Abstract
Background
Indices of multimorbidity and deficit accumulation (ie, frailty indices) have been proposed as markers of biological aging. If true, changes in these indices over time should predict downstream changes in cognition and physical function, and mortality.
Methods
We examined associations that 8-year changes in (i) a multimorbidity index comprised of 9 chronic diseases and (ii) a frailty index (FI) based on deficit accumulation in functional, behavioral, and clinical characteristics had with subsequent measures of cognitive and physical function over 10 years. We drew data from 3 842 participants in the Action for Health in Diabetes clinical trial. They were aged 45–76 years at baseline and at risk for accelerated biological aging due to overweight/obesity and type 2 diabetes mellitus.
Results
A total of 1 501 (39%) of the cohort had 8-year increases of 1 among the 9 diseases tracked in the multimorbidity index and 868 (23%) had increases of ≥2. Those with greatest increases in multimorbidity tended to be older individuals, males, and non-Hispanic Whites. Greater FI increases tended to occur among individuals who were older, non-Hispanic White, heavier, and who had more baseline morbidities. Changes in multimorbidity and FI were moderately correlated (r = 0.26; p < .001). Increases in both multimorbidity and FI were associated with subsequently poorer composite cognitive function and 400-m walk speed and increased risk for death (all p < .001).
Conclusions
Accelerated biological aging, as captured by multimorbidity and frailty indices, predicts subsequent reduced function and mortality. Whether intensive lifestyle interventions generally targeting multimorbidity and FI reduce risks for downstream outcomes remains to be seen.
Clinical Trials Registration Number: NCT00017953
Keywords: Aging, Deficit accumulation, Intensive lifestyle intervention
Aging is associated with increases in the number of chronic diseases and in the number of deficits in functional, behavioral, and clinical characteristics. These increases exhibit heterogeneity unaccounted for by chronologic age alone. Accordingly, indices of multimorbidity and deficit accumulation (ie, frailty indices) have been proposed as clinical markers that can distinguish biological age from chronological age and as important outcomes in clinical trials (1,2). One would expect that changes over time in these 2 indices should be correlated and share common risk factors, and that increases in these indices should be predictive of important downstream adverse health outcomes.
This paper uses 8-year changes in a multimorbidity index comprised of age-related chronic diseases and a frailty index (FI) based on deficit accumulation in a large cohort of individuals at risk for accelerated biological aging due to type 2 diabetes mellitus and overweight or obesity. These individuals were enrolled in a clinical trial testing a multidomain lifestyle intervention, which was shown to significantly slow 8-year increases in multimorbidity and FI compared with a control condition (3,4). We describe the distribution of changes in the 2 indices, their intercorrelation, and risk factor relationships. We then examine the degree that changes in these indices were associated with subsequent cognitive and physical function and mortality over the following 10 years and whether they accounted for any intervention effects on these outcomes. Our goal is to inform the use of these measures as outcomes in future clinical trials.
Method
The Look AHEAD design, methods, and Consolidated Standards of Reporting Trials diagram have been published previously (5,6). Look AHEAD was a multisite, single-blind randomized controlled clinical trial that recruited 5 145 individuals (during 2001–2004) from 16 U.S. centers. Protocols and consent forms were approved by local Institutional Review Boards.
Participants
Look AHEAD participants had type 2 diabetes and met the following criteria: 45–76 years of age, body mass index (BMI) >25 kg/m2 (>27 kg/m2 if on insulin), glycated hemoglobin (HbA1c) <97 mmol/mol (11%), systolic/diastolic blood pressure <160/<100 mmHg, triglycerides <600 mg/dL, and successful passing of a maximum graded exercise test. The current study is restricted to participants for whom 8-year changes in multimorbidity and FI were observed and for whom at least one 400-m walk and/or cognitive assessment was completed from 8 years until the end of follow-up (N = 3 842).
Interventions
Participants had been randomly assigned to intensive lifestyle intervention (ILI) or diabetes support and education (DSE). The multidomain ILI targeted reducing caloric intake and increasing physical activity to induce weight loss to average >7% at year 1 and to maintain this over time (7,8). Caloric consumption goals of 1 200–1 800 kilocalories/day were based on initial weight. Physical activity of >175 min/wk through activities similar in intensity to brisk walking was targeted, as was improved diet (<30% calories from fat, <10% calories from saturated fat, >15% calories from protein). Cardiometabolic risk factors (lipids, HbA1c, blood pressure) were monitored and participants were provided results, which were, in turn and with consent, shared with their clinicians. During the first 6 months, ILI participants attended 3 group meetings and 1 individual session per month. For the remainder of the first year, they were provided 2 group and 1 individual meeting per month. The intensity of the intervention gradually decreased thereafter (7).
DSE participants were invited to attend group sessions focused on diet, physical activity, and social support (8). Four meetings were offered during year 1, 3 per year during years 2–4, and 1 annually thereafter. Participants did not receive specific diet, activity, or weight goals, or information on behavioral strategies; however, the protocol for sharing risk factor information with participants and their physicians was the same as for ILI.
Interventions were terminated September 2012, when all participants’ planned follow-up was at least 8 years (minimum 8.6 years) (6).
