Customize your JAMA Network experience by selecting one or more topics from the list below.
Walston J, McBurnie MA, Newman A, et al. Frailty and Activation of the Inflammation and Coagulation Systems With and Without Clinical ComorbiditiesResults From the Cardiovascular Health Study. Arch Intern Med. 2002;162(20):2333–2341. doi:10.1001/archinte.162.20.2333
Copyright 2002 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2002
The biological basis of frailty has been difficult to establish owing to the lack of a standard definition, its complexity, and its frequent coexistence with illness.
To establish the biological correlates of frailty in the presence and absence of concurrent cardiovascular disease and diabetes mellitus.
Participants were 4735 community-dwelling adults 65 years and older. Frail, intermediate, and nonfrail subjects were identified by a validated screening tool and exclusion criteria. Bivariate relationships between frailty level and physiological measures were evaluated by Pearson χ2 tests for categorical variables and analysis of variance F tests for continuous variables. Multinomial logistic regression was performed to evaluate multivariable relationships between frailty status and physiological measures.
Of 4735 Cardiovascular Health Study participants, 299 (6.3%) were identified as frail, 2147 (45.3%) as intermediate, and 2289 (48.3%) as not frail. Frail vs nonfrail participants had increased mean ± SD levels of C-reactive protein (5.5 ± 9.8 vs 2.7 ± 4.0 mg/L), factor VIII (13 790 ± 4480 vs 11 860 ± 3460 mg/dL), and, in a smaller subset, D dimer (647 ± 1033 vs 224 ± 258 ng/mL) (P≤.001 for all, χ2 test for trend). These differences persisted when individuals with cardiovascular disease and diabetes were excluded and after adjustment for age, sex, and race.
These findings support the hypothesis that there is a specific physiological basis to the geriatric syndrome of frailty that is characterized in part by increased inflammation and elevated markers of blood clotting and that these physiological differences persist when those with diabetes and cardiovascular disease are excluded.
THE GERIATRIC syndrome of frailty has been defined as a wasting syndrome and a physiological state of vulnerability to increased morbidity and mortality.1,2 Frailty exacts an enormous toll on the health and well-being of older adults. Frail individuals are at much higher risk for falls, fractures, infections, the development of disabilities, hospitalization, institutionalization, and death compared with their age-matched nonfrail counterparts.2,3 Although frailty is associated with advanced age and increased disability, evidence suggests that neither old age nor disability alone identifies those at highest risk of adverse outcomes.3,4 In fact, frailty is predictive of adverse outcomes independent of these factors, suggesting an independent physiological etiology.3,5
Though postulated to have a distinct biological basis, the physiological changes underlying frailty have been difficult to characterize because of the complexity of these changes, the frequent coexistence with acute and chronic illness and with disability and the lack of specificity of the definition of frailty.6 Buchner and Wagner1 have suggested that alterations in several physiological systems, including decreased musculoskeletal functioning, neurologic control, and lower energy metabolism, in the presence of a physiological trigger influence the development of this disorder. Roubenoff and Rall7 have proposed physiological alterations in immune and endocrine systems that in turn have a profound impact on body composition. We have previously hypothesized a dynamic model of frailty in which inflammation, neuroendocrine dysregulation, and sarcopenia contribute to a spiraling decline in physiological processes and function.8
The lack of a standardized definition to identify frail older adults has slowed the study of the physiological basis of frailty, as has the inclusion of illness and disability in previous definitions.6,9 Frail older adults are most often recognized by a constellation of signs and symptoms including weight loss, muscle weakness, fatigue, declines in activity, and slow or unsteady gait.8,10 To arrive at a standardized screening tool for frailty, we previously synthesized the most commonly described attributes of frailty into definable clinical criteria. These screening criteria consist of weight loss, muscle weakness, fatigue, declines in activity, and slow or unsteady gait.3 We have validated this clinical criteria by showing its ability to predict disability, hospitalization, and mortality among participants of the Cardiovascular Health Study (CHS), a cohort study of 5888 community-dwelling persons 65 years and older at study inception.3,11
