Designing randomized, controlled trials aimed at preventing or delaying functional decline and disability in frail, older persons: a consensus report

The discovery of effective interventions to prevent or delay disability in older persons is a public health priority. Most likely to benefit from such interventions are frail individuals who are not yet disabled and those with early disability who are at high risk of progression. In spite of this frail older persons have often been excluded from research on the assumption that they would not tolerate testing or benefit from treatment. The Interventions on Frailty Working Group developed recommendations to screen, recruit, evaluate, and retain frail older persons in clinical trials. Specific recommendations are: Eligibility screening should include a multistage process, to quickly exclude those who are too well and those who are too sick. Inclusion criteria should target those most likely to benefit, be meaningful to clinicians, and reflect advancements in the frailty research area. Disability outcome measures should include self-reported, objective, and proxy measures. Strategies to improve retention and compliance and to monitor their effectiveness should be an integral part of the study design. Estimation of cost and sample size should contemplate high dropout rates and interference by competing outcomes. Additional research is needed to refine criteria for screening frail older persons, identify objective measures of disability that are reliable and valid in frail older persons, and improve the informed consent process for high-risk participants, recognizing that research in this subgroup is essential to improving their health outcomes.  

Coagulation and activation of inflammatory pathways in the development of functional decline and mortality in the elderly

PURPOSE: This study was performed to determine the effects of markers of inflammation (interleukin 6) and coagulation (D-dimer) on mortality and functional status in older persons. METHODS: Subjects were selected for the Duke Established Populations for Epidemiologic Studies of the Elderly. In 1992, 2569 subjects (age >71 years) were interviewed, of whom 1,723 had interleukin-6 and D-dimer measurements. Values of interleukin 6 and D-dimer were categorized into quartiles. Outcomes were mortality (through 5 years) and functional status (through 4 years). Relative risks were estimated with proportional hazards models that adjusted for potential confounders. RESULTS: The relative risk of mortality was 1.28 (95% confidence interval [CI]: 0.98 to 1.69; P = 0.06) for those with only interleukin-6 levels in the highest quartile, 1.53 (95% CI: 1.18 to 1.97; P = 0.001) for subjects with only D-dimer levels in the highest quartile, and 2.00 (95% CI: 1.53 to 2.62; P = 0.0001) for those with levels of both in the highest quartile, as compared with those who were not in either of the highest quartiles. Those with high interleukin-6 and high D-dimer levels had the greatest declines in all measures of function. CONCLUSION: Activation of the coagulation and inflammatory pathways is associated with mortality and decline in function, and may be part of the explanation for the development of a frailty phenotype in the elderly.  

Clinical Global Impression of Change in Physical Frailty: development of a measure based on clinical judgment

Objectives: To expand the ability to assess physical frailty by developing a Clinical Global Impression of Change in Physical Frailty (CGIC-PF) instrument.
Design: Qualitative and quantitative instrument development.
Setting: Academic centers.
Participants: Six expert panel members, 46 clinicians, 24 patients, and 12 caregivers.
Measurements: Literature review and structured group processes with experts, clinicians, and consumers were used to generate an initial list of domains and indicators. Structured interviews with clinical experts in the area of frailty were used to establish relevance and feasibility of measurement of domains. Interrater reliability was assessed through a Web-based study. Geriatricians pilot tested the feasibility of the baseline CGIC-PF with 10 patients.
Results: The CGIC-PF includes six intrinsic domains (mobility, balance, strength, endurance, nutrition, and neuromotor performance) and seven consequences domains (medical complexity, healthcare utilization, appearance, self-perceived health, activities of daily living, emotional status, and social status). Each domain has two to four clinical indicators. Change is scored on a 7-point scale from markedly worse to markedly improved. Average interrater reliability of the CGIC-PF for the Web-based cases was 0.97. Geriatricians completed a baseline CGIC-PF on their own patients in 10 minutes or less.
Conclusion: The CGIC-PF is a structured assessment of change in physical frailty with defined content and process. It has strong face validity, reliability, and feasibility for use in clinical research. It may be useful as one criterion of change and as an anchor for change in other measures.  

Effects of exercise training added to ongoing hormone replacement therapy on bone mineral density in frail elderly women

OBJECTIVES: To determine whether exercise training added to ongoing hormone replacement therapy (HRT) increases bone mineral density (BMD) in physically frail elderly women. DESIGN: Prospective controlled trial. SETTING: University-based research center. PARTICIPANTS: Twenty-eight women on HRT, aged 75 and older with physical frailty. INTERVENTIONS: Participants were assigned to 9 months of supervised (EXER) or home (HOME) exercise. The EXER program started with physical therapy and gradually incorporated resistance and endurance training. The HOME program consisted of flexibility exercises. MEASUREMENTS: Changes in BMD and body composition. RESULTS: There were larger increases in lumbar spine BMD in response to EXER than with HOME (3.5% vs 1.5%, P =.048), with a trend for larger increases in total body BMD (1.5% vs 0.2%, P =.058). There were no significant between-group differences in hip BMD. The EXER group had decreases in weight (-2.2 +/- 0.3 kg, P =.010) and fat mass (-2.7 +/- 0.4 kg, P =.018) and increases in muscle strength (9-30%, P <.05). CONCLUSION: In physically frail elderly women on HRT, relatively vigorous exercise training significantly increased lumbar spine BMD. The improved BMD and strength in response to exercise could reduce fracture risk in frail women already on HRT.  