Multimorbidity
Our multimorbidity index is a subset of 20 chronic diseases developed by the United States Department of Health and Human Services (9). This list includes 9 conditions ascertained at baseline and follow-up by Look AHEAD: cancer, cardiac arrhythmia, chronic kidney disease, congestive heart failure, coronary artery disease, depression, dyslipidemia, hypertension, and stroke (4). We excluded diabetes, which was common to all participants. The remaining chronic diseases from this list that we were not able to include (ie, were not ascertained at baseline and during the first 8 years of follow-up on all Look AHEAD participants) were: arthritis, asthma, autism, chronic obstructive pulmonary disease, dementia, hepatitis, human immunodeficiency virus, osteoporosis, schizophrenia, and substance abuse. In limiting our focus to chronic diseases, we excluded many important geriatric syndromes (eg, incontinence, falls, etc.); however, these contribute to the FI. Details are provided in Supplementary Exhibit S1.
Deficit Accumulation FI
We constructed a 38-item FI. It was modeled after the FI developed by the Systolic Blood Pressure Intervention Trial (SPRINT) (10), and which we augmented to include 9 additional deficits related to diabetes and obesity, which gave comparable results to the SPRINT FI and met the recommendation that at least 30 components be included in FI (Supplementary Exhibit S2) (3,11). Included were measures of behavior, medical history, clinical biomarkers, function, and abilities. Individual component scores range between 0 and 1, with higher scores being worse. The total FI is the ratio of the sum of the individual component scores divided by the number of components, ranging from 0 to 1.
Mortality
Mortality was adjudicated from death certificates, recent hospitalization records (discharge summaries only), outcomes interviews, and a National Death Index search. All deaths reported subsequent to participant’s year 8 follow-up visit through June 30, 2020 were included in analyses linking changes in FI and multimorbidity to subsequent risks for mortality. Those deaths occurring prior to this contributed to the multimorbidity index but are not included in the analyses of risks following the year 8 visit.
Other Measures
Demographic characteristics were collected by self-report. Self-reported race/ethnicity was included as a covariate in analyses to account for potential biases in cognitive assessment methods. BMI was based on standardized clinical measurements.
Cognitive assessment was performed among those enrolled in the postintervention observation study, 10–13 years after their Look AHEAD enrollment. A subset had 1 or 2 earlier assessments in the Look AHEAD Movement and Memory Study (4 clinics: years 8–11) (12), the Look AHEAD Brain MRI study (3 clinics: years 10–12) (13), and the Look AHEAD MIND Study (15 clinics; years 16–18) (14).
Assessments were performed by staff who were centrally trained, certified, and masked. A composite cognitive function measure was created by averaging standardized scores from tests of verbal learning and memory, speed of processing and working memory, executive function, and global cognitive functioning (12).
Usual 400-m walk speed (15) was assessed first at 8–9 years in a subset of participants, and in the full cohort at approximately 11 years (range, 9.5–13.2 years) after randomization, and again at years 16 and/or 18 during follow-up (16,17). Of those included in our analysis database, 6% had not undertaken at least one 400-m walk. The course was 10 or 20 m long, depending on clinic space, and marked by cones at each end and walks were terminated if not completed within 15 minutes. Participants were instructed to walk at their usual pace and time (to the nearest second) to complete the 400-m walk was recorded. Participants who were wheelchair-bound or were dependent on a walker or quad cane, or who had a cardiovascular disease event in the past 3 months, or whose blood pressure was >170/100 mmHg were excluded from testing. For participants unable to complete the walk, we used gait speed over the distance they covered. We also examined associations with the percent of participants who started the walk and were able to complete it within 15 minutes (as a dichotomous outcome).
Statistical Analysis
We grouped individuals according to 8-year changes in the multimorbidity index (0, 1, or ≥2) and changes in the FI (tertiles) and used chi-squared tests and analyses of variance to compare risk factor profiles among these groups. Correlation coefficients were computed to assess associations between baseline levels of the multimorbidity and frailty indices and between 8-year changes in these indices. Mixed effects models were fitted to compare, among these groups, the level and trajectories of composite cognitive function and 400-m walk speeds collected during years 8–18. We also used generalized estimating equations to compare the rates of individuals who completed the walk in 15 minutes. Because stroke and heart failure may be thought to drive associations that increases in multimorbidity have with cognitive function and walk speeds, we repeated analyses removing them from the multimorbidity index. We also repeated analyses using the FI based on the SPRINT trial without the additional components we added (10), but do not report these results because they did not materially differ from those using the augmented FI. Proportional hazards regression was used to compare mortality during this span of time. Stabilized inverse probability weighting was used to assess the sensitivity of results to missing data (18). No imputation was performed.
Results
At baseline, the mean (SD) multimorbidity index for the cohort was 2.17 (0.92) and the mean FI was 0.202 (0.069). The correlation between indices was r = 0.36 (p < .001)
As seen in Table 1, over the 8 years of follow-up, the number of the 9 tracked morbidities increased by 1 for 1 501 (39%) participants and by ≥2 for 868 (23%) participants. The overall mean (SD) increase was 0.89 (0.96). Older individuals, males, and non-Hispanic Whites tended to be overrepresented among those with the greatest increases. Assignment to the ILI was associated with smaller increases in morbidities (p = .007). While those with baseline coronary artery disease tended to have greater increases in multimorbidity, this was not the case for those with chronic kidney disease, depression, dyslipidemia, or hypertension, which resulted overall in an inverse relationship between baseline multimorbidity and 8-year expansions. There was little association between baseline FI and 8-year multimorbidity expansion (p = .41).
Table 1.