In the present study, we sought to explore the physiological correlates of frailty using this newly validated clinical criteria of frailty. In particular, we sought to assess the association of frailty with biomarkers previously found to be associated with other wasting syndromes and with components of several physiological systems hypothesized to be related to frailty, including inflammation and markers of metabolic processes. Because several clotting factors respond directly to inflammatory mediators, we also sought to define relationships between frailty and markers of clotting process in a subset of older adults without cardiovascular disease (CVD). Finally, because of the frequent use of lipid and albumin levels as markers of frailty in older adults and controversies over whether alterations in these markers are primary or secondary, we evaluated the relationship of the frailty clinical criteria to these biomarkers.12,13 We chose the CHS cohort for this study because the frailty examination was previously validated in this group, because of the well-characterized disease states of the subjects, and because of the large number of frail older adults who were part of this cohort.3,11
The CHS is a population-based, longitudinal study of adults 65 years and older that began data collection in 1989. The study was designed to identify the risk factors and outcomes of CVD in older adults.11 The cohort was sampled from Medicare eligibility lists of community-dwelling men and women from 4 geographic areas in the United States including Allegheny County, Pennsylvania; Forsyth County, North Carolina; Washington County, Maryland; and Sacramento County, California. Sampled subjects and age-eligible household members were invited to participate. Exclusion criteria included undergoing active treatment for a malignancy, being wheelchair bound in the home, likelihood of moving out of the area in the coming 3 years, and being unable to come for a baseline examination in a field center. Data presented here are drawn from the baseline evaluation of the original cohort conducted in 1989-1990 and included weight, timed walking speed over a distance of 15 ft (4.5 m) at usual pace, and maximal grip strength in dominant hand as measured by a hand held JAMAR dynamometer (Sammons Preston AbilityOne Corporation, Bolingbrook, Ill). The Center for Epidemiologic Studies Depression (CES-D) Scale was administered.14 Physical activity levels were assessed using the Modified Minnesota Leisure Time Activities questionnaire.15
Cardiovascular disease was defined as the validated and confirmed presence of any of the following 6 conditions: angina, myocardial infarction, congestive heart failure, transient ischemic attack, stroke, and intermittent claudication.11 Diabetes mellitus was defined according to the World Health Organization guidelines in existence at the start of the CHS (ie, fasting glucose level ≥140 mg/dL [7.8 mmol/L] or 2-hour glucose level after 75-g oral glucose tolerance test ≥200 mg/dL [11.1 mmol/L] and those with a previous diagnosis of diabetes currently receiving therapy with insulin or oral agents). To investigate the potential confounders of CVD on frailty, a subset of 400 individuals who were free of CVD at baseline was selected from the entire cohort for further analysis of blood proteins important in the clotting process.16 Hypertension was defined by a systolic blood pressure greater than 160 mm Hg or a diastolic blood pressure greater than 95 mm Hg (definition used at start of data collection) or a medical history of hypertension and currently receiving therapy with antihypertensive medications.
A 5-component screening examination to identify frail individuals was developed and validated by us based on observations of frail patients in a number of clinical settings, physiological hypothesis, and literature review to identify consensus in regard to characteristics of frailty and outcomes.3 The complex methodology and validation of this examination is described elsewhere.3 The components of this examination were part of a broader functional assessment examination performed at intake into the study. To be declared "frail" by these criteria, subjects had 3, 4, or 5 components present, consistent with the concept of frailty as a syndrome. Intermediate subjects met 1 or 2 of these criteria. To be declared "not frail," subjects had 0 positive frailty criteria. The screening tool consisted of the following measures:
Grip Strength. Grip strength was measured in the dominant hand using a JAMAR handheld dynamometer set at level 2. Three attempts at maximal squeeze were recorded. The average value was adjusted by the individual's body mass index (BMI). Those with grip strength in the bottom 20% were considered positive for this frailty criterion.
Walking Speed. Individuals were asked to walk 15 ft (4.5 m) at their usual pace, following a standardized protocol; and time was measured by a trained examiner. Height- and sex-adjusted time points were used, with the slowest 20% being considered positive for frailty by this criterion.
Weight Loss. Individuals reporting more than 10 lb (45 kg) of unintentional weight loss in the previous year were deemed positive by this criterion.
