Epidemiology & Public Health

Biography

Biography: Gary J Macfarlane

Abstract

Background: Observational studies have shown an association between alcohol consumption and the reporting of chronic musculoskeletal pain reporting. A “J” shaped curve is observed whereby risk is highest in non-drinkers and heavy drinkers. These patterns may arise through failure to account for changes in drinking due to ill health. This study investigated associations in those with stable patterns of alcohol consumption from a large population study. Methods: UK Biobank recruited approximately a half million persons 40-69 year-old people across Great Britain. Participants answered health/lifestyle questions by touch screen at assessment centres. Questions included pain at regional sites or ‘all-over’ in last month, and whether pain was chronic (i.e. had lasted at least 3 months). Alcohol questions were frequency/amount, previous drinking if non-drinkers, consumption change in 10 years, and reasons for stopping/reducing consumption. Participants were classed as: ‘no change in alcohol consumption’ or ‘stopped/reduced alcohol consumption’. Those classed ‘as no change in alcohol consumption’ were categorised as: non-drinkers, special occasions-only, 1-3 times/month, or units/week if at least once/week. Relative risk reduction (RRR) was calculated for reporting any chronic pain, and chronic pain all-over for each category compared to non-drinkers. RRRs were adjusted for potential confounders (age, gender, BMI, education, deprivation, social networks, loneliness, mood, and smoking (‘partially-adjusted’)) and self-reported health (‘fully-adjusted’). Results: In those not stopping/reducing, risk of any chronic pain was lowest in drinkers particularly those drinking 10-15 units/week (compared to non-drinkers, RRR partially-adjusted 16.1%, RRR fully-adjusted 8.2% [99% CI 5.7-10.8%]). For chronic pain all-over, reductions were greater but consistent across all consumption levels in those drinking at least once a week (at 10-15 units/week, compared to non-drinkers, RRR partially-adjusted 70.6%, RRR fully-adjusted 50.4% [99% CI 38.3-60.3%]). Conclusions: Associations between alcohol consumption and reporting chronic pain remained even when focussing only on persons with stable consumption, suggesting association is not explained by persons with chronic pain reducing alcohol. Although it has been hypothesised that metabolites of alcohol may facilitate pain inhibitory pathways, residual confounding cannot be ruled out.