NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.
Kane RL, Talley KMC, Shamliyan T, et al. Common Syndromes in Older Adults Related to Primary and Secondary Prevention [Internet]. Rockville (MD): Agency for Healthcare Research and Quality (US); 2011 Jul. (Evidence Syntheses/Technology Assessments, No. 87.)
Common Syndromes in Older Adults Related to Primary and Secondary Prevention [Internet].
Show detailsGeriatric syndromes can lead to age-related decline in well-being among elderly adults.1,2 The signs and symptoms encompassed by geriatric syndromes span multiple physiological systems related to functional dependency.3,4 A number of syndromes identified by longitudinal studies are associated with reduced function, quality of life and survival, and an increased risk of institutionalization.5–8 However, variations in syndrome definitions make systematic discussion of their effects difficult.
Routine clinical practice includes assessment of age-related chronic diseases based on accepted diagnostic criteria. In contrast, comprehensive geriatric assessment goes beyond examination for chronic diseases and focuses on functional independence in daily activities and optimal interventions to improve functional status and quality of life.9 Indeed, comprehensive geriatric assessment emphasizes functional status as a major quality of life factor for older adults.10
Quality of life improvements for older adults require addressing geriatric syndromes in addition to managing chronic disease.11 A geriatric syndrome’s definition, along with its combination with any chronic disease, affects the syndrome’s association with patient-centered outcomes, including quality of life, institutionalization, and survival.12–14 Certain factors are long known to affect patient-centered outcomes. For example, the persistently strong association between self-assessed health status and patient-centered outcomes remains a marvel.15 Similarly, dependency, defined as deficiencies in activities of daily living (ADLs), also associates strongly with patient-centered outcomes.16 Systematic reviews have yet to examine other syndromes such as cognitive impairment, frailty, poor nutrition status, or chronic inflammation for prevalence or association with institutionalization and survival.
This review examines what is known about common geriatric syndromes and their effect on the clinical course of older patients. Our analysis examines the extent to which varying definitions of each syndrome can affect determination of its prevalence and its association with patient-centered outcomes. In general, we anticipate a reciprocal relationship; the more inclusive the definition, the higher the prevalence. However, inclusivity should make the variable less predictive of adverse outcomes. More encompassing definitions or those with lower thresholds will inevitably raise prevalence estimates and be less precise in their predictive power than more stringent definitions with higher cut scores. For example, Manton applied ADL- and instrumental activities of daily living (IADL)-related measures for disability to describe a pattern of decline in prevalence over two decades.17 Systematic criteria to define multisystem complex geriatric syndromes are needed.1
Meanwhile, multiple operational definitions of the syndromes presented a challenge to summarizing the research on their prevalence and predictive power. Frailty, especially, persists as an elusive concept, despite efforts at consensus conferences on the topic.18–21 Frailty may be viewed as a specific phenotype or as an index of deficit accumulations.22,23 However, despite problems of definition and measurement, frailty demonstrates a potent association with outcomes. Different indices derived from frailty measures have shown associations with adverse events.24 Likewise, increasing frailty is typically associated with adverse events.25 Frailty and related components (such as ADL dependency, delirium, malnutrition risk, and comorbidity) are linked to increased mortality risk.26 More deficit accumulation is associated with worse outcomes.13 Frailty predicts mortality even after consideration of the effects of clinical and subclinical disease.27 Frailty’s predictive capacity also seems to hold up among various populations in different countries.28
Syndromes are also not independent;definitions and prevalence estimates overlap considerably. For example, sarcopenia is associated with frailty, but some view the former as a dimension of the latter.29–31 Frailty is associated with comorbidity and disability, although efforts to distinguish the latter emphasize frailty’s multisystem dysfunction and instability.32,33 Various geriatric conditions (such as cognitive impairment, falls, and ADL dependency) are associated with disability.34 Polypharmacy may indicate multiple morbidities, but overzealous prescription may also be a factor.35 Research suggests that inflammatory cytokines play a substantial r ole in age-related disease.36 Thus, separating the syndromes presents another challenge.
This report was commissioned by the U.S. Preventive Services Task Force (USPSTF) as background material to help them understand the impact of geriatric syndromes on well-being. The USPSTF opted not to consider disease as a risk factor for the purposes of this review. Our review does not address the suitability of preventing the examined syndromes or altering their courses.
The Technical Expert Panel selected geriatric syndromes (but not diseases) for this review according to how much each syndrome would affect the enthusiasm of clinicians for recommending prevention strategies. We addressed the eight syndromes that were most highly rated.
We included original epidemiologic studies that examined prevalence of the eligible syndromes in adults older than age 65 years. We defined young-old as ages 65–80 years, elderly as ages 80–90 years, and very old as ages 90 years and older. We defined age categories the same as they were defined in the original studies.
