Background UK state pension eligibility ages are linked to average life expectancy, which ignores wide socioeconomic disparities in both healthy and overall life expectancy.

Objectives Investigate whether there are occupational social class differences in the amount of time older adults live after they stop work, and how much of these differences are due to health.

Methods Participants were 76 485 members of the Office for National Statistics Longitudinal Study (LS), who were 50–75 years at the 2001 census and had stopped work by the 2011 census. Over 10 years of follow-up, we used censored linear regression to estimate mean differences in years between stopping work and death by occupational social class.

Results After adjustment for age, both social class and health were independent predictors of postwork duration (mean difference (95% CI): unskilled class vs professional: 2.7 years (2.4 to 3.1); not good vs good health: 2.4 years (1.9 to 2.9)), with LS members in the three manual classes experiencing ~1 additional year of postwork duration than professional workers (interaction p values all <0.001). Further adjustment for gender and educational qualifications was reduced but did not eliminate social class and postwork duration associations. We estimate the difference in postwork years between professional classes in good health and unskilled workers not in good health as 5.1 years for women (21.0 vs 26.1) and 5.5 years for men (19.5 vs 25.0).

Conclusions Lower social class groups are negatively affected by uniform state pension ages, because they are more likely to stop work at younger ages due to health reasons.