SIR—The concept of ‘frailty’ has been used to identify older adults at increased risk for disability [1, 2]. A major obstacle to developing primary prevention strategies for frailty is the lack of clinical markers of early frailty, especially among high functioning older adults. The Physical Performance Battery (PPB) has been reported to predict disability in older adults [3, 4], but its association with early stages of frailty is not well established. We examined the validity of PPB to identify frailty in nondisabled and nondemented community-residing older adults. Slow gait is the most easily identifiable feature of frailty [2, 3, 5], but may occur later in the disablement process [6–8]. To assess the ability of PPB to capture mild or very early stages of frailty, we further restricted our sample to older adults with normal walking speeds.
Methods
Study population
We undertook a cross-sectional study nested within a longitudinal community-based ageing study [9]. Potential subjects (age ≥70) from local neighbourhoods who were identified from population lists were contacted by letter and then by telephone. Subjects who gave verbal consent on the telephone were invited for in-person evaluation. Exclusion criteria included severe audiovisual loss, bed bound and institutionalization. Additional exclusion criteria for this study included presence of dementia [8] or disability (inability or requiring assistance to perform activities of daily living) [10]. Study protocols were approved by the local institutional review board, and written informed consents were obtained from subjects prior to enrollment.
Clinical assessment
Clinical assistants used structured questionnaires to elicit history of medical illnesses, medication use, falls in the previous year and depressive symptoms [8, 11, 12]. Presence of self-reported depression, diabetes, heart failure, hypertension, angina, myocardial infarction, strokes, Parkinson’s disease, chronic lung disease and arthritis were used to calculate a summary illness index [8, 13]. We consulted medical records and contacted subjects’ family members or physicians to verify details. General cognitive status was assessed by the Blessed Information-Memory-Concentration Test [14]. Blessed test scores range from 0 to 32 (higher worse) and a score >4 was used to indicate minor cognitive impairment in this nondemented sample. Study clinicians rated gaits as normal or abnormal using a clinical rating scale with good reliability [9, 12].
PPB
The PPB was done by a clinical assistant, and includes tests of balance, walking and chair rise [3]. The balance portion requires maintaining side-by-side, semi-tandem and tandem stance for 10 s each, with scores ranging from 0 to 4. Gait velocity (cm/s) was measured while walking at usual pace on a 28-ft computerised walkway (GAITRite, CIR systems) [8]. Time for subjects to get up from a chair with arms across their chest five times was recorded. Categorical scores (range 0–4) for walking and chair stand subtests were based on timed quartiles [2]. Inability to complete either task received a score of 0. The sum of the three components comprised the final PPB score (range 0–12, higher is better).
Velocity was also measured while subjects walked on the mat reciting alternate letters of the alphabet [15, 16]. The walking while talking (WWT) test predicts fall risk [15, 16] and was examined as an alternate predictor of frailty.
Frailty
Frailty assessment was done by a clinical assistant, and was defined when subjects met at least three out of the following four attributes: unintentional weight loss (>10% per year), muscle weakness (grip strength in dominant hand was tested with a Jamar handgrip dynamometer and weakness defined using established cutscores [2]), exhaustion (negative response to the question ‘do you feel full of energy?’ on the Geriatric Depression Scale [11]) and self-reported low physical activity levels [2]. Our goal was to identify early stages of frailty. Hence, we excluded subjects with slow gait, the remaining frailty criterion in the Fried definition [2]. We did gait evaluations in 154 subjects with clinically normal gaits at baseline. Based on this pilot study, we derived velocity cutscores as 1.5 standard deviations below age- and sex-specific means. Gait velocity cutscores (lower excluded) in women <75 years was 75.8 cm/s, ages 75–80 was 75.7 cm/s and age >80 years was 66.9 cm/s. Gait velocity cutscores for men age <75 years was 87.2 cm/s, ages 75–80 was 78.8 cm/s and age >80 years was 66.6 cm/s.
Data analysis
Comparisons between subjects with and without frailty were done with chi-square test for categorical variables and two-sample t test for continuous variables [17]. We used binary logistic regression analysis to study the cross-sectional associations of PPB and WWT with frailty, adjusting for age, gender, ethnicity, education, illness index, previous falls, Blessed test scores and clinical gait abnormalities [17]. Results are reported as odds ratio (OR) with 95% confidence interval (CI). We also examined the association of individual PPB components and WWT with frailty in separate models adjusted for the same covariates [3]. In the absence of a criterion standard for frailty, we also conducted a secondary analysis using the summary illness index as an alternate definition of frailty to verify the reliability of our results. The illness index was treated as count data, and Poisson regression model was used to examine its association with PPB scores adjusted for age, gender, ethnicity, education, previous falls, Blessed test scores and clinical gait abnormalities.