We retrieved 2,377 publications and excluded 1,865 that were not eligible for review. We included 509 publications of 123 studies. The majority of the studies were well designed prospective cohorts or national surveys conducted in the United States, including the National Health and Nutrition Examination Survey (NHANES), the National Health Interview Survey, and the National Survey of Self-Care and Aging (76 studies, 62 percent).
Key Question 1. What is the Definition and Prevalence of Common Syndromes/Conditions in Older Adults?
Definitions of a given syndrome vary, and the concepts underlying various syndromes overlap. For example, frailty measures often include disability and comorbidity.
Multiple Morbidities
The studies used a variety of definitions for multiple morbidities, including number of chronic diseases or conditions, high comorbidity score, polypharmacy, or self-perceived poor health. Prevalence estimates varied depending on definitions, with more than 20 percent of older adults suffering from multiple chronic conditions. One-third to one-half of older adults take more than five drugs. One-third of older adults reported fair or poor health.
Cognitive Impairment
Definitions of cognitive impairment varied. The most common definition required subjective complaint of memory impairment with objective memory impairment, normal general cognitive function, and intact cognitive ADLs/IADLs.37,38 Depending on diagnostic method, prevalence (defined as a score of <24 on the Mini-Mental State Examination [MMSE]) varied from 10.6 to 33.1 percent.14,39–45 Studies of the Established Populations for Epidemiologic Studies of the Elderly demonstrated that 7.1 percent of the tested elderly had cognitive impairment detected with the 10-item Short Portable Mental Status Questionnaire.46–48 Prevalence of self-reported cognitive impairment varied from 2.8 to 13.2 percent.49
Prevalence estimates varied substantially in the same study, from 42 percent having self-reported memory complaint to less than 1 percent with mild or moderate cognitive impairment or questionable dementia.50 The variation in prevalence estimates (concordance from 0 to 24 percent) indicates that each definition identified a unique group, and total prevalence of mild cognitive impairment without dementia may exceed 50 percent of community-dwelling older adults.50
Prevalence of dementia did not exceed 8 percent in persons older than age 65 years, but this estimate can be misleading because of substantial variation in prevalence across age subgroups.
Frailty
Wide variances in frailty definitions affects prevalence estimates. Using the framework from the Interventions on Frailty Working Group, we categorized frailty definitions into two groups: phenotype and accumulation of deficits. When the studies accepted the biologic syndrome model of frailty with five major criteria, including weight loss, fatigue and exhaustion, weakness, low physical activity and slowness, and mobility impairment, we categorized the estimates into phenotype definitions.23 When the studies accepted the burden model of frailty, including symptoms, diseases, conditions, and disability, we categorized the estimates into the accumulation of deficits definition.22 Accumulation deficit indices included up to 75 components.12,51,52 Separation of these two definition types and estimation of disability in frail persons were somewhat artificial. Prevalence was higher when using accumulation of deficits (24 percent) than phenotype definitions, such as low physical activity or fatigue (14 percent). The overlap in prevalence estimates using different definitions was small. The Health and Retirement Study examined prevalence of frailty using different definitions, including phenotype and accumulation deficit, and found that 30 percent of elderly people were frail according to at least one definition, but only 3 percent according to all three definitions.
Disability
The most common definitions of disability included having difficulty with or needing assistance with basic activities of daily living (BADLs) or with IADLs. Disability was defined as having “any” (limitations with one or more activities), “moderate” (limitations with one to two activities), or “severe” (three or more activity limitations) disability or having a limitation with an individual activity. Definitions were based on self-reported inability and need for assistance to perform particular tasks. IADL disabilities were more common than BADL disabilities. Reporting of IADL disability ranged in prevalence from 12 to 46.7 percent53–57 compared to 558 to 25.6 percent for BADL disability.41,53,55–64
In general, disability prevalence decreased as severity increased. For example, the prevalence of severe BADL and severe IADL disability was lower than the prevalence for moderate disability, which was lower than having any disability. The order of individual IADLs from most to least common was having limitations with driving, housekeeping, personal finances, shopping, meal preparation, using the telephone, and medication management. For individual BADL disabilities, the hierarchy of most to least common disability was walking, bathing, dressing, transferring, toileting, and eating.65
Sarcopenia
Sarcopenia was defined as a loss of skeletal muscle mass owing to any disease or condition.66 Operational definitions were based on lean body mass relative to skeletal size and total body mass. Sarcopenia was defined based on index values within the sex-specific distribution in a healthy, younger population67,68 or based on a linear regression modeling the relationship between lean mass with fat mass and height. Recently published recommendations from the European Working Group on Sarcopenia in Older People defined sarcopenia as the presence of both low muscle mass and low muscle function (strength or performance).69 Prevalence estimates varied depending on definition from 14 to 60 percent among age, sex, and ethnicity categories. Simple relative skeletal muscle index underestimated sarcopenia in obese older persons. Residual methods adjusting for fat mass may be a better method to identify sarcopenia in overweight and obese older adults.