Mortality Risk Associated with Physical and Verbal Abuse in Women Aged 50 to 79

OBJECTIVES: To investigate whether midlife and older women who reported prior-year physical abuse, verbal abuse, or both abuse types had higher mortality risk than peers who did not report prior-year abuse.
DESIGN: Retrospective analysis.
SETTING: Community.
PARTICIPANTS: One hundred sixty-thousand six hundred seventy-six community-dwelling women ages 50 to 79 at baseline enrolled in one of two major Women's Health Initiative (WHI) study components who responded to baseline abuse questions. Observational study enrollment was N=93,676 (1994–1998; 90 months average follow-up). Clinical trial enrollment was N=68,132 (1993–1998; 96 months average follow-up).
MEASUREMENTS: Total mortality was measured from 1993 to 2005 using all available data sources. Blinded physician adjudicators measured cause-specific mortality. Ninety-six percent of death records were adjudicated.
RESULTS: Prior-year self-reported abuse prevalence was 11.3%. Women who reported physical abuse had the highest age-adjusted mortality rate, followed by women who reported both abuse types. Abuse independently predicted mortality risk after controlling for age, education, ethnicity, and WHI component. High mortality risk remained for physically abused women (hazard ratio (HR)=1.54, 95% confidence interval (CI)=1.09–2.18) after adjusting for demographic and health-related factors. Further adjustment for psychosocial variables diminished this association (HR=1.40, 95% CI=0.93–2.11), but high risk remained.
CONCLUSION: Community-dwelling middle-aged and older women who reported prior-year physical, verbal, or both types of abuse had significantly higher adjusted mortality risk than women who did not report abuse. These findings highlight the need for longitudinal research into prevention of abuse in later life and accompanying excess mortality and emphasize the importance of abuse prevention in later life.  

Nutrition, inflammation, and leptin levels in aging and frailty

OBJECTIVES: To examine nutritional indices and levels of leptin and inflammatory markers across age and frailty.
DESIGN: Observational study.
SETTING: Continuing care wards and a day hospital in Cardiff, South Wales, United Kingdom.
PARTICIPANTS: Thirty dependent patients (mean age 84.9) needing continuing inpatient care, 40 patients with falls attending a day hospital (mean age 84.2), 40 independent controls (mean age 82.7), and 30 young controls (mean age 23.3).
MEASUREMENTS: Functional status, including the five frailty indicators proposed by Fried et al., anthropometry, and serum markers of nutrition and inflammation.
RESULTS: The continuing care patients were frail, all having three to five frailty indicators. Day hospital patients were of intermediate frailty (mean Fried score 2.97), and the independent group was fittest (0.83). Body mass index, triceps skinfold thickness (TSF), and mid-arm muscle area were lowest in continuing care patients. With increasing patient frailty, albumin levels fell significantly ( P <.005) and C-reactive protein (CRP) levels increased significantly ( P <.005). Continuing care patients had significantly lower leptin levels ( P <.005) and significantly higher interleukin (IL)-6 levels ( P <.005). There was a significant correlation between log transformed leptin and TSF for each patient group.
CONCLUSION: The frailest older people displayed features of cachexia. Their leptin levels were appropriately low given their low body fat, and IL-6 and CRP levels were high. The mechanism of their cachexia may therefore be similar to that proposed in heart failure and cancer: disturbed hypothalamic feedback of leptin or effects of proinflammatory cytokines.  

Inflammation and frailty in older women

OBJECTIVES: To evaluate relationships between white blood cell (WBC) count and interleukin-6 (IL-6) and prevalent frailty. DESIGN: Cross-sectional study. SETTING: Two population-based studies, the Women's Health and Aging Studies (WHAS) I and II, Baltimore, Maryland. PARTICIPANTS: Five hundred fifty-eight women aged 65 to 101 from WHAS I and 548 women aged 70 to 79 from the merged WHAS I and II cohorts. MEASUREMENTS: Frailty was determined using validated screening criteria. WBC counts and IL-6 levels were measured using standard laboratory methods. Odds ratios (ORs) for frailty were evaluated across tertiles of baseline WBC counts and IL-6 levels, adjusting for age, race, education, body mass index, and smoking status. RESULTS: In WHAS I, those in the top tertile of WBC count and IL-6 had ORs of 4.25 (95% confidence interval (CI)=1.89-9.58) and 3.98 (95% CI=1.76-9.00), respectively, for frailty (both P<.001). In the combined models, participants in the top tertile of WBC count had an OR of 3.15 (95% CI=1.34-7.41), adjusting for IL-6 (P<.01), and those in the top tertile of IL-6 had an OR of 2.81 (95% CI=1.19-6.64), adjusting for WBC count (P<.05). Furthermore, participants in the top tertiles of WBC count and IL-6 had an OR of 9.85 (95% CI=3.04-31.99), and those in the middle/top tertiles had an OR of 5.40 (95% CI=1.83-15.92) (P<.001, trend test) for frailty. These results were validated in the merged WHAS I and II. CONCLUSION: Higher WBC counts and IL-6 levels were independently associated with prevalent frailty in community-dwelling older women.  

