Notes
Background:Frailty is associated with older age and multimorbidity (two or more long-term conditions); however, little is known about its prevalence or effects on mortality in younger populations. This paper aims to examine the association between frailty, multimorbidity, specific long-term conditions, and mortality in a middle-aged and older aged population.
Methods:
Data were sourced from the UK Biobank. Frailty phenotype was based on five criteria (weight loss, exhaustion, grip strength, low physical activity, slow walking pace). Participants were deemed frail if they met at least three criteria, pre-frail if they fulfilled one or two criteria, and not frail if no criteria were met. Sociodemographic characteristics and long-term conditions were examined. The outcome was all-cause mortality, which was measured at a median of 7 years follow-up. Multinomial logistic regression compared sociodemographic characteristics and long-term conditions of frail or pre-frail participants with non-frail participants. Cox proportional hazards models examined associations between frailty or pre-frailty and mortality. Results were stratified by age group (37-45, 45-55, 55-65, 65-73 years) and sex, and were adjusted for multimorbidity count, socioeconomic status, body-mass index, smoking status, and alcohol use.
Findings:
493 737 participants aged 37-73 years were included in the study, of whom 16 538 (3%) were considered frail, 185 360 (38%) pre-frail, and 291 839 (59%) not frail. Frailty was significantly associated with multimorbidity (prevalence 18% [4435/25 338] in those with four or more long-term conditions; odds ratio [OR] 27.1, 95% CI 25.3-29.1) socioeconomic deprivation, smoking, obesity, and infrequent alcohol consumption. The top five long-term conditions associated with frailty were multiple sclerosis (OR 15.3; 99.75% CI 12.8-18.2); chronic fatigue syndrome (12.9; 11.1-15.0); chronic obstructive pulmonary disease (5.6; 5.2-6.1); connective tissue disease (5.4; 5.0-5.8); and diabetes (5.0; 4.7-5.2). Pre-frailty and frailty were significantly associated with mortality for all age strata in men and women (except in women aged 37-45 years) after adjustment for confounders.
Interpretation:
Efforts to identify, manage, and prevent frailty should include middle-aged individuals with multimorbidity, in whom frailty is significantly associated with mortality, even after adjustment for number of long-term conditions, sociodemographics, and lifestyle. Research, clinical guidelines, and health-care services must shift focus from single conditions to the requirements of increasingly complex patient populations.
Application 14151
Exploring multimorbidity in UK Biobank ? patterns of illness reporting and effects of comorbidity and multimorbidity on health outcomes
Our aim is to determine the extent and patterns of morbidity reporting and treatment use, and to investigate the impact of comorbidity, multimorbidity and polypharmacy on healthcare outcomes. We will also investigate prognostic factors and the specific impact of having chronic pain and/or depression on individuals with chronic illness. Research questions include but are not limited to:
- What are the patterns of morbidity reporting?
- What are the patterns of polypharmacy reporting among people with co/multimorbidity?
- What are health related outcomes of people with co/multimorbidity and do these vary by combination of conditions, e.g. pain and depression?
This research will investigate reporting of chronic illness, including heart disease, stroke, arthritis, osteoporosis, and depression, all of which are priorities for UK Biobank; and examine treatment, including polypharmacy, and outcomes for people with multiple conditions, taking into account sociodemographic and other factors. We will examine casual pathways and prognostic factors and this will allow us to gain a better understanding of multimorbidity and to consider potential management and treatment approaches for individuals with multimorbidity. The proposed research therefore aims to promote improved understanding and treatment of a wide range of illnesses. Initial statistical models will utilise assessment centre data to determine patterns of morbidity reporting and relationships with sociodemographic, lifestyle and other factors. We intend to use assay data to determine relationships between chronic illness reporting and biomarkers. Medications will be examined to determine the treatment burden that individuals with multiple conditions experience. Subsequent analysis will build on this initial work and allow investigation of the relationship over time between chronic illness and health-related outcomes using hospitalisation, primary care, and mortality data, taking into account other factors measured at the assessment centre, which may also affect these relationships. We require the full cohort for this research as information on medical conditions was gathered from all participants. This will allow the identification of morbidity groups, including a comparison group (1 or no long-term conditions).
Lead investigator: | Dr Barbara Nicholl |
Lead institution: | University of Glasgow |