Background: As global life expectancy increases, the number of people who live to develop multiple chronic diseases also increases, with cancer and cardiometabolic diseases as leading drivers of mortality. Socioeconomic position (SEP) is associated with multimorbidity (having 2+ conditions) and all-cause mortality. However, whether social inequalities in mortality persist after the onset of multimorbidity remains unclear. 

Objectives: We investigated the associations between SEP and mortality, before and after onset of cancer and cardiometabolic multimorbidity.

Methods: We analysed data from two large European cohorts: European Prospective Investigation into Cancer and Nutrition (EPIC), which enrolled participants aged 35-70 between 1992-2000, and UK Biobank (UKB), which enrolled participants aged 40-70 between 2006-2010. SEP was proxied by highest education attained in both cohorts, with sensitivity analyses in UKB using household income and Townsend Deprivation Index. Participants were followed through medical records, hospital, cancer, and death registries, and active follow-up to identify incident cases of primary cancers, cardiovascular diseases (CVD), type 2 diabetes (T2D), their multimorbidity (defined as the occurrence of 2 or more of these conditions), as well as all-cause mortality. Cox proportional hazards models were used to estimate the transitions from baseline to multimorbidity, baseline to mortality, and multimorbidity to mortality in multistate models. 

Results: Among 236,177 EPIC participants (median [IQR] follow-up: 17.0 [15.3,18.2] years), we observed 4,203 multimorbidity cases and 23,785 deaths; in 407,093 UKB participants (15.2 [14.5, 15.9] years), we observed 14,464 multimorbidity events and 33,736 deaths. Mortality rates (age-standardized) were five-fold greater amongst those who developed multimorbidity than those who did not. Low vs high SEP was associated with increased risk of multimorbidity (HR [95% CI] EPIC: 1.66 [1.51, 1.82], UKB: 1.61 [1.53, 1.70]), as well as with all-cause mortality (HR [95% CI] EPIC: 1.44 [1.38, 1.51], UKB: 1.44 [1.39, 1.50]), adjusting for individual diseases as time-varying covariates.  Low vs high SEP remained associated with mortality, even after onset of multimorbidity (HR [95% CI] EPIC: 1.33 [1.14, 1.55], UKB: 1.29 [1.17, 1.42]). Low SEP participants with 3+ incident conditions had a significantly greater risk of death compared to high SEP participant with no incident conditions (HR [95% CI] EPIC: 12.92 [10.65, 15.67], UKB: 15.48 [13.29, 18.02]). All associations were attenuated but persisted after adjustment for lifestyle behaviors. 

Conclusions: SEP was strongly associated with both increased risk of multimorbidity of cancer and cardiometabolic diseases and all-cause mortality. Significant socioeconomic disparities in mortality persisted even after onset of multimorbidity. Mortality hazards escalated cumulatively with both increasing socioeconomic deprivation and greater number of incident conditions. These findings underscore the need for health policies to address both the structural inequities in primary prevention of chronic diseases and multimorbidity as well as the continuing mortality inequities following the onset of multimorbidity. 

Multistate models (Hazard Ratios, 95% CI) of socioeconomic position (SEP) and incidence rates (age adjusted) per 1000 person-years for transitions from baseline to multimorbidity and mortality (stratified by sex, centre, and age in 5-year categories).