What is the prevalence and implications of frailty in diabetes? A systematic review

Talk Code: 
I.2
Presenter: 
Isabella Fauré
Co-authors: 
Peter Hanlon, Isabella Fauré, Neave Corcoran, Elaine Butterly, David McAllister, Frances S Mair
Author institutions: 
University of Glasgow

Problem

With rising life expectancy, the average age of people with diabetes mellitus is increasing. Diabetes is known to be associated with frailty. Frailty describes an age-related decline in physiological reserves leading to increased vulnerability to decompensation and poor health outcomes. Diabetes guidelines recommend relaxing glycaemic targets in people with frailty, citing increased risks from hypoglycaemia. However, there is no consensus on how frailty should be measured nor the effects of frailty on health outcomes in diabetes. We performed a systematic review of studies assessing the prevalence of frailty in people with diabetes and the association between frailty and clinical outcomes.

Approach

We searched three electronic databases: Medline, Embase and Web of Science. Abstracts and full texts were screened to identify studies meeting the following inclusion criteria: (i) include people with diabetes (type 1, type 2, or unspecified); (ii) assess frailty status using a recognized frailty measure; (iii) report the prevalence of frailty in people with diabetes and/or the association between frailty and clinical outcomes in the context of diabetes. We included observational studies in any setting (community, hospital, institutional care). Quality was assessed using the Newcastle Ottawa Scale.

Findings

We screened 3,007 abstracts and 267 full texts, from which 74 studies were identified as eligible for inclusion. Studies were heterogenous in the type of diabetes studied (n=28 type 2, n=46 unspecified), setting, age of participants, and frailty measure used. Median sample size was 189 (IQR 95-534). Twenty-one different frailty scales were used. The frailty phenotype was the most frequently used measure (n=29 studies) followed by the frailty index (n=11), FRAIL scale (n=5), clinical frailty scale (n=3), Edmonton frailty scale (n=3), comprehensive geriatric assessment (n=3). Fourteen other scales were used in one or two studies each. The median prevalence of frailty in community studies was 9% (IQR 4%-17%). Prevalence was higher in inpatient settings (median 36%, IQR 22%-48%). Frailty was associated with mortality (9/9 studies), hospitalization (5/6 studies), cardiovascular events (3/4), hypoglycaemia (2/2), neuropathy (1/1), depression (1/1) lower quality of life (1/1) and incident disability (1/1). Frailty was associated with higher HbA1c in 3/6 studies, with 3 studies also showing that a high proportion of frail people had HbA1c levels below target range. No studies assessed if the impact of glycaemic control varied by frailty status.

Consequences

Frailty is common in diabetes and associated with a range of adverse outcomes. There is little consensus over how best to measure frailty, and no studies exploring how the choice of measure impacts on the estimated prevalence and the relationship with outcomes. Few studies in diabetes have assessed frailty measures designed for clinical practice. Future work should clarify the impact of choice of measure as well as the implications of glycaemic control to help guide treatment recommendations.

Submitted by: 
Peter Hanlon
Funding acknowledgement: 
Peter Hanlon is funded by a Medical Research Council Clinical Research Training Fellowship