Exhaustion. Two items from the CES-D Scale were used to characterize exhaustion.14 These were (1) "I felt that everything I do is an effort" and (2) "I cannot get going." Individuals were asked to indicate if they felt that way 0 (none of the time), 1 (some of the time [1-2 days a week]), 2 (a moderate amount of time [3-4 days]), or 3 (most of the time). Subjects answering 2 or 3 to either of these questions were positive for the exhaustion criterion.
Physical Activity. This part of the screening was based on the Modified Minnesota Leisure Time Activities questionnaire and involved self-report regarding whether a person performed any of 18 activities in the prior week, along with the frequency and duration of these activities.15 Kilocalories of energy expended in a week on leisure time activity were calculated. Those in the bottom quartile of physical activity were deemed positive for this frailty criterion. Cut points were stratified by sex. Men exerting less than 383 kcal (1602.5 kJ) per week or women exerting less than 270 kcal (1129.7 kJ) per week in these activities were positive for this criterion.
At each CHS field center, baseline blood samples were drawn after an overnight fast, processed for storage, and shipped to a central laboratory at the University of Vermont (Colchester) following standardized protocols. Methods of phlebotomy sample handing and quality assurance have been described previously.17 White blood cells, hemoglobin, and hematocrit were measured at local laboratories. Glucose, insulin, albumin, fibrinogen, factor VII, factor VIII, and plasma lipids were measured at the central laboratory. A 75-g oral glucose tolerance test was performed on all subjects without diabetes, and fasting and 2-hour glucose and insulin levels were measured. Coagulation factors were measured with the Coag-A-Mate X2 instrument (Organon Teknika Corp, Durham, NC) using immunodeficient plasma (Baxter Dade AG, Dudingen, Switzerland) and human placental thromboplastin (Thromborel S; Dade Behring Inc, Deerfield, Ill) for factor VII and a partial thromboplastin reagent for factor VIII (Organon Teknika Corp). Values were reported as a percentage of normal plasma pool, and standardization was performed by assaying reference plasma from the World Health Organization. The mean monthly coefficients for variation for the factor VII and VIII assays were 5.31% and 9.67%, respectively. Plasma fibrinogen was measured using a semiautomated modified clot-rate method with a BBL Fibrometer (Becton Dickinson and Company, Bedford, Mass) and are reported in milligrams per deciliters. The mean monthly coefficient of variation for the fibrinogen assay was 3.09%. Albumin was assessed using the Kodak Ektachem 700 analyzer (Eastman Kodak, Rochester, NY) as part of the standard clinical chemistry assays.17 White blood cells were assessed at each of the 4 local CHS laboratories using automated counters (Coulter Stack S cell counter [Beckman Coulter Inc, Fullerton, Calif] or the SysmexNE8000 counter [Toa Electronics Inc, South San Francisco, Calif]).11 C-reactive protein (CRP) was assessed with a high-sensitivity enzyme-linked immunosorbent assay using purified protein and polyclonal anticlonal anti-CRP antibodies.18 The interassay coefficient of variation was 5.50%. The laboratory measurements for markers of clotting process in the subset without CVD has been previously published.16
Because of the potential overlap in manifestations of frailty with other medical conditions common in advanced age, all individuals with Parkinson disease, stroke, depression under treatment, and dementia with a Folstein Mini-Mental State Examination score less than 18 were excluded from these analyses, leaving 4735 subjects available for further study. To determine if physiological alterations were associated with the syndrome of frailty independent of disease status, we performed 2 analyses. The first used only the above-stated exclusion criteria because these criteria may cause a subject to appear frail when he or she is not. The second used these exclusion criteria but also excluded those participants with CVD and diabetes because these diseases may cause alterations in inflammatory markers, glucose metabolism, and blood clotting. Data are presented as mean ± SD for all other variables. Prevalence of 5 common diseases by frailty status at baseline were computed. To evaluate bivariate relationship of physiological measures with frailty level, Pearson χ2 tests were used for categorical variables and analysis of variance F tests were used for continuous variables. For data that were highly skewed, significance testing was done on a log-transformed scale.