Characteristics at Look AHEAD Enrollment by Changes in Multimorbidity From Baseline to Year 8: N (%) or Mean (SE)
| Number of New Morbidities by Year 8 | p Value* | |||
|---|---|---|---|---|
| None | 1 | ≥2 | ||
| N = 1 473 | N = 1 501 | N = 868 | ||
| Age, years | ||||
| 45–54 | 414 (28.1) | 367 (24.5) | 196 (22.6) | |
| 55–64 | 838 (56.9) | 862 (57.4) | 485 (55.9) | <.001 |
| 65–76 | 221 (15.0) | 272 (18.1) | 187 (21.5) | |
| Female | 924 (62.7) | 922 (61.4) | 493 (56.8) | .02 |
| Race/ethnicity | ||||
| African American | 274 (18.6) | 241 (16.1) | 120 (13.8) | |
| American Indian | 93 (6.3) | 75 (5.0) | 41 (4.7) | .003 |
| Hispanic | 165 (11.2) | 226 (15.1) | 107 (12.3) | |
| Non-Hispanic White | 895 (60.8) | 911 (60.7) | 574 (66.1) | |
| Other, multiple | 46 (3.1) | 48 (3.2) | 26 (3.0) | |
| Education | ||||
| High School or less | 822 (55.8) | 858 (57.2) | 503 (58.0) | |
| College graduate | 309 (21.0) | 330 (22.0) | 195 (22.5) | .55 |
| Postcollege | 302 (20.5) | 277 (18.5) | 150 (17.3) | |
| Other | 40 (2.7) | 36 (2.4) | 20 (2.3) | |
| BMI, kg/m2 | ||||
| 25–29 | 226 (15.3) | 234 (15.6) | 141 (16.2) | |
| 30–39 | 906 (61.5) | 956 (63.7) | 533 (61.4) | .55 |
| ≥40 | 341 (23.1) | 311 (20.7) | 194 (22.4) | |
| Baseline morbidity | ||||
| Cancer | 118 (8.0) | 112 (7.5) | 70 (8.1) | .81 |
| Cardiac arrhythmia | 0 | 0 | 0 | |
| Chronic kidney disease† | 75 (5.1) | 59 (3.9) | 22 (2.5) | .01 |
| Congestive heart failure | 8 (0.5) | 9 (0.6) | 6 (0.7) | .90 |
| Coronary artery disease‡ | 123 (8.4) | 124 (8.3) | 106 (12.2) | .003 |
| Depression§ | 584 (39.7) | 386 (25.7) | 121 (13.9) | <.001 |
| Dyslipidemia‖ | 1 420 (96.4) | 1 301 (86.7) | 634 (73.0) | <.001 |
| Hypertension¶ | 1 312 (89.1) | 1 111 (74.0) | 551 (63.5) | .002 |
| Stroke | 33 (2.2) | 29 (1.9) | 18 (2.1) | .84 |
| Baseline multimorbidity index | ||||
| Mean (SE) | 2.49 (0.02) | 2.09 (0.02) | 1.76 (0.03) | <.001 |
| Level | ||||
| 0 | 4 (0.3) | 24 (1.6) | 60 (6.9) | |
| 1 | 106 (7.2) | 337 (22.4) | 302 (34.8) | |
| 2 | 677 (46.0) | 727 (48.4) | 329 (38.0) | |
| 3 | 686 (46.6) | 413 (27.5) | 177 (20.4) | |
| Baseline FI | 0.201 (0.002) | 0.202 (0.002) | 0.205 (0.002) | .41 |
| ILI assignment | 787 (53.4) | 745 (49.6) | 407 (46.9) | .007 |
Notes: BMI = body mass index; FI = frailty index; ILI = intensive lifestyle intervention; SE = standard error.
†Estimated glomerular filtration rate < 60.
‡Myocardial infarction, coronary artery bypass, angina.
§Beck Depression Index-2 >10 or current antidepressant medication use.
‖Low-density lipoprotein cholesterol > 100 or current lipid-lowering drug use.
¶Blood pressures >140/90 mmHg or current antihypertensive medication use.
*t Test or chi-squared test.
Over 8 years, the overall mean (SD) increase in FI was 0.029 (0.074). Tertile groups were based on the following cut points: 33% of FI changes were less than −0.0047 and 33% were greater than 0.0054. Table 2 describes baseline characteristics of individuals grouped by tertile of FI change. Greater increases tended to occur among individuals who were older, non-Hispanic White, and heavier; who had greater baseline multimorbidity; and who had not been assigned to ILI. Greater increases also were seen for those with baseline history of cancer, chronic kidney disease, coronary artery disease, or hypertension. Interestingly, however, there was some evidence of regression to the mean, with individuals in the lowest tertile of FI changes having the greatest baseline FI scores.
Table 2.