Malnutrition
Studies defined malnutrition as unintended weight loss70–72 or low body mass index (BMI).46,72–76 Among biochemical markers, low blood albumin levels,73,75,77 anemia,77,78 and deficit of micronutrients79–81 may identify older persons with poor nutritional status. Several studies used composite nutritional scores based on self-reported dietary intake and habits to identify elderly persons with malnutrition.70,72,82–85 Prevalence estimates varied across definitions and were less than 3 percent when defined with low BMI, 6–10 percent with vitamin and micronutrient deficit, and 1–5 percent with a low composite nutritional score. The prevalence of low BMI and blood albumin level was highest in older American veterans (15 percent).
Homeostenosis (Impaired Homeostasis)
Very few studies examined the prevalence of impaired homeostasis in elderly persons. Several studies that examined impaired homeostenosis used allostatic load defined by elevated markers of chronic inflammation, low albumin levels, impaired creatinine clearance, increased blood pressure, hemoglobin A1C, homocysteine, total cholesterol, and triglycerides.86 NHANES defined impaired homeostasis using an allostatic load score and found that 1.4 percent of older adults in the United States had an allostatic load score of more than 4.86 The Duke Established Populations for Epidemiologic Studies of the Elderly defined plasma tonicity as a marker of impaired homeostasis46 and found that 10 percent of older adults had increased plasma tonicity.
Chronic Inflammation
Few studies provided information on the prevalence of unspecified chronic inflammation in older adults. Elevated C-reactive protein (CRP) was found in 24.4 percent of older adults,87 while 5 percent had elevated interleukin 6 (IL6), and 5 percent had elevated tumor necrosis factor-alpha.88
Key Question 2. What is the Prevalence of Common Syndromes/Conditions in Older Adults in Sex, Age, Race, Ethnicity, and Other Subgroups?
Multiple Morbidities
Prevalence of multimorbid conditions increased with age from 28 percent in adults ages 65 to 74 years to 37 percent in adults older than age 75 years. Women tended to have a higher prevalence than men of having more than three comorbidities (16 to 18.4 percent), polypharmacy (43 percent), and poor health (7 percent). Prevalence of more than three chronic diseases was higher in African American women (13.4 percent) than in Caucasian women (9.5 percent).89 Inconsistent definitions of outcomes across the studies made comparison of the estimates difficult.
Cognitive Impairment
Prevalence of cognitive impairment without dementia increased with age across all definitions and studies, from 18.8 percent in adults older than age 75 years90 to 44.1 percent among those older than age 90 years.91 Prevalence of dementia also increased with age across all definitions. Prevalence of senile dementia increased from 1.6 percent in those age 67.5 years to 36.7 percent in those older than age 95 years. Prevalence of Alzheimer’s disease increased from 0.4 percent in those age 67.5 years to 37.4 percent in those older than age 95 years.92,93
Older men and women had comparable prevalence of cognitive impairment. Prevalence of cognitive impairment in older men varied from 16 percent to 36 percent.59,94 No age associated increase was evident. Prevalence of cognitive impairment in older women varied from 10 to 12 percent using the Modified Mini-Mental State Examination (3MSE) questionnaire.59,95
Prevalence was higher (24 percent) when the authors defined cognitive impairment using an MMSE score of <24.42,90,94,96 Prevalence of dementia in older women was consistent across different countries, with an evident increase with age, from 0.25 percent in those ages 65–69 years to 5 percent in those older than age 80 years.97
About 20 to 40 percent of older persons with cognitive impairment developed dementia within 2 to 5 years of followup.98–100 Several studies demonstrated that older persons with cognitive impairment were at higher risk of developing dementia.