Because frailty is defined as a 3-level categorical outcome variable, multinomial logistic regression was performed to evaluate multivariable relationships between frailty level and physiological measures and the 3 frailty categories. Odds ratios (ORs) were estimated for the intermediate and frail groups with respect to the nonfrail group. Reported P values correspond to the null hypothesis that the ORs for both comparisons are simultaneously equal to 0, the alternative hypothesis being that at least 1 OR is statistically significantly (P<.01) different from 0. Individual 95% confidence intervals are also presented for each estimate. All regressions were adjusted for age, sex, and race. Continuous variables were coded into categories based on quartile or tertile values within the frail group. This strategy resulted in sufficient numbers in each category to provide relatively stable OR estimates and does not require the variable to have a linear relationship with the logistic function. Because of the large number of statistical tests performed, P values should not be strictly interpreted; rather, they are presented to illustrate the relative strength of association between frailty level and these physiological measures. Analyses were carried out using SPSS for Windows (SSPS Inc, Chicago, Ill) and Stata software (Stata Corp, College Station, Tex).
The CHS cohort was 57% female, 43% male, 95% white, and 5% African American; 71% of the members were recruited from Medicare eligibility lists, and 29% were age-eligible household members. Age, education, income, BMI, and disease status of CHS participants included in the analysis are given in Table 1. In the CHS, 6.3% of the population were identified as frail (7.3% of the women and 4.9% of the men), while 48.3% were not frail and 45.3% were intermediate. As expected, the prevalence of frailty rose with age, from 2.5% in the group aged 65 to 70 years to 32.0% in the group 90 years and older. The mean ± SD age was 77.2 ± 6.4 years for the frail group, 73.4 ± 5.9 years for the intermediate group, and 71.5 ± 4.6 years for the not frail group (P<.001 for trend). The mean ± SD BMI (calculated as weight in kilograms divided by the square of height in meters) was higher in the frail group (26.7 ± 6.9) compared with the intermediate (26.7 ± 4.7) and not frail (26.1 ± 4.1) groups (P<.001 for trend).
Because frailty is thought to be both frequently associated with disease states and potentially independent of disease, we first evaluated the relationship between frailty and baseline disease status of 5 conditions in older adults. Higher frailty levels were associated with increased prevalence of CVD, congestive heart failure, diabetes, and hypertension (Table 2). There was no significant difference for cancer, probably because individuals undergoing treatment for cancer were excluded from participation in the CHS.
To determine if any relationships existed between inflammation and the syndrome of frailty independent of disease status, we performed analyses in the full cohort and after excluding those with CVD and diabetes. We identified several significant positive relationships between frailty status and acute-phase reactants in both the entire cohort and in the subgroups of participants without CVD or diabetes diagnoses. Higher CRP levels were identified in the frail compared with the not frail groups with and without CVD and with and without diabetes (Table 3). Increased mean levels of factor VIII and fibrinogen were also associated with increasing frailty level (Table 3). Multinomial logistic regression analysis further demonstrated an association between an increased risk of being frail or intermediate and increased CRP level (Table 4). Higher levels of factor VIII and fibrinogen were associated with increased risks of frailty in the adjusted logistic models for the frail group relative to the not frail group (Table 4). Those with the highest levels of CRP, factor VIII, and fibrinogen had a particularly high risk. These associations all persisted when CVD and diabetes were excluded. We identified no relationship between frailty and factor VII level.
We identified several significant bivariate relationships between frailty status and elevated glucose and insulin levels in a fasting state as well as at 2 hours after a 75-g oral glucose tolerance test (Table 3). In the full cohort, mean values for these measures generally increased with increasing frailty status, with the lowest mean values associated with the not frail group and the highest values with the frail group, and, except for fasting glucose level, persisted when those with CVD and diabetes mellitus were excluded. Those relationships weakened in the logistic models, especially after adjustment for BMI (Table 4).