Frailty in older men: prevalence, progression, and relationship with mortality

OBJECTIVES: To describe the association between frailty and health status, the progression of frailty, and the relationship between frailty and mortality in older men. DESIGN: Cross-sectional and prospective cohort study. SETTING: Six U.S. clinical centers. PARTICIPANTS: Five thousand nine hundred ninety-three community-dwelling men aged 65 and older. MEASUREMENTS: Frailty was defined as three or more of the following: sarcopenia (low appendicular skeletal mass adjusted for height and body fat), weakness (grip strength), self-reported exhaustion, low activity level, and slow walking speed. Prefrail men met one or two criteria; robust men had none. Follow-up averaged 4.7 years. RESULTS: At baseline, 240 subjects (4.0%) were frail, 2,395 (40.0%) were prefrail, and 3,358 were robust (56.0%). Frail men were less healthy in most measures of self-reported health than prefrail or robust men. Frailty was somewhat more common in African Americans (6.6%) and Asians (5.8%) than Caucasians (3.8%). At the second visit, men who were frail at baseline tended to remain frail (24.2%) or die (37.1%) or were unable to complete the follow-up visit (26.2%); robust men tended to remain robust (54.4%). Frail men were approximately twice as likely to die as robust men (multivariate hazard ratio (MHR)=2.05, 95% confidence interval (CI)=1.55-2.72). Mortality risk for frail men was greater in all weight categories than for nonfrail men but was highest for normal-weight frail men (MHR=2.39, 95% CI=1.51-3.79, P for interaction=.01). The relationship between frailty and mortality was somewhat stronger in younger men than older men (P for interaction=.01). CONCLUSION: Frailty in older men is associated with poorer health and a greater risk of mortality.  

Improvement and decline in health status from late middle age: modeling age-related changes in deficit accumulation

In a prospective multi-panel cohort study, we investigated how, from late middle age, individuals' health status improves or declines. In the Canadian National Population Health Survey, transition probabilities between different health states were estimated for 4330 people (58.8% women) aged 55+ at baseline over 2-year intervals from 1994 to 2000. Health status was defined by a deficit count, using 33 health-related variables combined in a frailty index. For each time interval, the chance of accumulating deficits increased linearly with the number of deficits. Older survivors (aged 70-85) showed a slightly lower chance of stability or improvement (52%; 95% confidence interval 50-54%) compared with those in late middle age (56%; 54-58%). Changes in health states can be described with high accuracy (R2=0.92) by a modified Poisson distribution, using four parameters: the background odds of accumulating additional deficits, the chance of incurring more or fewer deficits, given the existing number, and the corresponding probabilities of dying. An age-invariant limit to deficit accumulation was observed at 22 deficits. From late middle age, transitions in health states occur with a regularity that is easily modeled. Improvements in health can occur at any age. At all ages, there is a limit to deficit accumulation.  

The value of serum albumin and high-density lipoprotein cholesterol in defining mortality risk in older persons with low serum cholesterol

OBJECTIVES: To investigate the relationship between low cholesterol and mortality in older persons to identify, using information collected at a single point in time, subgroups of persons with low and high mortality risk. DESIGN: Prospective cohort study with a median follow-up period of 4.9 years. SETTINGS: East Boston, Massachusetts; New Haven, Connecticut; and Iowa and Washington counties, Iowa. PARTICIPANTS: Four thousand one hundred twenty-eight participants (64% women) age 70 and older at baseline (mean 78.7 years, range 70-103); 393 (9.5%) had low cholesterol, defined as < or =160 mg/dl. MEASUREMENTS: All-cause mortality and mortality not related to coronary heart disease and ischemic stroke. RESULTS: During the follow-up period there were 1,117 deaths. After adjustment for age and gender, persons with low cholesterol had significantly higher mortality than those with normal and high cholesterol. Among subjects with low cholesterol, those with albumin> 38 g/L had a significant risk reduction compared with those with albumin < or =38 g/L (relative risk (RR) = 0.57; 95% confidence interval (CI) = 0.41-0.79). Within the higher albumin group, high-density lipoprotein cholesterol (HDL-C) level further identified two subgroups of subjects with different risks; participants with HDL-C <47 mg/dl had a 32% risk reduction (RR = 0.68; 95% CI = 0.47-0.99) and those with HDL-C > or =47 mg/dl had a 62% risk reduction (RR = 0.38; 95% CI = 0.20-0.68), compared with the reference category; those with albumin < or =38 g/L and HDL-C <47 mg/dl. CONCLUSIONS: Older persons with low cholesterol constitute a heterogeneous group with regard to health characteristics and mortality risk. Serum albumin and HDL-C can be routinely used in older patients with low cholesterol to distinguish three subgroups with different prognoses: (1) high risk (low albumin), (2) intermediate risk (high albumin and low HDL-C), and (3) low risk (high albumin and high HDL-C).