Frailty
Prevalence of frailty increased with age but differed within age subgroups depending on definitions. Prevalence of frailty in those ages 65 to 70 years ranged from 3 to 6 percent, using a phenotype definition, to 5 to 15 percent using an accumulation deficit definition.23,101–104 Prevalence among those ages 70 to 80 years varied from 5 to 12 percent, according to a phenotype definition, to 8 to 17 percent according to accumulated deficits.22,23,101,104,105 Prevalence was more than 16 percent in those older than age 80 years according to any definition.23,101–104
Among race groups, African Americans had the highest prevalence of frailty across different definitions. More than half of older African Americans were frail according to two studies.23,104,106,107 Two studies examined the prevalence of frailty in older Hispanics and reported that 8 to 20 percent met different frailty criteria.104,108 Prevalence of frailty in older Caucasians varied from 6 to 12 percent, using a phenotype definition, to 15 to 40 percent using an accumulation deficit definition.23,104,107,109 The large cohort studies of predominantly older Caucasians in the Survey of Health, Aging and Retirement in Europe demonstrated that 17 percent were frail according to phenotype criteria.110
Among sex groups, prevalence of frailty was somewhat higher in women than in men, and increased with age in both sexes. Prevalence of phenotype frailty was 7 percent in older men, and accumulation deficit frailty was 24 percent.23,103–105,107,109,111–115
Prevalence was higher in aging African American and Hispanic men compared to Asian or Caucasian men.23,113,116 Prevalence of phenotype frailty in older women was 13 percent, and accumulation deficit frailty was 26 percent.23,24,103–105,107,109,111,115–117 African American women had the highest prevalence of frailty, with 60 percent of adults older than age 85 years being frail.23,116
Disability
In general, women had higher rates of BADL disability than men. However, as the severity of disability increased, the prevalence for women and men became similar (i.e., 7.0 percent for severe BADL disability118 and 1.2 percent for eating disability119,120). Few studies reported sex differences in IADL disabilities. The prevalence of any IADL disability was higher in women than men.121 One study described changes in the prevalence of any IADL disability over 6 years between men and women.121 More women reported an IADL disability than men at any time period, but showed less change over time. In terms of individual IADL disabilities, more women had difficulty with housekeeping and meal preparation than men but less difficulty with shopping.119,122
Only two studies reported the prevalence of any BADL disability by ethnic group, and one study enrolled only older Hispanic Americans.41,60,62 The prevalence of any BADL disability, in order of highest to lowest, was African Americans (13.6 percent), American Indians (11.6 percent), Hispanic Americans (11 percent),41 and Caucasians (8.1 percent).62 Racial differences persisted after accounting for sex. Older African American women had the highest prevalence of having any BADL disability (10.7 percent), followed by African American men (7.5 percent), Caucasian women (5.2 percent), and Caucasian men (4.7 percent).60 There were no racial differences in eating disabilities.119
Two studies reported age differences in the prevalence of BADL disability.59,118 Reporting any disability and moderate disability was more prevalent in the oldest age groups and in older women. The prevalence of severe BADL disability ranged from 10 to 11 percent in people ages 80 years and older, was 6 percent in those ages 65 to 74 years, and did not differ significantly by sex.118
Sarcopenia
Prevalence of sarcopenia increased with age.68,123 Older African Americans had significantly lower odds of sarcopenia when compared to older Caucasian Americans.123 Odds of sarcopenia did not differ between Hispanic and non-Hispanic whites.68
Malnutrition
Age and sex differences in malnutrition were not consistent across the studies. Pooled prevalence of poor nutritional score was 18.3 percent in older men and 24 percent in older women. Women had a lower prevalence of decreased albumin levels77 but higher prevalence of low BMI.
Older African Americans had a significantly higher risk of malnutrition, defined as unintentional weight loss, compared to older Caucasian persons.71 Prevalence of anemia did not differ among Caucasians and non-Caucasians.78 Prevalence of unintentional weight loss did not differ in Hispanics and non-Hispanics.72 Older Hispanic women had a higher prevalence of poor nutritional scores compared to non-Hispanic women (30 percent vs. 17 percent, respectively).72
The studies did not report prevalence of homeostenosis and chronic inflammation in older subpopulations.
Key Question 3. What is the Association Between These Common Syndromes/Conditions and Mortality, Institutionalization, Hospitalization, and Activities of Daily Living?
We analyzed the association between outcomes and each syndrome and across all syndromes to identify links to mortality. Estimates of association varied depending on definitions of comorbidities, population subgroups, definitions of outcomes, and adjustment for correlated contributing factors. Not all analyses addressed the multifactorial nature of geriatric syndromes and the role of baseline diseases. For example, disability was an outcome but also a part of the definition of frailty. Adjustment for correlated multifactorial syndromes that ignored definitive primary cause of disability or death may give invalid estimation of the association between syndromes and mortality. No studies separately examined age and specific disease contributions.
Multiple Morbidities
We observed a consistent and significant positive association between multiple morbidities and mortality across the studies. Older persons with multiple morbidities had a 32 to 112 percent relative increase in death compared to those without multiple morbidities.39,124–129 The magnitude of the association was dose responsive, with an 85 percent relative increase in mortality for those with four to five diseases and 112 percent among those with six or more chronic conditions.39,124–129 The magnitude of the association decreased with time of followup, from a 100 percent relative increase at 10 years (odds ratio [OR], 2 [95% confidence interval (CI), 1.4–2.8]) to a 59 percent increase at 15 years of followup (OR, 1.6 [95% CI, 1.1–2.3]).125 However, this may be a statistical reflection of a greater denominator as outcomes accrue over time.