Mean albumin and low-density lipoprotein (LDL) cholesterol levels were lower in the frail group with and without diabetes or CVD. The relationships of mean total cholesterol, high-density lipoprotein (HDL), and triglyceride levels with frailty level were weak and nonlinear in the case of the latter two (Table 3). With the full cohort, the logistic models indicated that a decreasing risk of frail or intermediate status was associated with an increasing total, LDL and HDL cholesterol, and albumin levels, while the highest triglyceride level was associated with greater risk. When participants with diabetes or CVD were excluded, those with the highest levels of LDL and total cholesterol were less likely to be frail (Table 4).
Because of the evidence of increased levels of acute-phase reactants in the frail group and the known relationship between inflammation and hemostasis, we further studied the relationship between frailty status and previously obtained measures of coagulation in the previously described subset of CHS participants without CVD.16 Increasing frailty level was associated with greater mean levels in several markers of ongoing clotting, including D dimer and factor XIa α1-antitrypsin (Table 5). All of these relations persisted or were strengthened after participants with diabetes were excluded (Table 5). Because of the small number of cases in this subset, the logistic models were not fit.
The present study describes the cross-sectional association of the geriatric syndrome of frailty with several biochemical alterations related to increased inflammation, altered carbohydrate metabolism, and increased coagulant activities. Importantly, several of these associations were independent of two of the most common clinical comorbidities in older adults, CVD and diabetes. Given that these analyses are cross-sectional, we are unable to make causal links between these physiological alterations and frailty. However, our findings support the hypothesis that the geriatric syndrome of frailty has a specific and definable underlying physiological etiology, and that this physiology exists in both the presence and absence of specific disease states. The data also support the hypothesis, yet to be tested, that specific physiological processes may make frail older adults more vulnerable to disease processes, functional decline, and mortality.
We have demonstrated a significant association between several inflammatory markers (CRP, factor VIII, and fibrinogen) and the syndrome of frailty in both the presence and absence of 2 prevalent chronic diseases, diabetes and CVD. Serum levels of CRP increase in response to acute or chronic illness and inflammatory states.13,17,19 While its exact role remains controversial, CRP may act to amplify both inflammation and clotting through a number of pathways. It can activate the shedding of the receptor for the key proinflammatory cytokine, interleukin (IL) 6, from stimulated monocytes, amplifying the signal transduction potential of circulating IL-6. High IL-6 levels have been previously associated with CRP, which is directly up-regulated by IL-6 through an IL-6 responsive element in its promotor region.13 Our findings provide evidence that even in the absence of significant clinical morbidity, frailty is consistent with increased inflammation, which is potentially associated with up-regulated IL-6. This in turn suggests the potential for a disease-independent inflammatory mechanism for frailty. Alternatively, frailty and inflammation associated with frailty may be influenced by the subclinical versions of CVD and/or diabetes to a much greater degree than previously suspected.
We have also demonstrated a significant elevation in fasting insulin levels, as well as a significant elevation in insulin and glucose levels 2 hours after an oral glucose tolerance test in the frail cohort. This relationship was weakened when those with diabetes and CVD were excluded from the analyses and when BMI was included in the logistic regression model. Insulin resistance is frequently associated with increased fat mass and decreased lean body mass. We and others have previously hypothesized that decreased lean body mass is a critical contributing physiological component to the syndrome of frailty.8 In fact, a recent cross-sectional study of older adults suggests that increased fat mass and decreased lean body mass as measured by dual-energy x-ray absorptiometry scan may predict poor physical performance.20 The frail subset had a significantly higher BMI than the intermediate and not frail subset, despite the weight loss variable used to identify frailty in this cohort. Although we have no body composition data in this cohort, the increased BMI, the trend toward glucose intolerance, and the lower walking speed and strength measures in the frail subset suggest a lower percentage of lean body mass.
