Frailty in chronic diseases: prevalence and implications for clinical management

Hanlon, Peter (2022) Frailty in chronic diseases: prevalence and implications for clinical management. PhD thesis, University of Glasgow.

Full text available as:
[thumbnail of 2022HanlonPhD.pdf] PDF
Download (10MB)



A growing number of people worldwide live with frailty. Frailty describes an age-related state of reduced physiological reserve, characterised by increased vulnerability to decompensation in response to physiological stress. People living with frailty are at increased risk of adverse health outcomes including mortality and hospital admission. There is often uncertainty over clinical management of long-term conditions in the presence of frailty. This includes uncertainty over how frailty should be identified, how frailty influences the balance of risks and benefits arising from specific diagnostic and therapeutic choices, and over the applicability of trial evidence when trials rarely measure or report frailty. Three conditions in which frailty is common, and in which these uncertainties manifest, are type 2 diabetes, rheumatoid arthritis, and chronic obstructive pulmonary disease (COPD).

This thesis addresses the following aims in each of these exemplar conditions:
• assess the prevalence of frailty
• quantify the relationship between frailty and adverse clinical outcomes
• identify and measure frailty within randomised controlled trials for each condition.

Three approaches were used for each condition: systematic review of observational studies, analysis of observational data, and analysis of individual participant data from industry-sponsored randomised controlled trials. Systematic reviews included observational studies of adults with the condition of interest (each reviewed separately), using any frailty measure, in any setting, and assessing either frailty prevalence or the relationship between frailty and clinical outcomes relevant to the exemplar condition. Observational analyses used UK Biobank (all conditions) and the Scottish Early Rheumatoid Arthritis (SERA) cohort (rheumatoid arthritis only) and assessed frailty using the frailty phenotype and the frailty index. Analyses quantified the relationships between frailty and mortality and hospital admission (all conditions); major adverse cardiovascular events (MACE), falls and hypoglycaemia (type 2 diabetes); rheumatoid arthritis disease activity; and COPD exacerbations. Finally, a frailty index was constructed using individual participant data from industry-sponsored drug trials for type 2 diabetes, rheumatoid arthritis and COPD, the prevalence of frailty examined, and the relationship between frailty and Serious Adverse Events assessed.

Research question 1: Frailty prevalence

In each exemplar condition, a wide range of frailty measures were used in observational studies within the published literature (20 measures used in 118 studies of frailty in diabetes, 11 measures in 17 studies of frailty in rheumatoid arthritis, and 11 measures in 56 studies of frailty in COPD). For all conditions, the frailty phenotype was the most commonly used (69/118 diabetes studies, 5/17 rheumatoid arthritis studies, and 32/53 COPD studies). In all conditions, prevalence varied considerably by frailty measure (generally lower using the frailty phenotype compared to other measures), age (higher prevalence in studies with greater mean age) and setting (higher in residential care and inpatient settings, lower in community-based studies). However, even among community-based studies using similar frailty measures, prevalence estimates were highly heterogenous. For all three conditions, frailty was present in people under 65-years in all studies in which this was assessed.

Research question 2: Frailty and clinical outcomes

Among participants aged between 40 and 70, frailty was associated with a range of subsequent adverse health outcomes.

In type 2 diabetes frailty was associated with an increased risk of mortality, MACE, and hospital admission with fall or fracture or with hypoglycaemia after adjustment for sociodemographic factors. These findings were similar for the frailty phenotype and frailty index. At any given level of frailty, the absolute risk of each of these outcomes was greater for older participants. The association between higher HbA1c and mortality was stronger in people with frailty compared with pre-frail or robust participants according to the frailty phenotype.

In rheumatoid arthritis frailty was associated with mortality and hospital admission using both the frailty phenotype and frailty index after adjustment for sociodemographic factors and disease activity. In SERA, a higher frailty index was also associated with higher disease activity. However, in the two years following initial diagnosis and with initiation of disease-modifying antirheumatic therapy, the mean frailty index of SERA participants reduced indicating an improvement in frailty at the group level.

Both the frailty phenotype and frailty index were associated with increased risk of mortality, hospital admission, MACE, and COPD exacerbations in people with COPD. In each case, the magnitude of the association was similar before and after adjusting for the severity of airflow limitation (measured using forced expiratory volume in 1 second [FEV1]).

Research question 3: Frailty in clinical trials

Out of 39 trials for which individual participant data were obtained, 19 trials (7 type 2 diabetes, 8 rheumatoid arthritis, 4 COPD) provided sufficient data to construct a 40-item frailty index. Based on a cut-off of 0.24, frailty was common in trials for each condition (range 7-21% in type 2 diabetes trials, range 33-73% in rheumatoid arthritis trials and range 15-22% in COPD trials). The mean frailty index was highest in rheumatoid arthritis trials, followed by COPD then type 2 diabetes. The 99th centile of the frailty index in all trials was lower than is seen in most general populations-based estimates. For all three conditions, frailty was associated with increased risk of Serious Adverse Events during trial follow-up (incidence rate ratios per 0.1-point increase in frailty index were 1.46 (95% confidence interval 1.21–1.75), 1.45 (1.13–1.87), and 1.99 (1.43–2.76) for type 2 diabetes, rheumatoid arthritis, and COPD, respectively).


Frailty is common in each of the exemplar conditions, including in people aged under 65-years in whom it is far less frequently studied. Frailty in younger people is also associated with a range of clinically significant adverse health outcomes in each condition. However, the absolute risks associated with frailty are considerably lower among younger people. This, along with the observation that frailty can improve within individuals, highlights the need to individualise clinical decisions around the implications of frailty, taking into account factors such as age and clinical context, as the implications of frailty may differ depending on age as well as the nature and severity of underlying long-term conditions. Frailty is also identifiable within clinical trials, a field where frailty is rarely reported. This shows that it is feasible to report frailty for most trials. Doing so could help inform shared clinical decision making for people living with frailty.

Item Type: Thesis (PhD)
Qualification Level: Doctoral
Subjects: R Medicine > R Medicine (General)
Colleges/Schools: College of Medical Veterinary and Life Sciences > School of Health & Wellbeing > General Practice and Primary Care
Funder's Name: Medical Research Council (MRC)
Supervisor's Name: Mair, Professor Frances, McAllister, Professor David and Lewsey, Professor Jim
Date of Award: 2022
Depositing User: Theses Team
Unique ID: glathesis:2022-83340
Copyright: Copyright of this thesis is held by the author.
Date Deposited: 09 Jan 2023 08:31
Last Modified: 10 Jan 2023 09:04
Thesis DOI: 10.5525/gla.thesis.83340
Related URLs:

Actions (login required)

View Item View Item


Downloads per month over past year