Polypharmacy was significantly associated with mortality in two studies, with evidence of a dose response.54,130 Those with poor health had an increased risk of death in all studies that examined the association.27,103,131
The Longitudinal Study of Aging demonstrated a positive significant association between multiple morbidities and institutionalization.128 Those with poor health had a 10 to 80 percent relative increase in institutionalization.8,103,132
Older adults with multiple morbidities had increased odds of hospitalization (OR, 1.7 [95% CI, 1.1–2.9]).133 The association with hospitalization was dose responsive.126 The relative increase in odds of hospitalization was 37 percent in those with morbidity scores of 3 versus ≤2, 46 percent in those with scores of 4–5 versus ≤2, and 94 percent in those with scores of ≥6 versus ≤2.126
Polypharmacy was significantly associated with hospitalization.133,134 The Medicare Risk Demonstration cohort reported a 190 percent increase (OR, 2.9 [95% CI, 2.2–4.1) in odds of hospitalization among those with more than five prescriptions compared to those with fewer than five concurrent drugs.133 The Longitudinal Study of Aging reported a significant increase in risk of hospitalization among older adults with poor health.135–140
We concluded that multimorbid conditions and poor perceived health demonstrated a strong association with mortality. Poor perceived health was a strong predictor of institutionalization. The number of chronic conditions, polypharmacy, and poor perceived health was associated with hospitalization.
Cognitive Impairment
Cognitive impairment was associated with a significantly higher risk of mortality in all studies that examined this association. The largest relative increase of 250 percent in women and 280 percent in men was found in the Canadian Study of Health and Aging, which defined cognitive impairment as a score of <78 on the 3MSE scale.141,142 There was dose response association, with a 4 percent relative increase in mortality for each decrease by 1 point on the MMSE.39,44,54,143–148 The studies that estimated relative risk ratio (RR) or hazard rate ratio (HR) found a 37 percent39,54 and a 61 percent44,147 relative increase in risk of death, respectively.
Older women (pooled RR, 1.4 [95% CI, 1.11–1.7]) but not men (pooled RR, 1.2 [95% CI, 0.8–1.8]) with MMSE scores of <24 had a significant risk of death.39,94 Both men and women with severe cognitive impairment, defined as an MMSE score of <18, were at higher risk of death.
Dementia was associated with a significantly higher risk of mortality in the majority of the studies that examined this association.149 Overall, dementia was associated with a 163 percent relative increase in odds of death (pooled OR, 2.6 [95% CI, 2.2–3.2]).
Cognitive impairment was associated with a significant risk of institutionalization in the majority of the studies that examined this association. For cognitive function measured with MMSE, the association was dose responsive and significant even within the normal ranges of the scale. Older persons had a higher relative increase in institutionalization of 9 percent per each 1-point decrease on the MMSE.145 Those with dementia were at a significant risk of institutionalization in several large studies, including the Medicare Current Beneficiary Survey (OR, 34.9),150 the Canadian Study of Health and Aging (OR, 36.3),151 and the Marshfield Epidemiologic Study Area (HR, 5.1).152
We concluded that older persons with cognitive impairment had a higher risk of mortality and institutionalization. Magnitude of the association varied depending on the country, age, and sex of the participants, definitions of the cognitive impairment, and statistical estimates.
Frailty
Frailty was associated with mortality across a number of studies with varying definitions.14,23,111,115,148,153–156 The strength of the association was cumulative157 and dose responsive,14,156 with a greater risk among those with increasing numbers of frailty components.14,156 The association generally persisted over longer followup periods.23,156
The association was significant in men and women. Frail men had a relatively greater risk of death of 105 to 251 percent according to phenotype definitions and of 65 to 356 percent according to accumulation deficit definitions.113,114,158,159 Frail women had increased mortality across different studies and definitions of frailty.158–160
Frailty was associated with an increased risk of institutionalization155,160 and hospitalization.14,28,161 The studies demonstrated a 29 percent relative increase in risk161 and a 41 to 345 percent relative risk14,161 in odds of hospitalization.
Disability
Disability and hospitalization. The statistically significant association between disability and hospitalization was demonstrated in four studies (adjusted relative measures of association ranged from 1.8 to 16.0).136,137,162,163 Risk of hospitalization depended on the definition of disability, and, in general, risk increased along with the severity of disability. Older people who had any BADL (defined as having one or more BADL dependencies) had the lowest risk of hospitalization. Those with severe BADL disability (defined as having three or more BADL dependencies) had the highest risk. The risk for women was also greater than the risk for men. The manner in which BADL disability develops appears to influence the risk for hospitalization. Older people who experienced catastrophic severe disability (defined as the sudden onset of three or more BADL disabilities when no BADL disabilities existed before) were 16 times more likely to be hospitalized than those who had moderate BADL disability.163 When ADL disability was measured on a continuous scale, the risk for hospitalization was not statistically significant, nor was it significant when measured as having any BADL or IADL dependencies.
Disability and risk of death. In general, older people with BADL disabilities were at higher risk for death (OR range, 1.9–86.8) than those with IADL disabilities (OR range, 1.5–6.6) when compared to those without disability. Those with more BADL disabilities had a higher risk of death than those with fewer BADL disabilities. Severe BADL disabilities were associated with the highest risk of death,128,164 followed by moderate BADL disabilities.128,164 The lowest risk of death occurred when any BADL disability was reported.162 The risk of death associated with individual BADL disabilities was not reported in the studies, with one exception. The risk for death at 48 months doubled for older people with bathing disabilities.53 The risk of death associated with IADL disabilities was highest when any IADL disability was reported.128,164 Those with severe IADL disability had slightly higher risks of death (OR range, 1.6–2.2) than those with moderate IADL disabilities (OR range, 1.5–1.7). Those with difficulty managing personal finances were twice as likely to die as those without this disability.53 When disability was measured on a continuous scale, the per one-point increase in the disability score and the risk of death were the same whether the scale measured BADL, IADL, or BADL/IADL disability. Men with BADL disabilities had slightly higher risks of death than women.64 Caucasian men and women who were unable to prepare a meal had higher risks of death than African American men and women.119 Women suffer greater discrepancies in years of expected active life remaining than men if they have a BADL/IADL disability.40 In summary, older people with the most severe BADL disabilities had the highest risks of death. No studies examined how individual BADL and IADL disabilities increase the risk of death. Women fared worse than men in terms of expected active life remaining, but men with BADL disabilities had slightly higher death rates than women with BADL disabilities. Older Caucasian people were more likely to die than older African American people if they reported difficulty with preparing meals. Few studies reported differences in death rates between men and women or between older people of different ethnicities who have disabilities.
Sarcopenia
Limited evidence indicates that sarcopenia was associated with significantly higher odds of multiple disabilities68 but not mortality.165
Malnutrition
The association between malnutrition and mortality was consistent across the studies and different definitions of malnutrition.70,71,83,166–169 Low BMI166,167 and malnutrition identified using the Mini Nutritional Assessment70 were the strongest predictors for mortality.
Among biological markers that may be related to malnutrition, red cell distribution width (RDW) demonstrated strong and significant association with mortality in all examined age, sex, and race subgroups in a meta-analysis of individual subject data from seven community-based studies of 11,827 older adults.79 Red cell distribution was associated with mortality, however, in those with iron, folate, and/or vitamin B12 deficiencies (adjusted HR for 1 percent increment in RDW, 1.2 [95% CI, 1.1–1.2]), as well as in those without these deficiencies (adjusted HR for 1 percent increment in RDW, 1.2 [95% CI, 1.2–1.3]).79 Routinely measured as a part of the complete blood count, a red cell distribution width of >15 percent was associated with a 151 percent relative increase in risk of death (HR, 2.5 [95% CI, 2.2–2.9]).79 Very low albumin levels and very high pre-albumin levels (transthyretin >316mg/L or <258mg/L) were associated with increased mortality.167,170
Several studies analyzing composite measures of malnutrition and chronic inflammation found a significant positive association with mortality. Those with the highest levels of alpha-1-acid glycoprotein and the lowest levels of transthyretin had the highest risk of death, with a 364 percent relative increase in women and a 586 percent relative increase in men.170 Elevated composite measure of chronic inflammation and poor nutritional status was associated with an increased risk of death in older men but not women.170
Homeostenosis (Impaired Homeostasis)
Individual studies demonstrated a significant association between disability, mortality, and indicators of impaired homeostasis. The association of impaired homeostasis with clinical outcomes varied depending on the definitions of the exposure and population studied. Unstable BMI (HR, 1.3 [95% CI, 1.0–1.8]), pulse pressure (HR, 1.3 [95% CI, 1.0–1.7]), and fasting plasma glucose (HR, 1.6 [95% CI, 1.2–2.1]) were associated with a greater risk of mortality in an Italian cohort.171 The MacArthur Studies of Successful Aging demonstrated a significant association between increased allostatic load and mortality.124,172,173 The same study found a 27 percent relative increase in the odds of death (OR, 1.3 [95% CI, 1.0–1.5]) in those with an elevated stress hormone index.124 Consensus around the operational definition of homeostenosis and its biomarkers is necessary for a better interpretation of the results.