Declaration of sources of funding
This work was supported by the National Institute on Aging (grant RO1 AG025119). The funding source played no role in the design, execution, data analysis or writing of the study.
Results
Of the 713 subjects enrolled in the ageing study during a 47-month period from September 2004 till July 2008, 174 were excluded due to slow gait (n = 98), dementia (n = 18), disability (n = 8) or combinations of these conditions (n = 50). Average age of participants was 80.1 years and most were women (60.5%). Mean gait velocity was 101.2 cm/s. There was a low prevalence of illness and overall frailty scores were low (mean 1.6, maximum 4), supporting the high functional status of this sample.
Univariate associations
Of the 539 eligible subjects, 106 (19.7%) met frailty criteria [2]. Table 1 shows that the frail group included more women (67.9% vs 58.7%) and had lower education (13.3 vs 14.3 years). Frail subjects had more clinical gait abnormalities and weaker grip. They were slower both while walking at usual pace and during WWT. PPB scores were worse in frail subjects.
Table 1.
Baseline variables in overall group and by frailty status
| Variable | Overall group (n = 539) | Frail (n = 106) | Not frail (n = 433) | P-valuea |
|---|---|---|---|---|
| Age | 80.1 ± 5.2 | 80.5 ± 5.4 | 79.9 ± 5.2 | 0.33 |
| Female, % | 60.5 | 67.9 | 58.7 | 0.08 |
| Education, years | 14.1 ± 3.4 | 13.3 ± 3.1 | 14.3 ± 3.5 | <0.001 |
| Ethnicity, % | ||||
| Caucasian | 71.6 | 67.0 | 72.7 | 0.17 |
| African-American | 22.8 | 29.3 | 21.3 | |
| Frailty score (0–4), mean ± SD | 1.6 ± 1.1 | 3.2 ± 0.4 | 1.1 ± 0.7 | <0.001 |
| Illness index (0–10), mean ± SD | 1.1 ± 0.9 | 1.4 ± 1.1 | 1.1 ± 1.0 | 0.002 |
| Minor cognitive impairment, %b | 9.4 | 11.3 | 9.0 | 0.06 |
| Previous falls, % | 14.1 | 19.1 | 12.9 | 0.10 |
| GDS score (0–15)c | 2.0 ± 2.0 | 2.8 ± 2.2 | 1.9 ± 1.9 | <0.001 |
| Clinical gait abnormality, % | 12.9 | 17.9 | 11.8 | 0.09 |
| Grip strength, kg/cm2 | 23.3 ± 7.4 | 20.2 ± 6.1 | 24.1 ± 7.5 | <0.001 |
| Normal gait velocity, cm/s | 101.2 ± 17.7 | 95.7 ± 16.8 | 102.6 ± 17.7 | <0.001 |
| WWT velocity, cm/s | 72.4 ± 23.4 | 67.6 ± 24.9 | 73.5 ± 24.0 | 0.008 |
| PPB score (0–12) | 10.0 ± 1.7 | 9.5 ± 1.7 | 10.1 ± 1.6 | <0.001 |
P-values are for comparison of subjects with and without frailty.
Blessed test score >4.
Geriatric Depression Scale [11].
Multivariate associations
PPB (OR 0.86, 95% CI 0.76–0.97, P = 0.01) was associated with presence of frailty in this high functioning sample. Only the illness index (OR 1.33, 95% CI 1.07–1.65) and education (OR 0.96, 95% CI 0.86–0.98) among the remaining covariates included in the model were significantly associated with frailty. In secondary analysis, PPB scores (estimate per one-point increase in score—0.06, 95% CI −0.11 to −0.01, P = 0.01) were also associated with illness index, our alternate definition of frailty.
Table 2 presents sensitivity and specificity of various PPB scores.
Table 2.
Sensitivity and specificity of the Physical Performance Battery (PPB) in identifying frailty in high functioning adults with normal mobility
| PPB scores (range 0–12) | Sensitivity, % | Specificity, % |
|---|---|---|
| ≤4 | 1.9 | 100.0 |
| ≤6 | 3.8 | 96.8 |
| ≤8 | 51.9 | 69.6 |
| ≤10 | 100.0 | 0.0 |
None of the individual PPB components (examined in separate models) were significantly associated with frailty. Our alternate predictor, WWT velocity, was also not associated with frailty (OR 1.00, 95% CI 0.99–1.01).
Discussion
In this large, well-characterised cohort of community-residing nondisabled and nondemented high functioning older adults, a simple clinical battery identified signs of frailty. Each one-point increase in the PPB score was associated with a 14% decreased risk of frailty even after accounting for several potential confounders. Our findings extend the previously reported associations of PPB with disability in high functioning older adults to detecting frailty [3]. Our findings show that frailty can be identified in older adults with normal walking speeds. These individuals are traditionally considered to be at lower risk for frailty and accounted for a significant proportion of our sample [2]. Our finding suggests that PPB may detect early signs of frailty even before the occurrence of slow gait. Demonstrating the utility of PPB for detecting frailty in healthier and high functioning older adults with normal mobility has clinical implications such as identifying high risk subjects early for preventive interventions.
Slow gait has been reported to be the strongest predictor of disability among frailty criteria [5]. Gait velocity mainly accounts for the association of PPB with disability [3]. However, the overall PPB was better at identifying frailty than its individual components. This may reflect the lower sensitivity of components such as gait velocity when applied to older adults with walking speeds in the normal range. The WWT test was also not associated with frailty. The lack of significance may reflect the Cardiovascular Health Study criteria used [2], which does not include cognitive criteria.
Strengths of this study included the large sample and standardised evaluation procedures. We used a reduced frailty index (excluding slow gait) similar to other recent studies [18]. A simple frailty index consisting of weight loss, inability to rise from a chair and poor energy was reported to predict adverse outcomes as well as the five-item Cardiovascular Health Study index [18]. While there is an overlap in the physical domains assessed by PPB and those used in the Fried criteria, our results are supported by our secondary analysis using illness index as an alternate definition of frailty.
Several potential limitations need to be noted. The cross-sectional design limits inferences of causality. Previous studies have used different cutscores (some higher) to define slow gait [3, 4, 6, 19]. Unlike these studies, we derived normative data for gait velocity in subjects without clinical gait abnormalities. The resultant velocity cutscores are, hence, specific to this cohort and need to be validated on other samples. While we focused on velocity, other gait measures such as gait variability may have stronger associations with frailty [8]. But these variables have not been used to define frailty and require specialised equipment for their measurement [8].
Given the high functioning status of this cohort, modest sensitivity and specificity for any test of frailty is expected. In general, lower (worse) PPB scores had better specificity but lower sensitivity for identifying frailty, whereas higher PPB scores had high sensitivity with low specificity. A PPB score of ≤8 provided a sensitivity of 52% and specificity of 70% for detecting frailty. Our findings are supported by a prior study which reported that a PPB cutscore of ≤9 identified frailer subjects with increased illness burden and functional limitations among nondisabled and nondemented older adults recruited for a clinical trial to prevent disability [20]. However, individual investigators or clinicians may choose different PPB cutscores to maximise either sensitivity or specificity depending on their goals. The PPB is simple, easy and does not require specialised equipment or extensive training of testers. While some subjects could not complete some components of the PPB due to frailty, none of our subjects refused to do the PPB, suggesting a high degree of acceptability. A simple clinical battery identified early stages of frailty in high functioning older adults with normal mobility, and may help institute preventive measures early.
Key points
Identifying early stages of frailty among high functioning older adults is a challenge.
The physical performance battery detects early stages of frailty even among high functioning older adults with normal walking speeds.
The overall battery, but not the individual components, predicted frailty in high functioning older adults.
Conflicts of interest
None.
