The number of people living with ≥ two health conditions, termed multimorbidity, is increasing. We investigated the impact of multimorbidity on all-cause and cause-specific mortality risk in 502,370 UK Biobank participants aged 40 to 70 years. Participants attended an assessment centre between 2006 and 2010 and self-reported medical conditions during a nurse-led verbal interview. The presence of ≥ 2 long-term conditions from a preselected list of 43 conditions was used to define multimorbidity. In a training sample (80% of participants with multimorbidity), disease clusters were identified in four groups: women aged (1) 40-59 or (2) 60-70, and men aged (3) 40-59 or (4) 60-70. Mortality was ascertained from linkage to death records. Multivariate Cox proportional-hazards regression models were used to assess the association between multimorbidity and mortality adjusted for age, sex, ethnicity, socioeconomic status and education. Over a 16-year follow-up period (median = 13 years) dose-response associations were observed between number of multimorbid conditions and risk of all-cause mortality (n = 44,399 deaths), and particularly strong dose-response associations with cause-specific deaths due to cardiovascular and respiratory conditions. For women, a mental health/cancer/pain-related conditions cluster at ages 40-59 (Hazard Ratio [HR] = 2.61, 95% Confidence Interval [CI] 2.33-2.93), and a respiratory and pain-related conditions cluster at ages 60-70 (HR = 2.03, 95% CI 1.0-2.17), were associated with the greatest risk of mortality. For men, clusters of cardiometabolic conditions at ages 40-59 (HR = 3.43, 95% CI 3.14-3.74) and 60-70 (HR = 2.24, 95% CI 2.13-2.35) were associated with greater mortality risk. These findings suggest that understanding the impact of multimorbidity, and especially clusters of disease, is important for tailoring healthcare approaches for mortality risk reduction.