Chronic Inflammation
Studies consistently found positive significant associations between chronic inflammation and mortality. Among common definitions of chronic inflammation, elevated IL6 and CRP were associated with higher mortality. Those with elevated IL6 levels had a 42 percent relative increase in death (pooled RR, 1.4 [95% CI, 1.2–1.7]).124,174–177 Those with elevated CRP had a 42 percent relative increase in death (pooled RR, 1.4 [95% CI,1.3–1.6]).87,124,129,166,167,170,174,177–179
Among other individual markers of chronic inflammation, the strongest association with mortality was demonstrated for a combination of elevated CRP with low albumin (HR, 5.0 [95% CI, 2.3–11.0])170 or with elevated fibrinogen levels (HR, 9.56 [95% CI, 4.34–21.1]).178
We found no studies that examined the association between chronic inflammation and institutionalization or hospitalization.
Key Question 4. What Statistical and Decisionmaking Models Report Mortality Based on These Common Geriatric Syndromes/Conditions?
Models reporting mortality vary by complexity, by selection of predictors, and by time course. We found 28 studies that described prognostic indices for mortality in older adults. Previous indices are complex, time consuming, or have a lack of clinical applicability. However, recent studies have been designed to develop and validate easy-to-use indices using information readily available from administrative data, laboratory data, diagnoses, or self-reported health status data. Some indices were created for certain segments of the population (e.g., hospitalized elderly,26,180–183 community dwellers,4,23,27,148,155,184–188 or older individuals with acquired mental disorders).189 Others have been developed with the use of nationally representative samples.53,186,187,190 While the majority of studies have been conducted in the U.S. population ages 50 years and older, a few have been done in European countries14,134,180,189,191,192 and Canada.155,193,194
To examine overall effects of different syndromes on mortality in adults older than age 65 years, we estimated numbers of deaths per 1,000 from individual studies that provided death rates among persons with and without different syndromes (Table 1).195 We estimated that among frail older persons, 459 per 1,000 died within 1–2 years of followup.182,187 Disability in basic ADLs and IADLs were the strongest predictor of mortality.196
Table 1
Mortality Among Older Persons With Geriatric Syndromes.
Within 3 years, 500–600 older persons with malnutrition, 351 with cognitive impairment, and 534 with severe dementia died per 1,000 older persons.189,197,198
Within 5 years, 490 elderly persons with malnutrition, 513 with frailty, 530 with elevated CRP, and 827–941 with vascular dementia died per 1,000 older persons.70,91,141,186 Frailty and cognitive impairment were associated with 400–800 deaths per 1,000 during more than 5 years of followup.23,148,199 Such estimations may not reflect mortality in age, sex, or race subgroups but demonstrate a burden of geriatric syndromes.
We also estimated population risk of mortality attributable to geriatric syndromes (Table 2). When population prevalence and multivariate adjusted relative risks were taken into account, more than 7 percent of deaths were attributable to multiple morbidities and elevated CRP. We estimated that 3 to 5 percent of deaths among older persons could be delayed by preventing frailty; prevention of mild cognitive impairment could delay 5 to 6 percent of deaths. Overall, around 26 percent of deaths in older persons can be attributed to geriatric syndromes. Conversely, having these syndromes affects the likelihood of benefitting from preventive interventions.
Table 2
Population Risk of Mortality Attributable to Geriatric Syndromes, Sorted in Descending Order of Population Attributable Risk.
The prevalence and risk of mortality and institutionalization were almost inversely related. The prevalence of accumulation deficit frailty (which uses many components) was higher than phenotype frailty (which uses only a few components). The relative risk of mortality and institutionalization was higher for phenotype frailty (Figure 1). The same negative association was seen for more severe forms of the same syndrome. Prevalence of severe cognitive impairment and dementia were lower, but risk of mortality was higher when compared to mild cognitive impairment (Figure 2). A negative association between the prevalence of a syndrome and its effect on mortality was evident across those syndromes in which the more restricted definition defines a more severe state (Figure 3).
We estimated remaining life expectancy in individuals with syndromes using Centers for Disease Control and Prevention United States Life Tables and the relative risk of death from pooled analyses and individual studies. Increased levels of allostatic load (impaired homeostasis) and dementia were associated with the lowest survival among older persons when compared to the general U.S. population. The data shown in Figure 4 represent a merger of several data sets to yield general trends. The influence on survival of some factors is much greater than others. Poor health, malnutrition, and allostatic load (poor homeostasis) exert twice the influence of factors such as comorbidity and frailty. The size of the effect differs by age (and thus expected life expectancy) (Table 3). Relative risk is likely more useful than population attributable risk. In the young-old, ages 65–74 years, only the very few who are very ill (e.g., homeostenosis, poor health, or advanced dementia) or frail suffer significant alterations in predicted life expectancy. From ages 75–90 years, maximal heterogeneity of disease and geriatric syndromic states result in larger mortality deviations from unafflicted individuals than seen in other age groups. In the old-old, particularly past age 90 years, the added value of factoring in conditions and syndromes to predict mortality beyond 1 year is minimal.