Alternatively, our findings of glucose intolerance may be associated with frailty as a part of a chronic inflammatory state. Hypothalamic-pituitary-adrenal axis activity is increased in inflammatory conditions, and increased hypothalamic-pituitary-adrenal axis activity is known to powerfully influence glucose tolerance, as are the inflammatory cytokines IL-6 and tumor necrosis factor α.13,21 While we provide no supportive evidence for a direct relationship between increased hypothalamic-pituitary-adrenal axis activity and glucose intolerance with our present data set, indirect evidence here supports further hypothesis development about their relationship. In the CHS cohort of older adults with or without CVD or diabetes, we demonstrated a relatively weak but still significant association between the frail clinical criteria and lower LDL cholesterol, which is inconsistent with traditional views of CVD risk factors. This would be most consistent with the lower lipid levels reflecting and associating with frailty through an inflammation mechanism.22 There were significant minimal differences in albumin, triglyceride, and HDL cholesterol levels observed in the unadjusted analysis, which were substantially attenuated in the adjusted analyses. These values are also in the expected directions if driven by inflammatory mechanisms (Table 4). Both albumin and cholesterol levels are frequently used as nutritional and metabolic markers in older adults; low levels of both have been associated with increased morbidity and mortality.23 While we cannot be certain that changes in lipid and albumin levels do not have direct mechanistic implications, the pattern of these findings support a role for inflammation in the development of frailty.23,24
A number of markers of an ongoing coagulation process are elevated in a frail subset of CHS participants without clinical CVD and without diabetes. Although the power to detect significant differences between groups is limited in this subset by the number of frail individuals, the data are suggestive of a sustained ongoing mild activation of clotting as observed with increasing age.25 D dimer is known to be a potent transcription factor in and of itself and has been shown to induce the synthesis and release of biologically active IL-1β, IL-6, and plasminogen activator inhibitor.26 Fragment 1.2 and fibrinopeptide A, like D dimer, are markers of ongoing clotting activity.25 While the measurements taken in this study do not allow for the demonstration of direct pathological effects of these coagulation markers, our findings support the concept that a physiological state of heightened inflammation with secondary changes in coagulation exists in the frail cohort. It may also represent progressive degradation of vascular integrity characterized by chronic inflammation.27 These findings also provide further rationale for the study of these mechanisms as part of the increased vulnerability to morbidity and mortality observed in frail older adults.
While our data provide important new insights into the physiological correlates of frailty and that these correlates exist independent of two of the most common chronic disease states of older adults, they do not prove causal influence. Several questions remain to be explored: What mechanism(s) trigger frailty and its physiological correlates in a subset of older adults? Which of the biochemical changes identified herein play important mechanistic roles? Would interventions targeted at these pathways decrease the risk of being frail?
Clearly, subclinical disease can and does trigger chronic inflammatory changes, and frail older adults are more likely to have subclinical disease.28 However, a substantial subset of frail older adults have no identified triggers. Our findings, along with recent studies detailing the relationship between elevated IL-6 level and the development of disability and early mortality in healthy older adults, provide supportive evidence for further etiologic studies.13 Although we have no direct measurements of IL-6 or other inflammatory cytokines in our study, data from these previous population studies, together with our data, lend support to the hypothesis that age-related inflammation system changes may underlie the development of both the clinical criteria of frailty and the physiological vulnerability observed in frail older adults. Many of our findings support this hypothesis, with increased acute-phase reactants, glucose intolerance, and increased blood clotting all influenced by increased inflammation and/or inflammatory cytokines.13,19,29
There are several important limitations to this study. First, the CHS cohort is on average healthier than the age-matched general population. While the number of frail adults identified in this cohort may be fewer than the number in the general population, this is not likely to have affected the observed relationships between frailty and biochemical measures. Second, the subset analysis used to study clotting markers contained only 29 frail individuals, limiting power. Third, there are no dual-energy x-ray absorptiometry measurements of fat or lean body mass and no direct measures of immune or neuroendocrine functions that may influence inflammation and clotting in this cohort, limiting us to speculation in these areas.
Several potential confounding variables that may impact our findings were beyond the scope of these initial analyses. For example, the use of statins, corticosteroids, or angiotensin-converting enzyme inhibitors may influence or prevent frailty or influence the physiological variable of interest. In addition, variables such as hemoglobin levels may influence the frailty criteria of fatigue and walking speed. These potential confounding variables highlight the need for future analyses in which medication use and hemoglobin variables are included in the models.
Despite several limitations, the large, well-characterized cohort of older adults has allowed us to identify several important relationships that provide rationale for further physiological and molecular etiologic studies in frail older adults. Future studies of other physiological pathways as well as inclusion of additional potential confounding variables into the analytical models may help further define important etiologic pathways. Such etiologic understanding may in turn foster model interventions targeted at this most vulnerable subset of older adults.