Characteristics at Look AHEAD Enrollment by Changes in FI Scores at Year 8: Mean (SE) or N (%)
| Tertile Change in FI From Baseline to Year 8 | p Value* | |||
|---|---|---|---|---|
| <−0.0047 | −0.0047 to 0.0540 | ≥0.0540 | ||
| N = 1 281 | N = 1 280 | N = 1 281 | ||
| Age, years | ||||
| 45–54 | 371 (29.0) | 320 (25.0) | 286 (22.3) | |
| 55–64 | 745 (58.2) | 722 (56.4) | 718 (56.0) | <.001 |
| 65–76 | 165 (12.9) | 238 (18.6) | 277 (21.6) | |
| Female | 797 (62.2) | 773 (60.4) | 769 (60.0) | .48 |
| Race/ethnicity | ||||
| African American | 223 (17.4) | 218 (17.0) | 194 (15.1) | |
| American Indian | 87 (6.8) | 79 (6.2) | 43 (3.4) | |
| Hispanic | 194 (15.1) | 135 (10.6) | 169 (13.2) | <.001 |
| Non-Hispanic White | 751 (58.6) | 801 (62.6) | 828 (64.6) | |
| Other, multiple | 26 (2.0) | 47 (3.7) | 47 (3.7) | |
| Education | ||||
| Not college graduate | 730 (57.0) | 711 (55.6) | 742 (57.9) | |
| College graduate | 267 (20.8) | 279 (21.8) | 288 (22.5) | .57 |
| Postcollege | 252 (19.7) | 255 (19.9) | 222 (17.3) | |
| Other | 32 (2.5) | 35 (2.7) | 29 (2.3) | |
| BMI, kg/m2 | ||||
| 25–29 | 216 (16.8) | 201 (15.7) | 184 (14.4) | |
| 30–39 | 812 (63.4) | 815 (63.7) | 768 (60.0) | .003 |
| >40 | 253 (19.8) | 264 (20.6) | 329 (25.7) | |
| Baseline morbidity | ||||
| Cancer | 82 (6.4) | 102 (8.0) | 116 (9.1) | .04 |
| Cardiac arrhythmia | 0 | 0 | 0 | — |
| Chronic kidney disease† | 48 (3.8) | 37 (2.9) | 71 (5.5) | .002 |
| Congestive heart failure | 4 (0.3) | 13 (1.0) | 6 (0.5) | .05 |
| Coronary artery disease‡ | 84 (6.6) | 121 (9.4) | 148 (11.6) | <.001 |
| Depression§ | 401 (31.3) | 315 (24.6) | 375 (29.3) | <.001 |
| Dyslipidemia‖ | 1 122 (87.6) | 1 102 (86.1) | 1 131 (88.3) | .23 |
| Hypertension¶ | 962 (75.1) | 969 (75.7) | 1 043 (81.4) | <.001 |
| Stroke | 17 (1.3) | 31 (2.4) | 32 (2.5) | .07 |
| Total multimorbidity index | ||||
| Mean (SE) | 2.12 (0.03) | 2.10 (0.03) | 2.28 (0.03) | <.001 |
| Level | ||||
| 0 | 30 (2.3) | 40 (3.1) | 18 (1.4) | |
| 1 | 269 (21.0) | 266 (20.8) | 210 (16.4) | |
| 2 | 570 (44.5) | 595 (46.5) | 548 (44.3) | |
| 3 | 412 (32.2) | 379 (29.6) | 485 (37.9) | |
| Baseline FI | 0.219 (0.002) | 0.192 (0.002) | 0.196 (0.002) | .002 |
| ILI Assignment | 729 (56.9) | 619 (48.4) | 591 (46.1) | <.001 |
Notes: BMI = body mass index; FI = frailty index; ILI = intensive lifestyle intervention; SE = standard error.
†Estimated glomerular filtration rate < 60.
‡Myocardial infarction, coronary artery bypass, angina.
§Beck Depression Index-2 >10 or current antidepressant medication use.
‖Low-density lipoprotein cholesterol > 100 or current lipid-lowering drug use.
¶Blood pressures >140/90 mmHg or current antihypertensive medication use.
*t Test or chi-squared test.
Overall, the correlation between 8-year changes in multimorbidity and FI was r = 0.26 (p < .001). Correlations were similar across subgroups based on gender, age, BMI, and intervention assignment, ranging from r = 0.23 to r = 0.29.
Across years 8–18, the number of cognitive assessments varied among participants: 28.1% had only 1, 50.6% had 2, 15.5% had 3, and 5.8% had 4. The number of 400-m walk assessments also varied among participants: 20.8% had only 1, 22.3% had 2; 44.4% had 3; and 12.4% had 4. Of the 3 842 participants included in our analyses, 535 (13.9%) died following their year 8 visit through year 18. Table 3 describes the associations that 8-year changes in multimorbidity and FI had with subsequent composite cognitive function, 400-m gait speed, and mortality risk. For each of these outcomes, there was a significant graded relationship: increases in multimorbidity and FI were associated with worse functional outcomes and increased mortality risk (all p < .001). Relationships with the functional outcomes were slightly stronger for FI changes. Figure 1 portrays the trajectories traced by mean composite cognition and gait speed by FI tertile across 3 spans of follow-up: years 8–9, years 10–13, and years 14–18. For both outcomes, individuals with the greatest 8-year increases in FI had worse initial profiles and differences were maintained throughout longer term follow-up.
Table 3.
Associations That Changes in Multimorbidity and Frailty Have With Subsequent Cognitive and Physical Function Scores and Risk of Death With Adjustment for Gender, Current Age, Education, Race/Ethnicity, and Randomization Assignment, Number of Prior Cognitive Assessments, and Baseline BMI, Multimorbidity, and FI
| Mean (SE) Composite Cognitive Function Scores | Mean (SE) 400-m Walk Speed (m/s) | Hazard Ratio [95% CI] Mortality | |
|---|---|---|---|
| Change in multimorbidity | |||
| None | −0.16 (0.02) | 0.965 (0.005) | 1.00 |
| 1 | −0.19 (0.02) | 0.938 (0.005) | 1.46 [1.18,1.82] |
| ≥2 | −0.29 (0.02) | 0.894 (0.007) | 2.04 [1.61, 2.58] |
| p Value* | <.001 | <.001 | <.001 |
| Tertile change in FI | |||
| Least increase | −0.129 (0.018) | 0.992 (0.005 | 1.00 |
| Midlevel increase | −0.172 (0.018) | 0.945 (0.005) | 1.82 [1.43, 2.32] |
| Greatest increase | −0.300 (0.019) | 0.877 (0.005) | 2.32 [1.84, 2.94] |
| p Value* | <.001 | <.001 | <.001 |
Notes: BMI = body mass index; CI = confidence interval; FI = frailty index; SE = standard error.
*Mixed effects models or proportional hazards regression.
Figure 1.
Mean composite cognition (A) and gait speed (B) by frailty index (FI) tertile across 3 spans of follow-up: years 8–9, years 10–13, and years 14–18, with adjustment for age, gender, race/ethnicity, education, intervention assignment, baseline frailty, and baseline multimorbidity.
We also examined associations with the percent of individuals who undertook the 400-m walk that were able to complete it within the required 15 minutes. Among all 400-m walks over time, these ranged from 89.4%, 86.6%, and 82.2% across the 3 groups based on multimorbidity increases and from 91.5%, 89.3%, and 78.2% across the tertiles of FI changes and (both p < .001, with similar covariate adjustment as for Table 3).
Figure 2 portrays the incidence of deaths over time, stratified by change in multimorbidity and FI tertiles. For both indices, there was a graded relationship with more adverse changes being related to greater mortality risk (p < .001).
Figure 2.
Kaplan–Meier plots of times until incidence of death by change in multimorbidity and frailty indices tertiles with adjustment for age, gender, education, race/ethnicity, intervention assignment, baseline body mass index (BMI), and baseline multimorbidity.
When both changes in multimorbidity and FI were included in the above models, each remained independently related to the functional and mortality outcomes (all p < .01).
During the 8 years of follow-up, 125 (3.3%) participants reported a stroke, which might have strongly affected cognitive and physical functioning. We examined the impact of removing strokes from the change scores for multimorbidity. This had little overall effect on findings. For example, the associations between increases in multimorbidity and subsequent composite cognitive function remained statistically significant (p < .001), with fitted means (SE) of −0.16 (0.02) for no increase, −0.20 (0.02) for 1 morbidity increase, and −0.28 (0.02) for increases of >2. Similarly, 400-m walk speeds differed among groups with strokes removed (p < .001): 0.964 (0.005) m/s for no increase, 0.936 (0.005) m/s for 1 morbidity increase, and 0.897 (0.007) m/s for increases of 2 or more. Removing the 41 cases of incident congestive heart failure also had no material effect on results.
As described in Supplementary Exhibit S3, we used stabilized inverse probability weighting to assess the impact that differential attrition across 8 years might have on our findings. Overall, we found no evidence that results were biased, based on these analyses.
With covariate adjustment for baseline FI, multimorbidity, obesity, age, gender, education, and race/ethnicity, mean (SE) 400-m walk speed over time varied between intervention groups (p < .001): 0.927 (0.004) m/s for DSE and 0.952 (0.004) m/s for ILI. Inclusion of changes in multimorbidity did not attenuate these differences, which remained highly significant (p < .001): 0.928 (0.004) m/s for DSE and 0.950 (0.004) m/s for ILI. Inclusion of tertile change in FI attenuated them only slightly (p = .004): 0.931 (0.004) m/s for DSE and 0.948 (0.004) m/s for ILI. There were no overall differences between intervention groups following the year 8 visit in mean composite cognitive function (p = .73) or mortality (p = .90).
Discussion
Multimorbidity and frailty, as assessed by deficit accumulation, are overlapping constructs that are associated cross-sectionally in Look AHEAD and in other cohorts (19,20). We have shown that they also track over time, and although 8-year longitudinal changes in multimorbidity and FI are not strongly correlated, that is, in the range between r = 0.20 to r = 0.30, associations are fairly consistent across age, gender, body weight, and the two Look AHEAD intervention conditions.
Greater increases in both indices were seen among older and non-Hispanic White individuals; however, other risk factor relationships diverged. Being male was associated with greater increases in multimorbidity, but not in FI. Obesity was associated with greater increases in FI, but not multimorbidity. Somewhat surprisingly, for both indices relationships between baseline levels and subsequent changes were inverse, which may reflect a regression to the mean, a ceiling effect, or differences in clinical care as new disease or deficits arise. For multimorbidity, this inverse relationship appeared to be driven by depression, dyslipidemia, and hypertension, conditions that were improved by the Look AHEAD ILI intervention (21,22). FI increases were greater among individuals with greater multimorbidity, but increases in multimorbidity were unrelated to baseline FI.
It is widely reported that greater levels of multimorbidity and FI are associated with poorer trajectories of physical and cognitive function and increased risk of death (19,23–28). To our knowledge, we are the first to report that increases in multimorbidity and FI over an extended period of time are also associated with poorer subsequent trajectories of cognitive function and gait speed, even with covariate adjustment for baseline multimorbidity and FI. Moreover, for both indices, these adverse associations remained highly statistically significant, even with covariate adjustment for changes in the other index. Our supporting analyses indicate that these associations are not overly influenced by differential attrition (including mortality) and that the associations with poorer cognitive and physical function are not driven by stroke and heart failure incidence. There has recently been another report that increases in FI are associated with greater risks for deaths, with each 0.01/year increase in FI being associated with 24% to 71% increased hazards for mortality across 4 large cohorts (29).
Both multimorbidity and frailty are reasonable, patient-centered targets for clinical trials. It is clear that improvements in these outcomes would be associated with better health-related quality of life and likely fewer requirements for medical care. Our findings from the Look AHEAD trial, however, suggest that improvement in these indices may not lead to long-term benefits in other important outcomes. Across 8 years, the Look AHEAD ILI appeared to slow increases in multimorbidity and FI corresponding roughly to what was observed over 1 year in the control group (3,4). ILI was also associated with modest, but sustained benefits for gait speed (16,17); however, as we have shown, these benefits in mobility were largely independent of the benefits in multimorbidity and FI. And, the ILI benefits on these indices did not translate to overall improvements in cognitive function (12,30), nor risks for cognitive impairment (31) or mortality (6). It is possible that the magnitude of benefits on the indices was too small or the time frame was too short for them to be expressed in other age-related outcomes. It is also possible that the characteristics of the cohort, that is, type 2 diabetes or obesity, or the nature of the intervention presented a milieu that disassociated outcomes. For example, weight loss may have a variable impact on outcomes depending on chronological age (32,33), raising the possibility that as one’s age-related health status declines, weight loss may engender responses that countermand and even reverse its overall benefit. An example of this arises from Look AHEAD. When participants were grouped by baseline FI tertile, there were significant differences in the relative impact of ILI on the incidence of major cardiovascular events across tertiles (p = .03), with potential benefits for participants with low FI (hazard ratio = 0.73; 95% confidence interval [0.55, 0.98]) which were dissipated and possibly reversed for individuals with the highest baseline FI tertile (hazard ratio = 1.15 [0.94, 1.42]) (34). In a similar fashion, the ILI was found to have differential effects among individuals grouped at baseline into diabetes subtypes based on biomarker profiles (35).
Overall, our findings are consistent with an enriched view for the design of clinical trials based on the geroscience hypothesis. While it is recognized that there is an underlying process of biological aging leading to loss of function and increased disease, changes in multimorbidity and FI appear to be only partial expressions of this process, and thus may not serve as general surrogates for biological aging.
Limitations
While our study is strengthened by a large, diverse study cohort with high retention and standardized data collection, we note several limitations. First, as volunteers to a clinical trial of lifestyle intervention, the Look AHEAD cohort may not represent other populations and our findings may not generalize to younger or older individuals and those without type 2 diabetes or overweight/obesity. The Look AHEAD eligibility requirement (eg, completing a graded exercise test, behavioral run-in) may have compressed the range of the FI and cognitive function. Secondly, we relied on self-report for many components of multimorbidity and frailty indices, which may have variable reliability. The multimorbidity index we use is less inclusive than some used in other studies (24,36,37).
The number of assessments of cognitive and gait and retention varied among participants nonrandomly, and were censored by deaths, which may attenuate relationships. We have previously reported that lost follow-up was linked to poorer health profiles (4,17); thus, it is likely that this led to underestimation of the strength of associations we report. Mortality is a competing risk for declines in cognitive and physical function. The specific FI we use has not been validated outside of the Look AHEAD cohort; however, it meets standard recommendations for the construction of FIs (11). The age range of the Look AHEAD cohort (45–76 years at entry) limits generalizability to older age groups.
Summary
Increases in multimorbidity and frailty augur poorer trajectories of cognitive and physical functioning and increased risk for death. They represent overlapping, but not interchangeable, markers of aging and appear to be important targets for intervention.
Supplementary Material
Contributor Information
Mark A Espeland, Sticht Center for Healthy Aging and Alzheimer’s Prevention, Wake Forest School of Medicine, Winston-Salem, North Carolina, USA; Department of Biostatistics and Data Science, Wake Forest School of Medicine, Winston-Salem, North Carolina, USA.
Jamie Nicole Justice, Sticht Center for Healthy Aging and Alzheimer’s Prevention, Wake Forest School of Medicine, Winston-Salem, North Carolina, USA.
Judy Bahnson, Department of Biostatistics and Data Science, Wake Forest School of Medicine, Winston-Salem, North Carolina, USA.
Joni K Evans, Department of Biostatistics and Data Science, Wake Forest School of Medicine, Winston-Salem, North Carolina, USA.
Medha Munshi, Joslin Geriatric Diabetes Program, Joslin Diabetes Center, Boston, Massachusetts, USA.
Kathleen M Hayden, Department of Social Sciences and Health Policy, Wake Forest School of Medicine, Winston-Salem, North Carolina, USA.
Felicia R Simpson, Department of Mathematics, Winston-Salem State University, Winston-Salem, North Carolina, USA.
Karen C Johnson, Department of Preventive Medicine, University of Tennessee Health Science Center, Memphis, Tennessee, USA.
Craig Johnston, Department of Health and Human Performance, University of Houston, Houston, Texas, USA.
Stephen R Kritchevsky, Sticht Center for Healthy Aging and Alzheimer’s Prevention, Wake Forest School of Medicine, Winston-Salem, North Carolina, USA.
Funding
The work was supported, as follows. Look AHEAD trial was supported by the National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Department of Health and Human Services (grant number U01DK057136). A diversity supplement to this parent award provided support for F.R.S. (grant number 3U01DK057136-19S1). The Look AHEAD MIND ancillary study was funded by the National Institute on Aging (grant number AG058571). The Look AHEAD Brain MRI ancillary study was supported by the National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Department of Health and Human Services (grant numbers DK092237-01 and DK092237-02S2). The Look AHEAD Movement and Memory ancillary study was supported by the National Institute on Aging, National Institutes of Health, Department of Health and Human Services (grant number AG03308701). The Look AHEAD Aging study is funded by the National Institute on Aging, National Institutes of Health, Department of Health and Human Services, AG073697. Additional support was provided through a grant from the Glenn Foundation for Medical Research and the Wake Forest Older Americans Independence Center (grant number P30-021332). Other funding sources and the Look AHEAD study group are listed in Supplementary Exhibit S4.
Conflict of Interest
None declared.
Author Contributions
The sponsors played no role in the development of this manuscript. All coauthors have made substantial contributions to the manuscript. M.A.E. obtained funding, supervised analyses, and wrote the initial draft of the manuscript. J.B. and K.M.H. obtained funding and reviewed drafts and participated in the writing of the manuscript. J.N.J. and S.R.K. provided critical scientific motivation and coauthored sections of the manuscript. J.K.E. conducted statistical analyses and reviewed manuscript drafts. M.M., F.R.S., K.C.J., and C.J. all reviewed and edited several drafts of the manuscript and contributed significantly to its discussion.
References
- 1. Espeland MA, Crimmins EM, Grossardt BR, et al. ; Multimorbidity Clinical Trials Consortium . Clinical trials targeting aging and age-related multimorbidity. J Gerontol A Biol Sci Med Sci. 2017;72(3):355–361. doi: 10.1093/gerona/glw220 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Theou O, van der Valk AM, Godin J, et al. Exploring clinically meaningful changes for the frailty index in a longitudinal cohort of hospitalized older patients. J Gerontol A Biol Sci Med Sci. 2020;75(10):1928–1934. doi: 10.1093/gerona/glaa084 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Simpson FR, Pajewski NM, Nicklas B, et al. ; Indices for Accelerated Aging in Obesity and Diabetes Ancillary Study of the Action for Health in Diabetes (Look AHEAD) Trial . Impact of multidomain lifestyle intervention on frailty through the lens of deficit accumulation in adults with type 2 diabetes mellitus. J Gerontol A Biol Sci Med Sci. 2020;75(10):1921–1927. doi: 10.1093/gerona/glz197 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Espeland MA, Gaussoin SA, Bahnson J, et al. Impact of an 8-year intensive lifestyle intervention on an index of multimorbidity. J Am Geriatr Soc. 2020;68(10):2249–2256. doi: 10.1111/jgs.16672 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. The Look AHEAD Research Group. Design and methods for a clinical trial of weight loss for the prevention of cardiovascular disease in type 2 diabetes. Control Clin Trials. 2003;24:610–628. doi: 10.1016/s0197-2456(03)00064-3 [DOI] [PubMed] [Google Scholar]
- 6. The Look AHEAD Research Group. Cardiovascular effects of intensive lifestyle intervention in type 2 diabetes. New Engl J Med. 2013;369:145–154. doi: 10.1056/NEJMoa1212914 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. The Look AHEAD Research Group. The Look AHEAD Study: a description of the lifestyle intervention and the evidence supporting it. Obesity. 2006;14:737–752. doi: 10.1038/oby.2006.84 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. The Look AHEAD Research Group. The development and description of the diabetes support and education (comparison group) intervention for the Action for Health in Diabetes (Look AHEAD) trial. Clin Trials. 2011;8:320–329. doi: 10.1177/1740774511405858 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Goodman RA, Posner SF, Huang ES, Parekh AK, Koh HK. Defining and measuring chronic conditions: imperatives for research, policy, program, and practice. Prev Chronic Dis. 2013;10:E66. doi: 10.5888/pcd10.120239 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Pajewski NM, Williamson JD, Applegate WB, et al. ; SPRINT Study Research Group . Characterizing frailty status in the systolic blood pressure intervention trial. J Gerontol A Biol Sci Med Sci. 2016;71(5):649–655. doi: 10.1093/gerona/glv228 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Searle SD, Mitnitski A, Gahbauer EA, Gill TM, Rockwood K. A standard procedure for creating a frailty index. BMC Geriatr. 2008;8:24. doi: 10.1186/1471-2318-8-24 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Espeland MA, Rapp SR, Bray GA, et al. ; Action for Health in Diabetes (Look AHEAD) Movement and Memory Subgroup; Look AHEAD Research Group . Long-term impact of behavioral weight loss intervention on cognitive function. J Gerontol A Biol Sci Med Sci. 2014;69(9):1101–1108. doi: 10.1093/gerona/glu031 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Espeland MA, Erickson K, Neiberg RH, et al. Brain and white matter hyperintensity volumes after ten years of random assignment to lifestyle intervention. Diabetes Care. 2016;39(5):764–771. doi: 10.2337/dc15-2230 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Espeland MA, Yassine H, Hayden KD, et al. ; Action for Health in Diabetes (Look AHEAD) Research Group . Sex-related differences in cognitive trajectories in older individuals with type 2 diabetes and overweight or obesity. Alzheimers Dement. 2021;7(1):e12160. doi: 10.1002/trc2.12160 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Simonsick EM, Montgomery PS, Newman AB, Bauer DC, Harris T. Measuring fitness in healthy older adults: the Health ABC Long Distance Corridor Walk. J Am Geriatr Soc. 2001;49(11):1544–1548. doi: 10.1046/j.1532-5415.2001.4911247.x [DOI] [PubMed] [Google Scholar]
- 16. Houston DK, Leng X, Bray GA, et al. ; Action for Health in Diabetes (Look AHEAD) Movement and Memory Ancillary Study Research Group . A long-term intensive lifestyle intervention and physical function: the Look AHEAD Movement and Memory Study. Obesity (Silver Spring). 2015;23(1):77–84. doi: 10.1002/oby.20944 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Houston DK, Neiberg RH, Miller ME, et al. Physical function following a long-term lifestyle intervention among middle aged and older adults with type 2 diabetes: the Look AHEAD study. J Gerontol A Biol Sci Med Sci. 2018;73(11):1552–1559. doi: 10.1093/gerona/glx204 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Austin PC, Stuart EA. Moving towards best practice when using inverse probability of treatment weighting (IPTW) using the propensity score to estimate causal treatment effects in observational studies. Stat Med. 2015;34(28):3661–3679. doi: 10.1002/sim.6607 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Guaraldi G, Malagoli A, Theou O, et al. Correlates of frailty phenotype and frailty index and their associations with clinical outcomes. HIV Med. 2017;18(10):764–771. doi: 10.1111/hiv.12527 [DOI] [PubMed] [Google Scholar]
- 20. Battaggia A, Scalisi A, Novelletto BF, Fusello M, Michiele R, Cancian M. Prevalence of frailty in older people in Veneto (Italy). J Drug Assess. 2019;8:1–12. doi: 10.1080/21556660.2018.1563549 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Rubin RR, Gaussoin SA, Peyrot M, et al. ; Look AHEAD Research Group . Cardiovascular disease risk factors, depression symptoms and antidepressant medicine use in the Look AHEAD (Action for Health in Diabetes) clinical trial of weight loss in diabetes. Diabetologia. 2010;53(8):1581–1589. doi: 10.1007/s00125-010-1765-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. The Look AHEAD Research Group. Long-term effects of a lifestyle intervention on weight and cardiovascular risk factors in individuals with type 2 diabetes mellitus: four year results of the Look AHEAD trial. Arch Intern Med. 2010;170(17):1566–1575. doi: 10.1001/archinternmed.2010.334 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Clegg A, Young J, Iliffe S, Rikkert MO, Rockwood K. Frailty in elderly people. Lancet. 2013;381(9868):752–762. doi: 10.1016/S0140-6736(12)62167-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Calderón-Larrañaga A, Fratiglioni L. Multimorbidity research at the crossroads: developing the scientific evidence for clinical practice and health policy. J Intern Med. 2019;285(3):255–271. doi: 10.1111/joim.12872 [DOI] [PubMed] [Google Scholar]
- 25. Wallace LM, Ferrara M, Brothers TD, et al. Lower frailty is associated with successful cognitive aging among older adults with HIV. AIDS Res Hum Retroviruses. 2017;33(2):157–163. doi: 10.1089/AID.2016.0189 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Warwick J, Falaschetti E, Rockwood K, et al. No evidence that frailty modifies the positive impact of antihypertensive treatment in very elderly people: an investigation of the impact of frailty upon treatment effect in the HYpertension in the Very Elderly Trial (HYVET) study, a double-blind, placebo-controlled study of antihypertensives in people with hypertension aged 80 and over. BMC Med. 2015;13:78. doi: 10.1186/s12916-015-0328-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Wei MY, Mukamal KJ. Multimorbidity, mortality, and long-term physical functioning in 3 prospective cohorts of community-dwelling adults. Am J Epidemiol. 2018;1:103–112. doi: 10.1093/aje/kwx198 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Yarnall AJ, Sayer AA, Clegg A, Rockwood K, Parker S, Hindle JV. New horizons in multimorbidity in older adults. Age Ageing. 2017;46(6):882–888. doi: 10.1093/ageing/afx150 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Stolz E, Hoogendijk EO, Mayerl H, Freidl W. Frailty changes predict mortality in 4 longitudinal studies of aging. J Gerontol A Biol Sci Med Sci. 2021;76(9):1619–1626. doi: 10.1093/gerona/glaa266 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Rapp SR, Luchsinger JA, Baker LD, et al. ; Look AHEAD Research Group . Effect of a long-term intensive lifestyle intervention on cognitive function: Action for Health in Diabetes study. J Am Geriatr Soc. 2017;65(5):966–972. doi: 10.1111/jgs.14692 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Espeland MA, Luchsinger JA, Baker LD, et al. ; Look AHEAD Study Group . Effect of a long-term intensive lifestyle intervention on prevalence of cognitive impairment. Neurology. 2017;88(21):2026–2035. doi: 10.1212/WNL.0000000000003955 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Ma C, Avenell A, Bolland M, et al. Effects of weight loss interventions for adults who are obese on mortality, cardiovascular disease, and cancer: systematic review and meta-analysis. BMJ. 2017;359:j4849. doi: 10.1136/bmj.j4849 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Coker RH, Wolfe RR. Weight loss strategies in the elderly: a clinical conundrum. Obesity (Silver Spring). 2018;26(1):22–28. doi: 10.1002/oby.21961 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Bancks MP, Chen H, Balasubramanyam A, et al. Type 2 diabetes subgroups, risk for complications, and differential effects due to an intensive lifestyle intervention. Diabetes Care. 2021;106(5):e2106–e2115. doi: 10.2337/dc20-2372 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Simpson FR, Pajewski NM, Beavers KM, et al. Does the impact of intensive lifestyle intervention on cardiovascular disease risk vary according to frailty as measured via deficit accumulation? J Gerontol A Biol Sci Med Sci. 2021;76(2):339–345. doi: 10.1093/gerona/glaa153 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Barnett K, Mercer SW, Norbury M, et al. Epidemiology of multimorbidity and implications for health care, research, and medical education: a cross-sectional study. Lancet. 2012;360:37–43. doi: 10.1016/S0140-6736(12)60240-2 [DOI] [PubMed] [Google Scholar]
- 37. Xu X, Mishra GD, Jones M. Evidence on multimorbidity from definition to intervention: an overview of systematic reviews. Ageing Res Rev. 2017;37:53–68. doi: 10.1016/j.arr.2017.05.003 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.