AIM: There is a lack of studies on the prevalence of frailty, and the association between frailty and mortality in a Norwegian general population. Findings regarding sex differences in the association between frailty and mortality have been inconsistent. The aim of the present study was to investigate the association between the frailty phenotype and all-cause mortality in men and women in a Norwegian cohort study. METHODS: We followed 712 participants (52% women) aged ≥70 years participating in the population-based Tromsø 5 Study in 2001-2002 for all-cause mortality up to 2016. The frailty status at baseline was defined by a modified version of Fried's frailty criteria. Cox regression models were used to analyze the association between frailty and mortality with adjustment for age, sex, disability, comorbidity, smoking status and years of education. RESULTS: In total, 3.8% (n = 27) of participants were frail (women 4.4%, men 3.2%) and 38.1% (n = 271) were pre-frail (women 45.8%, men 29.9%). During follow-up (mean 10.1 years), 501 (70%) participants died. We found an increased risk of mortality for frail older adults (multivariable-adjusted HR 4.16, 95% CI 2.40-7.22) compared with non-frail older adults. In sex-stratified analysis, the adjusted HR was 7.09 (95% CI 3.03-16.58) for frail men and 2.93 (95% CI 1.38-6.22) for frail women. Results for pre-frailty showed an overall weaker association with mortality. CONCLUSIONS: While frailty was more prevalent in women than in men, the findings suggest that the association between frailty and mortality is stronger in men than in women. Geriatr Gerontol Int 2018; 18: 1200-1205.