Table 3
Differences in Remaining Life Expectancy Between Older Persons From the General Population and Older Persons With Geriatric Syndromes.
Models reporting mortality vary by complexity, by selection of predictors, and by time course. Some models strive for simplicity, with few predictors that are easily measured, much or all of which could be gained by culling administrative data. Other much more complex models rely on data gathered from clinician and/or patient assessments. For purposes of anticipating the benefit of preventive services, a simpler approach, based on some crude classifications around average life expectancy (based on age and sex), serves better than more complex models.
Geriatrics teaches that age is a good general predictor, but great care is needed to look within older people to distinguish other risk factors. In this case, the evidence for added benefit from factoring in syndromic information is mixed; syndromic information is helpful for younger old persons, but adds little insight for the very old. The most potent predictors are the rarest syndromes. Absolute risk differences and remaining life expectancy in comparison groups should be taken into consideration when analyzing predictive value of syndromes in different age subgroups.
Some basic relationships hold regardless of the measure used. They can be summarized as follows:
- Simple disease-based measures, such as number of comorbid illnesses or measures of inflammation, add modestly to the relative risk of mortality provided by age and sex alone but account for a more population-based mortality burden due to their high prevalence.
- Advanced dementia is one specific condition that confers significantly added mortality risk.
- More complex syndromic measures, such as those assessing frailty or incorporating functional status (e.g., allostatic load and dementia), better capture increased mortality risk (indicated by higher relative risk) than simpler measures, as they more selectively identify the relatively few (indicated by lower population attributable risk) sickest patients most likely to experience deterioration in health and death.
- Simpler measures that reflect the severity of individual diseases, such as indicators of advanced dementia, or the overall impact of multiple conditions, such as assessments of overall health, also identify the fewer and sicker patients at higher risk of mortality.
In conclusion, complex mortality models add comparatively little understanding to more simply measured and calculated models. Measures of the impact of conditions and syndromes on overall health and functioning provide greater discrimination among individual patients for assessing mortality risk. Mortality predictors appear to be relatively consistent across short- and long-range models. The greatest added advantage of mortality models over simple remaining life expectancy was observed among patients ages 75 to 90 years. No models considered psychosocial factors, such as resilience, or the role and quality of health care for elderly patients with syndromes. Decisionmaking models are based on various assumptions and simulation techniques that need careful sensitivity analysis and validation.
Our review does not address the extent to which the presence of a syndrome adds predictive power over and above the presence of specific diseases. It seems likely that the syndromes represent intermediate states between the disease and mortality, but their specific additive explanatory power remains unknown. For clinicians, the syndromes offer summative approaches that can help in some instances to improve the estimate of the risk of mortality.
Ideally, we would consider other outcomes besides mortality, but the measures used present large problems of endogeneity. Measures of frailty and disability contain elements central to quality of life. They also frequently provide the basis for institutionalization.
All syndromes had overlapping definitions or closely related pathophysiology. The multifactorial interactive nature of syndromes should be analyzed with interaction models rather than adjustment. The majority of the studies, however, provided multivariate adjustment for known confounding factors, causes of death, and presence of other syndromes. The models that analyzed the association between syndromes and mortality grouped primary causes of death into larger categories of cancer or cardiovascular diseases to adjust for them. The syndromes associated with decompensated chronic diseases, including inability to maintain homeostasis, poor general health, or low BMI, had the strongest association with mortality and institutionalization.
Evidence suggests that, despite differences in definitions, common geriatric syndromes can be examined and constructed from a variety of different measures. By almost all definitions used, evidence suggests a considerable disease burden as population age increases. Evidence about how geriatric syndromes may modify the efficacy of preventive or other interventions is needed but was outside of our scope. Future research should examine effectiveness of screening, preventive treatment, and disease management strategies in elderly adults with common geriatric syndromes.
- What is the Definition and Prevalence of Common Syndromes/Conditions in Older Adults?
- What is the Prevalence of Common Syndromes/Conditions in Older Adults in Sex, Age, Race, Ethnicity, and Other Subgroups?
- What is the Association Between These Common Syndromes/Conditions and Mortality, Institutionalization, Hospitalization, and Activities of Daily Living?
- What Statistical and Decisionmaking Models Report Mortality Based on These Common Geriatric Syndromes/Conditions?
- Executive Summary - Common Syndromes in Older Adults Related to Primary and Seco...Executive Summary - Common Syndromes in Older Adults Related to Primary and Secondary Prevention
Your browsing activity is empty.
Activity recording is turned off.
See more...