Elsevier

Ageing Research Reviews

Volume 21, May 2015, Pages 78-94
Ageing Research Reviews

Review
Modifications to the frailty phenotype criteria: Systematic review of the current literature and investigation of 262 frailty phenotypes in the Survey of Health, Ageing, and Retirement in Europe

https://doi.org/10.1016/j.arr.2015.04.001Get rights and content

Highlights

  • The frailty phenotype has been frequently modified among studies.

  • Physical inactivity and weight loss were the phenotypic criteria most often modified.

  • Modifications have great impact on its classification and predictive ability.

  • Among included phenotypes, frailty prevalence ranged from 12.7 to 28.2%.

  • Areas under receiver operating characteristic curves for predicting mortality ranged from 0.607 to 0.668.

Abstract

We conducted a systematic review to determine variability in how the criteria of the frailty phenotype (grip strength, weight loss, exhaustion, walking speed, physical activity) were assessed. We then evaluated the impact on estimating prevalence and mortality of modifying the criteria, using the Survey of Health, Ageing, & Retirement in Europe (SHARE). Five databases were searched for original research articles published after 2000, which evaluated frailty using the phenotypic criteria. Among the 264 included studies, 24 studies provided enough information to demonstrate that all criteria were assessed as proposed in the original frailty phenotype study by Fried et al. (2001). Physical inactivity and weight loss were the criteria most often modified. We then created 262 phenotypes from SHARE based on common modifications found in the review. Among these phenotypes, frailty prevalence ranged from 12.7% to 28.2%. Agreement with the primary frailty phenotype ranged from 0.662 to 0.967 and internal consistency ranged from 0.430 to 0.649. Women had 2.1–16.3% higher frailty prevalence than men. Areas under receiver operating characteristic curves for discriminating five-year mortality ranged from 0.607 (95% CI: 0.583–0.630) to 0.668 (0.645–0.691). The frailty phenotype often has been modified, and these modifications have important impact on its classification and predictive ability.

Introduction

Frailty challenges healthcare professionals and has pervasive impact on health and the outcomes of health care. It has been proposed that frailty should always be considered when treating the older patient (Theou and Rockwood, 2012). Various instruments have been developed to measure frailty (deVries et al., 2011); the “frailty phenotype” is the most commonly used way. The phenotype determines frailty based on five criteria: weight loss, exhaustion, physical inactivity, slow walking speed, and weak grip strength. Individuals with three or more of these problems are classified as frail, those with one or two are said to be “prefrail” and those with none are nonfrail, or robust. Since the first publication of this frailty measurement tool in 2001 (Fried et al., 2001), the five phenotypic criteria have been measured in different ways across studies, including by its originators (Bandeen-Roche et al., 2006, Makary et al., 2010, Eckel et al., 2011), which potentially limits the comparability of studies using this definition. Various studies (Johansen et al., 2014, Eckel et al., 2011, Saum et al., 2012, Freiheit et al., 2011, Shamliyan et al., 2013) showed that modifying the phenotypic criteria could change estimates of the prevalence of frailty and the predictive ability of the phenotype, leading to potentially different classifications and results. Kutner and Zhang (2013) recently commented on a study using a modified frailty phenotype, stating that “assignment of frailty classification is very dependent on the particular criteria and operational measures that are applied”. In particular, they commented on the replacement of the performance-based measures (grip strength, walking speed) in the original frailty phenotype definition with self-reported items. This modification is very common due to the inconvenience and infeasibility of assessing performance-based measures in the measurement of frailty, especially in clinical settings. Therefore, the aims of this study were to (1) conduct a systematic review of the literature to determine variability in measurement methods for the frailty phenotype’s five criteria, and (2) compare whether modifying the phenotypic criteria available in the Survey of Health, Ageing, & Retirement in Europe (SHARE) would have an impact on the classification and predictive ability of the frailty phenotype.

Section snippets

Systematic review

We searched Medline, Embase, Psycinfo, Scopus and Eric up to November 2012. Two arms of the search strategy were developed and intersected using the Boolean term “AND”: Frail AND all reasonable descriptors of the frailty phenotype (see Appendix 1 of Supplementary material for Medline search strategy). The database search results were uploaded into Refworks, which was used to manage the screening process. Two members of the review team independently screened the title and abstracts of the

Systematic review

The preliminary search yielded 2198 citations. After an initial screening of all titles and abstracts, 1051 articles remained from which full texts were obtained. Of these articles, 264 included some version of the frailty phenotype (Fig. 1; see Appendix 2 of Supplementary material for the list of articles). The total number of participants in these 264 studies was 483,039, ranging from 4 to 40,657. There were 187 studies that included both men and women, 22 studies examined only men and 55

Discussion

We conducted a systematic review to determine variability in the measurement of frailty phenotype criteria across studies, and then conducted secondary analysis of SHARE to compare whether modifying the phenotypic criteria has an impact on the properties of the frailty phenotype. We found that the great majority of studies modified the frailty phenotype criteria and that modifying the criteria in SHARE showed substantial differences in frailty prevalence, sex differences in frailty, internal

Acknowledgments

We would like to thank Fernando Pena, Yoko Ishigami, Byung Wook Chang, Roxanne Sterniczuk, and Henriëttevan der Roest for helping with the translation of the non-English articles included in the systematic review.

This review and analyses were supported by operating grant MOP209888 to KR from the Canadian Institutes of Health Research (CIHR) and by the Fountain Family Innovation Fund of the Queen Elizabeth II Health Care Foundation, Halifax, Nova Scotia Canada. OT is supported by a Banting

References (29)

  • A. Clegg et al.

    Diagnostic test accuracy of simple instruments for identifying frailty in community-dwelling older people: a systematic review

    Age Ageing

    (2015)
  • S.P. Eckel et al.

    Surrogate screening models for the low physical activity criterion of frailty

    Aging Clin. Exp. Res.

    (2011)
  • A. Etman et al.

    Socio-demographic determinants of worsening in frailty among community-dwelling older people in 11 European countries

    J. Epidemiol. Community Health

    (2012)
  • E.A. Freiheit et al.

    Operationalizing frailty among older residents of assisted living facilities

    BMC Geriatri.

    (2011)
  • Cited by (0)

    1

    Present address: Centre d’imagerie cérébrale, Douglas Mental Health Institute, McGill University, 6875 Boulevard LaSalle, Montreal, Quebec, Canada.

    View full text