models with: fibrinogen, glucose tolerance, total cholesterol, apoB, hypertension, factor VII, von Willebrand factor, electrocardiographic evidence of coronary heart disease and reported angina. CONCLUSIONS: In this population of office workers, only modest support was found for an atherothrombotic component to back pain sickness absence. However, the young age of participants at baseline and the lack of distinction between different types of back pain are likely to bias the findings toward null. Further research is required to ascertain whether a population sub-group of atherothrombotic back pain can be identified.