The Combined Effect of Cancer and Cardiometabolic Conditions on the Mortality Burden in Older Adults

The Combined Effect of Cancer and Cardiometabolic Conditions on the Mortality Burden in Older Adults Abstract Background The number of older people living with cancer and cardiometabolic conditions is increasing, but little is known about how specific combinations of these conditions impact mortality. Methods A total of 22,692 participants aged 65 years and older from four international cohorts were followed-up for mortality for an average of 10 years (8,596 deaths). Data were harmonized across cohorts and mutually exclusive groups of disease combinations were created for cancer, myocardial infarction (MI), stroke, and diabetes at baseline. Cox proportional hazards models for all-cause mortality were used to estimate the age- and sex-adjusted hazard ratio and rate advancement period (RAP) (in years). Results At baseline, 23.6% (n = 5,116) of participants reported having one condition and 4.2% (n = 955) had two or more conditions. Data from all studies combined showed that the RAP increased with each additional condition. Diabetes advanced the rate of dying by the most years (5.26 years; 95% confidence interval [CI], 4.53–6.00), but the effect of any single condition was smaller than the effect of disease combinations. Some combinations had a significantly greater impact on the period by which the rate of death was advanced than others with the same number of conditions, for example, 10.9 years (95% CI, 9.4–12.6) for MI and diabetes versus 6.4 years (95% CI, 4.3–8.5) for cancer and diabetes. Conclusions Combinations of cancer and cardiometabolic conditions accelerate mortality rates in older adults differently. Although most studies investigating mortality associated with multimorbidity used disease counts, these provide little guidance for managing complex patients as they age. Multimorbidity, Epidemiology, Risk factor, Aging The global population of people aged 60 years and older is increasing in almost all regions of the world and is projected to double in size over the next 35 years, reaching nearly 2.1 billion by 2050 (1–3). This transition is accompanied by an increase in prevalence of many chronic diseases (1,4), with most older adults having two or more conditions coexisting at the same time (5,6). Of these, cancer, ischemic heart disease, stroke, and diabetes were identified as the main drivers of mortality in developed countries by the Global Burden of Disease Study 2013, accounting for almost 25 million deaths worldwide (1). The accumulation of chronic conditions serves as an important indicator for the progressive loss of resilience and functional independence (7) and has been suggested as an early marker of accelerated biological aging (8). Furthermore, having two or more chronic conditions has been shown to be an independent predictor of mortality (7) and is associated with marked reductions in life expectancy (9). Researchers, however, seldom examine the impact of specific combinations of chronic conditions on survival, even though results from single-disease studies suggest that specific disease combinations are likely to have a much stronger association with mortality than others. Therefore, better understanding the consequences of specific combinations of chronic conditions on mortality is a critical first step toward optimizing the public health and health care needs of older adults. The purpose of this study is to determine how specific combinations of the main drivers of mortality, including cancer, myocardial infarction (MI), stroke, and diabetes, accelerate mortality rates in older adults. Methods Study Design and Participants Our study uses longitudinal individual-level harmonized data from four cohort studies, three from the Consortium on Health and Ageing: Network of Cohorts in Europe and United States (CHANCES) (10), and one from Canada, the Canadian Study on Health and Aging (CSHA). The cohorts’ key design characteristics are summarized in Table 1. Table 1. Characteristics of the Four Cohort Studies for the Analyses on Cancer and Cardiometabolic Multimorbidity and All-Cause Mortality Studya Countryb Period of enrollment Mortality follow-up Participants 65+ y, N Complete mortality FU, N Complete chronic disease data, N Complete confounder data, Nc EPIC-Elderly (13) (selected centers) DK, GR, NL, ES, SE 1992–2000 1992–2011 10,309 10,079 9,956 9,771 ESTHER (14) DE 2000–2002 2000–2013 3,845 3,842 3,421 3,228 Tromsø Study (15) NO 1994–1995 1994–2010 4,286 4,286 4,251 4,211 CSHA (16) CA 1991 1991–2001 9,008 8,743 5,627d 5,482 Studya Countryb Period of enrollment Mortality follow-up Participants 65+ y, N Complete mortality FU, N Complete chronic disease data, N Complete confounder