References
- 1.Fried LP, Ferrucci L, Darer J, et al. Untangling the concepts of disability, frailty, and comorbidity: implications for improved targeting and care. J Gerontol A Biol Sci Med Sci. 2004;59:255–63. doi: 10.1093/gerona/59.3.m255. [DOI] [PubMed] [Google Scholar]
- 2.Fried LP, Tangen CM, Walston J, et al. Frailty in older adults: evidence for a phenotype. J Gerontol A Biol Sci Med Sci. 2001;56:M146–56. doi: 10.1093/gerona/56.3.m146. [DOI] [PubMed] [Google Scholar]
- 3.Guralnik JM, Ferrucci L, Pieper CF, et al. Lower extremity function and subsequent disability: consistency across studies, predictive models, and value of gait speed alone compared with the short physical performance battery. J Gerontol A Biol Sci Med Sci. 2000;55:M221–31. doi: 10.1093/gerona/55.4.m221. [DOI] [PubMed] [Google Scholar]
- 4.Studenski S, Perera S, Wallace D, et al. Physical performance measures in the clinical setting. J Am Geriatr Soc. 2003;51:314–22. doi: 10.1046/j.1532-5415.2003.51104.x. [DOI] [PubMed] [Google Scholar]
- 5.Rothman MD, Leo-Summers L, Gill TM. Prognostic significance of potential frailty criteria. J Am Geriatr Soc. 2008;56:2211–116. doi: 10.1111/j.1532-5415.2008.02008.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Montero-Odasso M, Schapira M, Soriano ER, et al. Gait velocity as a single predictor of adverse events in healthy seniors aged 75 years and older. J Gerontol A Biol Sci Med Sci. 2005;60:1304–9. doi: 10.1093/gerona/60.10.1304. [DOI] [PubMed] [Google Scholar]
- 7.Verbrugge LM, Jette AM. The disablement process. Soc Sci Med. 1994;38:1–14. doi: 10.1016/0277-9536(94)90294-1. [DOI] [PubMed] [Google Scholar]
- 8.Verghese J, Wang C, Lipton RB, et al. Quantitative gait dysfunction and risk of cognitive decline and dementia. J Neurol Neurosurg Psychiatry. 2007;78:929–35. doi: 10.1136/jnnp.2006.106914. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Verghese J, LeValley A, Hall CB, et al. Epidemiology of gait disorders in community-residing older adults. J Amer Geriatr Soc. 2006;54:255–61. doi: 10.1111/j.1532-5415.2005.00580.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Verghese J, Lipton RB, Wang C. Quantitative gait markers and incident fall risk in older adults. J Gerontol A Biol Sci Med Sci. 2009;64:896–901. doi: 10.1093/gerona/glp033. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Yesavage JA, Brink TL, Rose TL, et al. Development and validation of a geriatric depression screening scale: a preliminary report. J Psychiatr Res. 1982;17:37–49. doi: 10.1016/0022-3956(82)90033-4. [DOI] [PubMed] [Google Scholar]
- 12.Verghese J, Lipton RB, Hall CB, et al. Abnormality of gait as a predictor of non-Alzheimer's dementia. N Engl J Med. 2002;347:1761–8. doi: 10.1056/NEJMoa020441. [DOI] [PubMed] [Google Scholar]
- 13.Holtzer R, Friedman R, Lipton RB, et al. The relationship between specific cognitive functions and falls in aging. Neuropsychology. 2007;21:540–8. doi: 10.1037/0894-4105.21.5.540. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Blessed G, Tomlinson BE, Roth M. The association between quantitative measures of dementia and of senile change in the cerebral grey matter of elderly subjects. Br J Psychiatry. 1968;114:797–811. doi: 10.1192/bjp.114.512.797. [DOI] [PubMed] [Google Scholar]
- 15.Verghese J, Kuslansky G, Holtzer R, et al. Walking while talking: effect of task prioritization in the elderly. Arch Phys Med Rehabil. 2007;88:50–3. doi: 10.1016/j.apmr.2006.10.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Verghese J, Kuslansky G, Katz M, et al. J Neurol Sci. 2001. A divided attention task differentially slows elderly with dementia; p. S48. [Google Scholar]
- 17.Altman DG. Practical Statistics for Medical Research. London: Chapman & Hall/CRC; 2006. [Google Scholar]
- 18.Ensrud KE, Ewing SK, Taylor BC, et al. Comparison of 2 frailty indexes for prediction of falls, disability, fractures, and death in older women. Arch Intern Med. 2008;168:382–9. doi: 10.1001/archinternmed.2007.113. [DOI] [PubMed] [Google Scholar]
- 19.Cesari M, Kritchevsky SB, Penninx BWHJ, et al. Prognostic value of usual gait speed in well-functioning older people—results from the Health, Aging and Body Composition Study. J Am Geriatr Soc. 2005;53:1675–80. doi: 10.1111/j.1532-5415.2005.53501.x. [DOI] [PubMed] [Google Scholar]
- 20.Bandinelli S, Lauretani F, Boscherini V, et al. A randomized, controlled trial of disability prevention in frail older patients screened in primary care: the FRASI study. Design and baseline evaluation. Aging Clin Exp Res. 2006;18:359–66. doi: 10.1007/bf03324831. [DOI] [PMC free article] [PubMed] [Google Scholar]