Accepted for publication May 8, 2002.
This research was supported by contracts NO1-HC-85079 through NO1-HC-85086 and Georgetown Echo RC-HL 35129 JHU MRI RC-HL 15103 from the National Institutes of Health National Heart, Lung and Blood Institute, Bethesda, Md.
Dr Walston is a Paul Beeson Physician Faculty Scholar. The opinions and assertions expressed herein are those of the authors and are not to be construed as reflecting the views of the Uniformed Services University of the Health Sciences or the US Department of Defense.
The authors wish to thank D. Baldwin for her assistance in manuscript preparation.
Wake Forest University School of Medicine (Forsyth County, North Carolina): Gregory L. Burke, Sharon Jackson, Alan Elster, Curt D. Furberg, Gerardo Heiss, Dalane Kitzman, Margie Lamb, David S. Lefkowitz, Mary F. Lyles, Cathy Nunn, Ward Riley, John Chen, Beverly Tucker; Wake Forest University—ECG Reading Center (Forsyth County): Farida Rautaharju, Pentti Rautaharju; University of California, Davis (Sacramento County, California): William Bonekat, Charles Bernick, Michael Buonocore, Mary Haan, Calvin Hirsch, Lawrence Laslett, Marshall Lee, John Robbins, William Seavey, Richard White; Johns Hopkins University (Washington County, Maryland): M. Jan Busby-Whitehead, Joyce Chabot, George W. Comstock, Adrian Dobs, Linda P. Fried, Joel G. Hill, Steven J. Kittner, Shiriki Kumanyika, David Levine, Joao A. Lima, Neil R. Powe, Thomas R. Price, Jeff Williamson, Moyses Szklo, Melvyn Tockman; Johns Hopkins University—MRI Reading Center (Washington County): Norman Beauchamp, R. Nick Bryan, Douglas Fellows, Melanie Hawkins, Patrice Holtz, Naiyer Iman, Michael Kraut, Cynthia Quinn, Grace Lee, Carolyn C. Meltzer, Larry Schertz, Earl P. Steinberg, Scott Wells, Linda Wilkins, Nancy C. Yue; University of Pittsburgh (Allegheny County, Pennsylvania): Diane G. Ives, Charles A. Jungreis, Laurie Knepper, Lewis H. Kuller, Elaine Meilahn, Peg Meyer, Roberta Moyer, Anne Newman, Richard Schulz, Vivienne E. Smith, Sidney K. Wolfson. University of California, Irvine—Echocardiography Reading Center (baseline): Hoda Anton-Culver, Julius M. Gardin, Margaret Knoll, Tom Kurosaki, Nathan Wong; Georgetown University Medical Center—Echocardiography Reading Center (Washington, DC) (follow-up): John Gottdiener, Eva Hausner, Stephen Kraus, Judy Gay, Sue Livengood, Mary Ann Yohe, Retha Webb; New England Medical Center, Boston—Ultrasound Reading Center (Boston, Mass): Daniel H. O'Leary, Joseph F. Polak, Laurie Funk; University of Vermont—Central Blood Analysis Laboratory (Colchester): Elaine Cornell, Mary Cushman, Russell P. Tracy; University of Arizona, Tucson—Pulmonary Reading Center: Paul Enright; University of Washington, Seattle—Coordinating Center: Alice Arnold, Annette L. Fitzpatrick, Richard A. Kronmal, Bruce M. Psaty, David S. Siscovick, Will Longstreth, Patricia W. Wahl, David Yanez, Paula Diehr, Susan Heckbert, Corrine Dulberg, Bonnie Lind, Thomas Lumley, Ellen O'Meara, Jennifer Nelson, Charles Spiekerman; National Heart, Lung, and Blood Institute Project Office (Bethesda, Md): Diane Bild, Teri A. Manolio, Peter J. Savage, Patricia Smith
Corresponding author and reprints: Jeremy Walston, MD, Johns Hopkins Geriatrics Center, 5505 Bayview Cir, Baltimore, MD 21224 (e-mail: email@example.com).
Create a personal account or sign in to: