Abstract
Objectives
To identify potentially modifiable factors associated with overall and low-quality-of-life (QoL) survival among physically frail older women.
Design
Prospective study with 7 years of follow-up to examine mortality, and among survivors with a QoL measurement within next 3 years to examine low- versus high-QoL patterns of survival.
Setting
Women’s Health Initiative Observational Study (WHI OS).
Participants
11,070 frail women with average age 72.6 (range 65–82).
Measurements
Frailty was defined using modified Fried’s criteria. Overall and low-QoL survival (score <7 on global QoL scale) indicated study outcomes. Risk factors were measured at the first follow-up clinic visit for WHI OS participants in 1997–2001.
Results
Of 11,070 frail women, 1,487 (13%) died. After 2,677 survivors with low or unknown QoL were excluded at study baseline, 3,153 (46%) reported high- and 1,263 (18%) reported low QoL at the end of study follow-up; QoL measures for 2,490 (38%) were unavailable. Older age, history of cardiovascular disease, diabetes, low self-rated health, body mass index under 25 kg/m2, waist circumference over 88cm, systolic blood pressure over 140 mmHg, high somatic symptoms, smoking, and low education were associated with a greater likelihood of low- versus high-QoL survival. Cumulative baseline risk scores demonstrated approximately linear increase in probability of low-QoL survival with an increase in risk factors. The probability of low-QoL survival among those with zero to two risk factors is 0.19 (95% CI 0.15–0.22) versus 0.40 (95% CI 0.35–0.44) among those with six or more risk factors.
Conclusion
We highlighted several potentially modifiable risk factors for overall and low-QoL survival in a large group of aged frail women.
INTRODUCTION
Frailty is currently recognized as a distinct geriatric syndrome with phenotypical representation.1 Frail older adults are at increased risks of disability, morbidity, and mortality compared with their nonfrail counterparts.1–3 The most widely used operational definition of frailty includes indicators of muscle loss, weight change, and fatigue.1 Although frail older adults develop age-related deficits, their clinical outcomes are not universally poor, and vary in progression even among older adults having the same levels of frailty initially. Conceptual model by Fried and colleagues (2001) that outlined pathophysiology of frailty might partially explain this disparity in health outcomes. However, empirical evidence on a constellation of factors associated with overall survival in frail older adults is still lacking. Moreover, because longevity per se has been reported as less important to older adults than attaining high-quality-of-life (QoL) survival, it is important to promote aging well in the face of frailty.
In the current report, we used data from the Women’s Health Initiative Observational Study (WHI OS) participants who screened positively for frailty to examine potential biological, lifestyle, and medication-related factors present at the time of frailty assessment during the period 1997–2001 with regard to overall survival in 7 years of follow-up and among survivors with regard to high- versus low-QoL patterns of survival within the next 3 years of follow-up.
METHODS
Study Population
The Women’s Health Initiative Observational Study (OS) comprised of 93,676 women age 50 to 79 years at baseline (1993–1998) from 40 US clinical centers with the aim of investigating risk factors for several chronic diseases. Details of the study design and baseline characteristics are reported elsewhere.4
At the baseline visit, subjects completed questionnaires on medical and psychosocial characteristics. The first follow-up clinic visit for OS participants was at 3 years and took place during the period 1997–2001. The Year 3 visit is considered the baseline in this analysis. After closeout of the main WHI study in 2005, surviving WHI participants were invited to join the 2005–10 Extension Study (ES). ES participants returned annual mailed forms pertaining to their QoL. Due to differences in WHI recruitment protocols, the lag between main-study closeout and the first available QoL measure in ES participants could be as much as 3 years.
The study sample was based on 43,190 women age 65 to 84 at our baseline who were enrolled in the WHI OS and had available information on five frailty criteria. Of the WHI cohort, 11,070 (26%) OS participants screened positively for frailty and composed the analytical sample. For the purpose of defining vital status, we identified 1,487 frail participants who died and 9,583 who survived within 7 years of follow-up. Among the survivors, the first assessment of QoL was used to indicate high- versus low-QoL categories of survival (Supplementary Figure 1). To examine patterns of survival, we excluded women with unknown or low QoL at baseline to create a cleaner population of women who had the opportunity to demonstrate decline in QoL throughout study follow-up
Frailty Criteria
Frailty was operationally defined consistently with Fried’s1 criteria as the presence of three or more of the following: muscle weakness; slow walking speed, exhaustion, low physical activity, and unintentional weight loss. For each criterion, 1 or 2 points was assigned if the participant’s assessment fell below a criterion-specific cut-point. Then the number of frailty components was then summed with a threshold of three or more defining frailty.1 Poor physical function scored as 2 points because the scale measured the muscle-strength and walking-ability components. Criteria-specific cut-points were based on self-reported measures, as indicated elsewhere.3 Briefly, RAND-36 Physical Function Scale scores below 75 indicated slowness and weakness; RAND-36 Vitality Scale scores below 55 indicated exhaustion; the lowest quartile on the WHI physical-activity questionnaire assessing frequency and duration of walking and mild, moderate, and strenuous activities indicated low physical activity; and weight loss of less than 5% between Year 3 and WHI baseline and a “yes” response to the question “in the past two years, did you lose five or more pounds not on purpose at any time?” indicated shrinking.
Risk-Factor Measures
Data on demographic variables (race/ethnicity [White; NonWhite], highest level of education [High school or less; More than high school]), marital status (Currently married or intimate; Not married or intimate), smoking status (Nonsmoker; Current smoker), and history of cancer (excluding nonmelanoma skin cancer), diabetes, cardiovascular disease, hypertension, and arthritis were obtained by self-report at the WHI baseline. Dietary information, medication regimen, self-rated health, somatic and depressive symptoms, and complete blood count were collected at the study baseline. Diet quality was assessed using the 2010 Alternate Healthy Eating Index (HEI; with scores ranging from 2.5 to 87.5 and higher scores indicating a healthier diet). Lipid-lowering, nonsteroidal anti-inflammatory drugs (NSAIDs) were retrieved using Medication Therapeutic Classes codes from the WHI medication-data repository and used to indicate users versus non-users. Self-rated health was assessed using an ordinal scale with five levels of health ranging from excellent to poor. These categories were then dichotomized to indicate poor/fair versus good/very good/excellent self-rated health. Somatic symptoms were assessed with a 34-item checklist to measure the occurrence and severity of bodily symptoms such as dizziness, heartburn, and neck pain in the last 4 weeks. Scores ranged from 0 to 3; a higher score indicated more-severe symptoms. Depression symptoms were assessed using the short form of the Center for Epidemiological Studies Depression Scale. Scores ranged from 0 to 1; a higher score indicated a greater likelihood of depression with a cut-point of 0.06 and higher, in accordance with Burnam’s algorithm suggestive of depression.5 Physical examination at study baseline included assessments of weight, height, waist circumference (WC), and seated blood pressure. Weight and height were used to compute body mass index (BMI, kg/m2). WC was measured at the natural waist or narrowest part of the torso. Blood pressure was measured by certified staff with a mercury sphygmomanometer after the participant was seated for 5 minutes. Hematocrit (Hct) and hemoglobin measures were analyzed by certified laboratories. Frequency of alcohol consumption was dichotomized to indicate those who consumed alcohol daily and more versus less-frequent drinkers. To facilitate interpretability and clinical relevance of the risk factors, continuous measures such as BMI, WC, HEI, systolic blood pressure (SBP), diastolic blood pressure (DBP), Hct, hemoglobin, and somatic symptoms scores were dichotomized using either specific criteria from national expert panels 6,7 or median split. Dichotomized scores were calculated for the following items: high BMI (≥25 kg/m2), central obesity (WC ≥88 cm), low diet quality (HEI ≤4.1), high SBP (≥140 mmHg), high DBP (≥90 mmHg), low Hct (≤38%), anemia (hemoglobin ≤12.0 gm/dL), and high somatic symptoms (≥0.62).
QoL Measure
The global QoL measure was collected at study baseline and mailed annually to the WHI ES participants. Respondents rated their QoL on 0-to-10 scales, with 0 representing worse QoL and 10 representing best QoL. Based on prior research, a high global QoL was defined as a score above 7.8
All-Cause Mortality
Mortality data were collected annually by mail. Participants’ death records were centrally adjudicated by study physicians using hospital records, autopsy reports, and death certificates. Periodic checks of the National Death Index (NDI) for all participants, including those lost to follow-up, were performed; the latest NDI search, at the time of this analysis, was updated through 2011.
Statistical Approach
Analysis of covariance was used to compare baseline characteristics by vital status and in survivors according to membership in low-, high- or missing-QoL categories, adjusting for age at the study baseline. Age-adjusted odds ratios (ORs) for mortality versus survival and, among survivors, for high versus low QoL were estimated using logistic regression models. Because a large number of variables were considered in the analyses, forward–backward, stepwise selection algorithms were used. Only variables that yielded the best-fitted model in terms of the Akaike information criterion (AIC) were included in the next set of logistic regressions. Stepwise procedures were performed separately for survival versus nonsurvival and high-QoL versus low-QoL models. Furthermore, risk scores based on a simple sum of risk factors highlighted in the stepwise procedures were created. The probability of death and low-QoL survival from age 65 was simulated using coefficients and standard-error estimates from logistic regression adjusted for an average age at the beginning of study follow-up and using dummy variables that corresponded to the levels of risk scores.
The statistical software STATA version 11.2 (StataCorp, College Station, TX) and R version (3.1.2) were used for the statistical analysis.
RESULTS
A total of 11,070 frail women with average age 72.6 (SD = 4.6) met the inclusion criteria for this analysis. Of these, 1,487 died within 7 years of follow-up (average age at death 77.5 [SD = 5.1]). Of the 9,583 survivors, 6,906 (72%) had high QoL at study baseline. Among these, 3,153 (46%) reported high QoL within 7 to 10.2 years of follow-up (average 7.7 years [SD = 0.5]); 1,263 (18%) reported low QoL; and 2,490 (36%) did not have QoL scores available at the follow-up and subsequently were excluded from regression analyses.
Table 1 displays age-adjusted baseline characteristics of the four survival categories: nonsurvivors, survived with high QoL, survived with low QoL, and survived with unknown QoL. Focusing on frailty, nonsurvivors as compared with survivors were more likely to screen positively for low physical activity and unintentional weight loss.
Table 1.
Baseline Characteristics by Pattern of Survival in Frail WHI OS Participants
| Characteristic | Survival Set (n = 11,070)
|
QoL Set (n = 6,906)
a |
||||||
|---|---|---|---|---|---|---|---|---|
| Overall Sample (n = 11,070) | Nonsurvived (n = 1,487) | Survived (n = 9,583) | P-value | High QoL (n = 3,153) | Low QoL (n = 1,263) | QoL NA (n = 2,490) | P-value | |
| Age, yr (SD) | 72.55 (4.56) | 73.9 (4.66) 7 |
72.3 (4.51) 3 |
71.69 (4.25) | 72.4 (4.45) 1 |
73.1 (4.654) | ||
|
| ||||||||
| Comorbidities | ||||||||
|
| ||||||||
| Hypertension, n (%) | 5,39 (49.7) 3 |
838 (57.6) | 4,55 (48.551) | <0.001 | 1,349 (43.4) | 625 (50.6) | 1,20 (49.42) | <0.001 |
|
| ||||||||
| Cardiovascular disease, n (%) | 3,46 (32.0) 0 |
627 (43.2) | 2,83 (30.3) 3 |
<0.001 | 792 (25.6) | 405 (32.8) | 705 (29.2) | <0.001 |
|
| ||||||||
| Diabetes, n (%) | 825 (7.5) | 187 (12.6) | 638 (6.7) | <0.001 | 120 (3.8) | 96 (7.6) | 174 (7.0) | <0.001 |
|
| ||||||||
| Cancer, n (%) | 1,85 (16.9) 4 |
347 (23.5) | 1,50 (15.9) 7 |
<0.001 | 461 (14.7) | 198 (15.6) | 371 (15.0) | 0.664 |
|
| ||||||||
| Arthritis, n (%) | 7,99 (72.9) 5 |
1,05 (71.8) 7 |
6,93 (73.1) 8 |
0.13 | 2,246 (71.6) | 955 (76.1) | 1,75 (71.14) | 0.002 |
|
| ||||||||
| Low self-rated health, n (%) | 3,44 (31.1) 1 |
646 (43.6) | 2,79 (29.2) 5 |
<0.001 | 407 (12.0) | 296 (23.5) | 530 (21.3) | <0.001 |
|
| ||||||||
| Anthropometric and nutrition | ||||||||
|
| ||||||||
| Height, cm (SD) | 159. (6.76) 85 |
159. (6.66) 42 |
159. (6.77) 92 |
0.65 | 160.5 (6.545) | 160. (6.73) 22 |
159. (6.8840) | <0.001 |
|
| ||||||||
| Weight, kg (SD) | 75.3 (17.89) 8 |
72.2 (18.558) | 75.8 (17.794) | <0.001 | 76.70 (17.34) | 76.3 (17.661) | 74.0 (17.081) | 0.003 |
|
| ||||||||
| BMI, kg/m2 (SD) | 29,3 (6.44) 5 |
28.3 (6.91) 4 |
29.5 (6.35) 0 |
<0.001 | 29.62 (6.15) | 29.6 (6.39) 1 |
29.0 (6.110) | 0.1 |
|
| ||||||||
| High BMI (≥25 kg/m2) n (%) | 8,13 (74.0) 6 |
972 (65.6) | 7,16 (75.3) 4 |
<0.001 | 2,413 (77.1) | 944 (75.4) | 1,82 (73.97) | 0.33 |
|
| ||||||||
| Waist circumference, cm (SD) | 90.9 (14.30) 5 |
90.0 (15.141) | 91.1 (14.106) | 0.68 | 91.02 (13.79) | 91.6 (14.332) | 89.9 (13.764) | 0.02 |
|
| ||||||||
| High waist (≥ 88 cm) n (%) | 6,09 (55.8) 3 |
756 (51.6) | 5,33 (56.4) 7 |
0.1 | 1,742 (55.8) | 738 (59.0) | 1,32 (54.02) | 0.03 |
|
| ||||||||
| Healthy Eating Index 2005 (SD) | 3.85 (1.16) | 3.76 (1.21) | 3.86 (1.15) | <0.001 | 3.96 (1.11) | 3.91 (1.11) | 3.85 (1.17) | <0.001 |
|
| ||||||||
| Low HEIb, n (%) | 5,49 (50.0) 5 |
778 (52.7) | 4,71 (49.6) 7 |
0.01 | 1,453 (46.4) | 603 (48.0) | 1,23 (49.82) | 0.02 |
|
| ||||||||
| Hemodynamics | ||||||||
|
| ||||||||
| Systolic blood pressure, mmHg (SD) | 131. (18.05) 18 |
131. (20.2987) | 131. (17.6077) | 0.62 | 129.5 (16.874) | 131. (17.9587) | 132. (17.9247) | <0.001 |
|
| ||||||||
| High systolic pressurec, n (%) | 3,33 (30.1) 4 |
472 (31.8) | 2,86 (29.9) 2 |
0.68 | 829 (26.3) | 397 (31.4) | 793 (31.9) | <0.001 |
|
| ||||||||
| Diastolic blood pressure, mmHg (SD) | 72.1 (9.74) 0 |
71.2 (10.651) | 72.2 (9.58) 4 |
0.03 | 72.28 (9.23) | 72.0 (9.78) 1 |
72.4 (9.684) | 0.05 |
|
| ||||||||
| High diastolic pressured, n (%) | 447 (4.0) | 76 (5.1) | 371 (3.9) | 0.009 | 112 (3.6) | 45 (3.56) | 102 (4.1) | 0.41 |
|
| ||||||||
| Medications | ||||||||
|
| ||||||||
| NSAID med, n (%) | 3,94 (36.2) 5 |
471 (32.1) | 3,47 (36.8) 4 |
0.09 | 1,193 (38.5) | 473 (37.8) | 860 (35.2) | 0.3 |
|
| ||||||||
| Antilipid med, n (%) | 2,56 (24.1) 8 |
319 (22.2) | 2,24 (24.4) 9 |
0.14 | 725 (23.9) | 297 (24.1) | 588 (24.9) | 0.49 |
|
| ||||||||
| Hematological | ||||||||
|
| ||||||||
| Hematocrit, % (SD) | 39.4 (3.75) 7 |
38.8 (4.39) 1 |
39.5 (3.63) 7 |
<0.001 | 39.64 (3.52) | 39.4 (3.79) 9 |
39.5 (3.287) | 0.44 |
|
| ||||||||
| Low hematocrite, n (%) | 3,47 (32.6) 9 |
569 (40.3) | 2,91 (31.5) 0 |
<0.001 | 930 (30.4) | 387 (31.9) | 743 (30.9) | 0.7 |
|
| ||||||||
| Hemoglobin, g/dL (SD) | 13.2 (1.26) 3 |
12.9 (1.61) 8 |
13.2 (1.19) 7 |
<0.001 | 13.31 (1.22) | 13.2 (1.18) 7 |
13.2 (1.117) | 0.52 |
|
| ||||||||
| Anemiaf, n (%) | 1,29 (12.2) 7 |
290 (20.6) | 1,00 (10.9) 7 |
<0.001 | 295 (9.65) | 132 (10.9) | 267 (11.1) | 0.22 |
|
| ||||||||
| Health habits | ||||||||
|
| ||||||||
| Smoke now, n (%) | 558 (5.1) | 127 (8.6) | 431 (4.5) | <0.001 | 88 (2.8) | 54 (4.3) | 139 (5.6) | <0.001 |
|
| ||||||||
| Alcohol daily, n (%) | 847 (14.0) | 140 (18.2) | 707 (13.4) | 0.001 | 276 (14.1) | 113 (15.4) | 166 (12.5) | 0.16 |
|
| ||||||||
| Somatic and mood symptoms | ||||||||
|
| ||||||||
| Depressive symptoms (SD) | 0.06 (0.17) | 0.06 (0.16) | 0.06 (0.17) | 0.45 | 0.03 (0.11) | 0.05 (0.13) | 0.04 (0.12) | <0.001 |
|
| ||||||||
| Depression, n (%) | 1,80 (16.7) 1 |
253 (17.7) | 1,54 (16.6) 8 |
0.07 | 288 (9.3) | 159 (12.9) | 288 (11.9) | <0.001 |
|
| ||||||||
| Somatic symptoms (SD) | 0.65 (0.31) | 0.66 (0.33) | 0.65 (0.31) | 0.2 | 0.57 (0.27) | 0.66 (0.29) | 0.61 (0.29) | <0.001 |
|
| ||||||||
| High symptomsg, n (%) | 4,67 (46.7) 8 |
635 (48.3) | 4,04 (46.5) 3 |
0.08 | 1,074 (36.5) | 553 (48.1) | 914 (41.0) | <0.001 |
|
| ||||||||
| Sociodemographic | ||||||||
|
| ||||||||
| White, n (%) | 9,46 (85.8) 1 |
1,30 (88.2) 6 |
8,15 (85.4) 5 |
0.03 | 2,871 (91.3) | 1,129 (89.89) | 1,98 (79.92) | <0.001 |
|
| ||||||||
| High school or less, n (%) | 4,53 (41.2) 2 |
564 (38.3) | 3,96 (41.7) 8 |
0.01 | 1,060 (33.8) | 521 (41.4) | 1,15 (46.74) | <0.001 |
|
| ||||||||
| Not married, n (%) | 5,49 (49.7) 4 |
867 (58.5) | 4,62 (48.4) 7 |
<0.001 | 1,358 (43.1) | 588 (46.6) | 1,26 (50.92) | 0.001 |
|
| ||||||||
| Frailty criteria | ||||||||
|
| ||||||||
| Low physical functioning, n (%) | 10,9 (99.1) 67 |
1,47 (99.0) 2 |
9,49 (99.1) 5 |
0.55 | 3,123 (99.1) | 1,255 (99.4) | 2,46 (99.06) | 0.51 |
|
| ||||||||
| Low physical activity, n (%) | 6,40 (57.8) 1 |
930 (62.5) | 5,47 (57.1) 1 |
<0.001 | 1,831 (58.1) | 668 (52.9) | 1,47 (59.11) | <0.001 |
|
| ||||||||
| Low energy, n (%) | 8,33 (75.3) 1 |
1,13 (76.3) 4 |
7,19 (75.1) 7 |
0.23 | 2,070 (65.7) | 979 (77.5) | 1,71 (69.09) | <0.001 |
|
| ||||||||
| Weight loss, n (%) | 1,70 (15.4) 1 |
382 (25.7) | 1,31 (13.8) 9 |
<0.001 | 357 (11.3) | 164 (13.0) | 387 (15.5) | 0.002 |
WHI OS = Women’s Health Initiative Observational Study; QoL = Quality of life; NA = Not available; BMI = Body Mass Index; HEI = Healthy Eating Index; NSAID = Nonsteroidal anti-inflammatory drug; Med = Medication; SD= Standard deviation
After excluding 2,677 participants with unknown or low QoL at frailty screening;
low diet quality (HEI ≤4.1);
high systolic blood pressure (≥140 mmHg);
high diastolic blood pressure (≥90 mmHg);
low hematocrit (≤38%);
anemia (hemoglobin ≤12.0 gm/dL);
high somatic symptoms (≥0.62).
Supplementary Table 1 shows age-adjusted ORs from logistic regression analyses for dichotomous factors that were associated with nonsurvival versus survival, and among survivors with having low versus high QoL. Consistently high ORs for nonsurvival and low-QoL survival were associated with being a smoker. Anthropometrically and nutritionally, increased mortality was associated with having BMI below 25 kg/m2 and worse nutritional quality as expressed by a low HEI 2010 index, whereas central, but not overall, obesity was associated with low-QoL survival. Depression and high somatic symptoms were consistently associated with low-QoL survival without being associated with mortality.
Table 2 shows variables that yielded the best-fitted model after stepwise procedure in terms of the AIC. Older age, history of cardiovascular disease, diabetes, low self-rated health, BMI below 25, and smoking were significantly associated with greater likelihood of nonsurvival and low-QoL survival. Central obesity, high SBP, and high somatic symptoms were uniquely selected for low- versus high-QoL models.
Table 2.
Stepwise Logistic Regression Model of Risk of Death or Low-QoL Survival in Frail WHI OS Participants
| Characteristic | Nonsurvival vs. Survival (n = 1,487 vs. 9,583) | P-value | Low vs. High QoL (n = 1,263 vs. 3,153) | P-value |
|---|---|---|---|---|
| OR (95% CI) | OR (95% CI) | |||
| Age, yr | 1.08 (1.06–1.09) | <0.001 | 1.04 (1.02–1.06) | <0.001 |
| Comorbidities | ||||
| Hypertension | 1.51 (1.31–1.73) | <0.001 | not selected | |
| Cardiovascular disease | 1.51 (1.32–1.73) | <0.001 | 1.28 (1.09–1.51) | 0.003 |
| Diabetes | 1.82 (1.45–2.28) | <0.001 | 1.67 (1.21–2.31) | 0.002 |
| Cancer | 1.45 (1.23–1.69) | <0.001 | not selected | |
| Arthritis | 0.83 (0.72–0.97) | 0.02 | 1.20 (1.00–1.43) | 0.04 |
| Low self-rated health | 1.62 (1.41–1.86) | <0.001 | 1.88 (1.54–2.29) | <0.001 |
| Anthropometric and nutrition | ||||
| High BMI (≥25) | 0.67 (0.59–0.79) | <0.001 | 0.75 (0.60–0.94) | 0.01 |
| High waist (≥ 88) | not selected | 1.33 (1.10–1.62) | 0.003 | |
| Low HEIa | 1.15 (1.00–1.31) | 0.05 | not selected | |
| Hemodynamics | ||||
| High systolic pressureb | not selected | 1.20 (1.02–1.42) | 0.03 | |
| High diastolic pressurec | 1.43 (1.04–1.94) | 0.02 | not selected | |
| Medications | ||||
| NSAID med | 0.90 (0.78–1.04) | 0.14 | not selected | |
| Antilipid med | 0.74 (0.63–0.87) | <0.001 | not selected | |
| Hematological | ||||
| Low hematocritd | not selected | not selected | ||
| Anemiae | 1.97 (1.65–2.34) | <0.001 | not selected | |
| Health habits | ||||
| Smoke now | 2.45 (1.89–3.15) | <0.001 | 1.56 (1.02–2.38) | 0.03 |
| Somatic and mood symptoms | ||||
| Depression | not selected | not selected | ||
| High symptomsf | not selected | 1.42 (1.22–1.66) | <0.001 | |
| Sociodemographic | ||||
| White | 1.49 (1.20–1.86) | <0.001 | not selected | |
| High school or less | 0.84 (0.73–0.97) | 0.01 | 1.25 (1.07–1.46) | <0.001 |
| Not married | 1.26 (1.09–1.44) | <0.001 | not selected | |
WHI OS = Women’s Health Initiative Observational Study; QoL = Quality of life; OR = Odds Ratio; CI = Confidence Interval; BMI = Body Mass Index; HEI = Healthy Eating Index; NSAID = Nonsteroidal anti-inflammatory drug; Med = Medication.
low diet quality (HEI ≤4.1);
high systolic blood pressure (≥140 mmHg);
high diastolic blood pressure (≥90 mmHg);
low hematocrit (≤38%);
anemia (hemoglobin ≤12.0 gm/dL);
high somatic symptoms (≥0.62).
Table 3 shows the estimated probabilities that an average-age 72.5-year-old frail WHI study participant would survive in the next 7 years of follow-up according to cumulative risk factors from baseline examinations. The probability of nonsurvival for those with zero to two risk factors is 0.06 (95% CI 0.04–0.07), and it increases to 0.60 (95% CI 0.15–1.03) for those with more than nine risk factors. Similarly, the probability of low-QoL survival for those with fewer than two risk factors is 0.19 (95% CI 0.15–0.22) and increases to 0.40 (95% CI 0.35–0.44) in those with six or more risk factors.
Table 3.
Age-Adjusted Probabilities of Survival and Low-QoL Survival by Risk Scores in Frail WHI OS Participants
| Number of Risk Factors | Risk of Deatha PP (95% CI) |
P-value | Number of Risk Factors | Risk of Low QoLb PP (95% CI) |
P-value |
|---|---|---|---|---|---|
| ≤2 | 0.06 (0.04–0.07) | <0.001 | <2 | 0.19 (0.15–0.22) | <0.001 |
| 3–4 | 0.11 (0.10–0.12) | <0.001 | 2–3 | 0.27 (0.24–0.29) | <0.001 |
| 5–6 | 0.15 (0.14–0.16) | <0.001 | 4–5 | 0.29 (0.26–0.31) | <0.001 |
| 7–8 | 0.26 (0.22–0.30) | <0.001 | >6 | 0.40 (0.35–0.44) | <0.001 |
| >9 | 0.60 (0.16–1.04) | 0.008 |
QoL = Quality of life; WHI OS = Women’s Health Initiative Observational Study; PP = Predicted Probabilities; CI = Confidence interval
In death set, risk factors are history of hypertension, cardiovascular disease, diabetes, cancer, arthritis, fair/poor self-rated health, normal/underweight, low health eating index, high diastolic pressure, no anti-lipids medication therapy, anemia, smoking, white, not married.
In QoL set, risk factors are history of cardiovascular disease, diabetes, arthritis, fair/poor self-rated health, normal/underweight, high waist, high systolic pressure, smoking, high somatic symptoms, high school education or less.
DISCUSSION
In the present study we focused on the frail population to explore predictors for longevity and, more important, high-QoL survival. Not surprisingly, older age, a history of chronic conditions such as cardiovascular disease and diabetes, and smoking were significantly associated with greater likelihood of both nonsurvival and low-QoL survival. These findings resonated in other studies.9–11
We also showed that low self-rated health was an independent predictor of both nonsurvival and low-QoL survival. Self-rated health has long been known as a valid reflection of health status and as a robust predictor of mortality in older adults.12,13 Interestingly, the association between self-rated health and mortality persists after adjusting for indicators of objective health, implying an independent predictive value of self-health perception, as demonstrated in the past.12
A higher BMI (≥25) was significantly associated with both lower likelihood of nonsurvival and low-QoL survival. These findings are in line with several recent publications and meta analyses indicating that overweight in older adults is not a risk factor and may even protect against mortality.14–16 Despite the protective effect of being overweight, higher waist circumference, an indicator of central visceral adipose tissue deposition,17 was associated with low-QoL survival, emphasizing its independent role in metabolic and health statuses.18,19
Marital status was associated only with mortality, in that being unmarried was a risk factor for nonsurvival but was not related to high or low QOL survival. Similarly, in men, being unmarried was associated with an increased risk of nonsurvival, but having a marital partner did not improve the odds of being healthier in very old age.19,20 Being unmarried may lead to increased feeling of loneliness, which among participants 60 and older was predictive of mortality.21
Although depression was associated with low QoL in the age-adjusted model, multivariate adjustment attenuated the strength of this relationship. Others found depression to be strongly correlated with QoL in old age;22 therefore, it is plausible to assume that QoL may be affected more by concurrent than by past depressive symptoms.
Surprisingly, educational attainment had opposite effects on survival and QoL: less education was associated with a decreased risk of death but with low-QoL survival. In previous research, less education was linked to lower health-related QoL23 and lower odds of exceptional survival, whereas no association was demonstrated with general survival.19,20 Higher educational attainment might serve as a proxy for higher socioeconomic status, which in turn might not warrant longer life but may help those who survive maintain better health and a higher QoL.
The positive dose-response relationship between the number of risk factors at older age and the probability of both nonsurvival and low-QoL survival in the next decade indicates their cumulative effects. Similar associations have been demonstrated in men with regard to overall and healthy survival20 and with regard to exceptional longevity to age 90.11 These findings emphasize the importance of prevention and education to a healthy lifestyle, since every additional risk factor counts in shaping the course and nature of late-life survival.
The strengths of the study are its large sample size, long follow-up, and data completeness. However, this study has several potential limitations. First, since our study population consists of women only, the external validity of these findings for men should be confirmed in another study. Second, although we adjusted for many factors potentially associated with survival and QoL, residual or hidden confounding may still exist. Third, because not all main-study participants were willing to join the ES, extensive missingness in QoL measures at follow-up might represent, on one hand, a loss of follow-up due to study protocols. On the other hand, it is also plausible that frail women with lower levels of functioning were less likely than their higher-functioning counterparts to be interested in the ES. Loss to follow-up and exclusion of women with low-QoL at the study baseline may underestimate the proportion of older women at risk of being in low-QoL categories.
In conclusion, we have identified several potentially important risk factors, some of which are modifiable, for aging well survival in a large group of frail old women. These can be monitored in both clinical and research settings and may also be modifiable for clinical, health policy, and epidemiological purposes.
Supplementary Material
Acknowledgments
The authors thank the Women’s Health Initiative (WHI) investigators and staff for their dedication and the study participants for making the program possible. A listing of WHI investigators can be found at: https://cleo.whi.org/researchers/Documents%20%20Write%20a%20Paper/WHI%20Investigator%20Short%20List.pdf The WHI program is funded by the National Heart, Lung and Blood Institute, National Institute of Health, U.S. Department of Health, and Human Services through Contracts N01WH22110, 24152, 32100-2, 32105-6, 32108-9, 32111-13, 32115, 32118-32119, 32122, 42107-26, 42129-32, and 44221.
Footnotes
Conflict of Interest: The editor in chief has reviewed the conflict of interest checklist provided by the authors and has determined that the authors have no financial or any other kind of personal conflicts with this paper.
Author Contributions
Drs. Zaslavsky and Zelber-Sagi had full access to all of the data used in the study and take responsibility for the integrity of the data and accuracy of data analysis. Zaslavsky, Woods and Zelber-Sagi: study concept, design, interpretation of data, and preparation of manuscript. Zaslavsky and Zelber-Sagi,: data analysis and visualization. Zaslavsky, Woods, LaCroix, Cauley, Johnson, Cochrane and Zelber-Sagi: preparation of manuscript.
Sponsor’s Role
The funding agencies had no role in the design and conduct of this study, the analysis or interpretation of the data, or the preparation of the manuscript.
References
- 1.Fried LP, Tangen CM, Walston J, et al. Frailty in older adults: Evidence for a phenotype. J Gerontol A Biol Sci Med Sci. 2001;56:M146–156. doi: 10.1093/gerona/56.3.m146. [DOI] [PubMed] [Google Scholar]
- 2.Ensrud KE, Ewing SK, Taylor BC, et al. Frailty and risk of falls, fracture, and mortality in older women: The study of osteoporotic fractures. J Gerontol A Biol Sci Med Sci. 2007;62:744–751. doi: 10.1093/gerona/62.7.744. [DOI] [PubMed] [Google Scholar]
- 3.Woods NF, Cochrane BB, LaCroix AZ, et al. Toward a positive aging phenotype for older women: Observations from the Women’s Health Initiative. J Gerontol A Biol Sci Med Sci. 2012;67:1191–1196. doi: 10.1093/gerona/gls117. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Anderson GL, Manson J, Wallace R, et al. Implementation of the Women’s Health Initiative study design. Ann Epidemiol. 2003;13:S5–S17. doi: 10.1016/s1047-2797(03)00043-7. [DOI] [PubMed] [Google Scholar]
- 5.Burnam MA, Wells KB, Leake B, et al. Development of a brief screening instrument for detecting depressive disorders. Med Care. 1988;26:775–789. doi: 10.1097/00005650-198808000-00004. [DOI] [PubMed] [Google Scholar]
- 6.Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure, National Institutes of Health (U.S.), National Heart, Lung, and Blood Institute et al. The seventh report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure. Bethesda, MD: 2004. [PubMed] [Google Scholar]
- 7.National Institutes of Health. Clinical guidelines on the identification, evaluation, and treatment of overweight and obesity in adults: The evidence report. Obes Res. 1998;6:51S–209S. [PubMed] [Google Scholar]
- 8.Guyatt GH, Osoba D, Wu AW, et al. Methods to explain the clinical significance of health status measures. Mayo Clin Proc. 2002;77:371–383. doi: 10.4065/77.4.371. [DOI] [PubMed] [Google Scholar]
- 9.Roos NP, Havens B. Predictors of successful aging: A twelve-year study of Manitoba elderly. Am J Public Health. 1991;81:63–68. doi: 10.2105/ajph.81.1.63. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Kempen GI, Brilman EI, Ranchor AV, et al. Morbidity and quality of life and the moderating effects of level of education in the elderly. Soc Sci Med. 1999;49:143–149. doi: 10.1016/s0277-9536(99)00129-x. [DOI] [PubMed] [Google Scholar]
- 11.Yates LB, Djousse L, Kurth T, et al. Exceptional longevity in men: Modifiable factors associated with survival and function to age 90 years. Arch Intern Med. 2008;168:284–290. doi: 10.1001/archinternmed.2007.77. [DOI] [PubMed] [Google Scholar]
- 12.Mossey JM, Shapiro E. Self-rated health: A predictor of mortality among the elderly. Am J Public Health. 1982;72:800–808. doi: 10.2105/ajph.72.8.800. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Miilunpalo S, Vuori I, Oja P, et al. Self-rated health status as a health measure: The predictive value of self-reported health status on the use of physician services and on mortality in the working-age population. J Clin Epidemiol. 1997;50:517–528. doi: 10.1016/s0895-4356(97)00045-0. [DOI] [PubMed] [Google Scholar]
- 14.Flegal KM, Kit BK, Orpana H, et al. Association of all-cause mortality with overweight and obesity using standard body mass index categories: A systematic review and meta-analysis. JAMA. 2013;309:71–82. doi: 10.1001/jama.2012.113905. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Hughes V. The big fat truth. Nature. 2013;497:428–430. doi: 10.1038/497428a. [DOI] [PubMed] [Google Scholar]
- 16.Winter JE, MacInnis RJ, Wattanapenpaiboon N, et al. BMI and all-cause mortality in older adults: A meta-analysis. Am J Clin Nutr. 2014;99:875–890. doi: 10.3945/ajcn.113.068122. [DOI] [PubMed] [Google Scholar]
- 17.Pouliot MC, Despres JP, Lemieux S, et al. Waist circumference and abdominal sagittal diameter: Best simple anthropometric indexes of abdominal visceral adipose tissue accumulation and related cardiovascular risk in men and women. Am J Cardiol. 1994;73:460–468. doi: 10.1016/0002-9149(94)90676-9. [DOI] [PubMed] [Google Scholar]
- 18.Wannamethee SG, Shaper AG, Morris RW, et al. Measures of adiposity in the identification of metabolic abnormalities in elderly men. Am J Clin Nutr. 2005;81:1313–1321. doi: 10.1093/ajcn/81.6.1313. [DOI] [PubMed] [Google Scholar]
- 19.Bell CL, Chen R, Masaki K, et al. Late-life factors associated with healthy aging in older men. J Am Geriatr Soc. 2014;62:880–888. doi: 10.1111/jgs.12796. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Willcox BJ, He Q, Chen R, et al. Midlife risk factors and healthy survival in men. JAMA. 2006;296:2343–2350. doi: 10.1001/jama.296.19.2343. [DOI] [PubMed] [Google Scholar]
- 21.Perissinotto CM, Stijacic Cenzer I, Covinsky KE. Loneliness in older persons: A predictor of functional decline and death. Arch Intern Med. 2012;172:1078–1083. doi: 10.1001/archinternmed.2012.1993. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Brett CE, Gow AJ, Corley J, et al. Psychosocial factors and health as determinants of quality of life in community-dwelling older adults. Qual Life Res. 2012;21:505–516. doi: 10.1007/s11136-011-9951-2. [DOI] [PubMed] [Google Scholar]
- 23.Chapman B, Duberstein P, Lyness JM. Personality traits, education, and health-related quality of life among older adult primary care patients. J Gerontol B Psychol Sci Soc Sci. 2007;62:P343–352. doi: 10.1093/geronb/62.6.p343. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.