Sueiras i , for the NNTC Objectives: Multimorbidity and frailty are consequences of aging with HIV, yet not everyone with medical disease is frail. Our objective was to identify factors associated with frailty in a multimorbid HIV-infected cohort. Design: Analysis of a prospective, observational, longitudinal cohort. Methods: Three hundred and thirty-two participants in the medically advanced NNTC were categorized as frail, prefrail, or robust with the Fried Frailty Index. A series of logistic regression analyses (first univariate, then multivariable) were conducted to determine whether medical comorbidities, immunologic and virologic parameters, and/or neuropsychiatric variables predicted increased odds of frailty. Results: The mean number of medical comorbidities per participant was 2.7, mean CD4 þ T-cell count was 530 cells/ml, and 77% had undetectable HIV RNA in blood. Twenty-two percent were frail, 55% prefrail, and 23% robust. Significant predictors of frailty in multivariable analysis were cognitive diagnosis rendered by Frascati criteria, depressive symptoms, diabetes mellitus, chronic obstructive pulmonary disease (COPD), and sex. Men were less likely to be frail than women. Higher odds of frailty were seen with: symptomatic, but not asymptomatic, cognitive impairment (compared with cognitive normals); more depressive symptoms; diabetes mellitus; and COPD. Conclusion: Neuropsychiatric illness increased odds of being frail on a predominantly physical/motoric measure, but only when symptomatic. Lack of association with asymptomatic impairment may reflect the importance of functional limitation to frailty, or possibly a unique resilience phenotype. Understanding why sex and symptomatic neuropsychiatric illness are associated with frailty will be important in managing HIVassociated morbidity in aging populations.