data, Nc EPIC-Elderly (13) (selected centers) DK, GR, NL, ES, SE 1992–2000 1992–2011 10,309 10,079 9,956 9,771 ESTHER (14) DE 2000–2002 2000–2013 3,845 3,842 3,421 3,228 Tromsø Study (15) NO 1994–1995 1994–2010 4,286 4,286 4,251 4,211 CSHA (16) CA 1991 1991–2001 9,008 8,743 5,627d 5,482 Note: aEPIC: European Prospective Investigation Into Cancer and Nutrition; ESTHER: Epidemiological Study on the Chances of Prevention, Early Recognition and Optimised Treatment of Chronic Diseases in the Older Population; CSHA: Canadian Study on Health and Aging. bCanada (CA); Denmark (DK); Germany (DE); Greece (GR); Netherlands (NL); Norway (NO); Spain (ES); Sweden (SE). cData on smoking (ever smoker, never smoker) and level of education (less than high school, high school, more than high school). dOne thousand four hundred and thirty-four CSHA participants did not complete the risk factor questionnaire and another 1,682 participants did not have complete exposure data (i.e. history of the four chronic conditions under study). View Large Table 1. Characteristics of the Four Cohort Studies for the Analyses on Cancer and Cardiometabolic Multimorbidity and All-Cause Mortality Studya Countryb Period of enrollment Mortality follow-up Participants 65+ y, N Complete mortality FU, N Complete chronic disease data, N Complete confounder data, Nc EPIC-Elderly (13) (selected centers) DK, GR, NL, ES, SE 1992–2000 1992–2011 10,309 10,079 9,956 9,771 ESTHER (14) DE 2000–2002 2000–2013 3,845 3,842 3,421 3,228 Tromsø Study (15) NO 1994–1995 1994–2010 4,286 4,286 4,251 4,211 CSHA (16) CA 1991 1991–2001 9,008 8,743 5,627d 5,482 Studya Countryb Period of enrollment Mortality follow-up Participants 65+ y, N Complete mortality FU, N Complete chronic disease data, N Complete confounder data, Nc EPIC-Elderly (13) (selected centers) DK, GR, NL, ES, SE 1992–2000 1992–2011 10,309 10,079 9,956 9,771 ESTHER (14) DE 2000–2002 2000–2013 3,845 3,842 3,421 3,228 Tromsø Study (15) NO 1994–1995 1994–2010 4,286 4,286 4,251 4,211 CSHA (16) CA 1991 1991–2001 9,008 8,743 5,627d 5,482 Note: aEPIC: European Prospective Investigation Into Cancer and Nutrition; ESTHER: Epidemiological Study on the Chances of Prevention, Early Recognition and Optimised Treatment of Chronic Diseases in the Older Population; CSHA: Canadian Study on Health and Aging. bCanada (CA); Denmark (DK); Germany (DE); Greece (GR); Netherlands (NL); Norway (NO); Spain (ES); Sweden (SE). cData on smoking (ever smoker, never smoker) and level of education (less than high school, high school, more than high school). dOne thousand four hundred and thirty-four CSHA participants did not complete the risk factor questionnaire and another 1,682 participants did not have complete exposure data (i.e. history of the four chronic conditions under study). View Large The CHANCES Consortium The CHANCES consortium is a large collaborative project which harmonized data from on-going prospective cohort studies in Europe and the United States to study aging-related health characteristics and determinants of healthy aging (10). Variables were harmonized using predetermined standardized procedures (11). Three cohorts were included in the present analyses: the European Prospective Investigation Into Cancer and Nutrition–Elderly (EPIC-Elderly) Study (12) from Spain, the Netherlands, Greece, Sweden, and Denmark; the Epidemiological Study on the Chances of Prevention, Early Recognition and Optimised Treatment of Chronic Diseases in the Older Population (ESTHER) Study (13) from Germany; and the Tromsø Study (14) from Norway. The current analysis includes all individuals 65 years and older who completed a self-administered questionnaire on lifestyle characteristics, chronic disease status, and potential chronic disease risk factors at baseline (Table 1) (n = 17,210; 9,771 [EPIC-Elderly], 3,228 [ESTHER], and 4,211 [Tromsø]). The CSHA The CSHA is a national, population-based study of cognitive impairment and other aspects of health in Canadian adults aged 65 years and older (15). The CSHA consists of 10,263 participants in community (n = 9,008) and institutional settings (n = 1,255) who were recruited from the 10 Canadian provinces. In the first wave of the CSHA in 1991, face-to-face interviews were conducted with all community participants to screen for dementia. Participants who were cognitively normal were asked to complete a self-administered risk factor questionnaire on demographic characteristics, lifestyle, and chronic disease status. The risk factor questionnaire was completed by a proxy for participants found to be cognitively impaired. CSHA variables were harmonized following CHANCES procedures. The current analysis includes community-dwelling individuals who completed a risk factor questionnaire (n = 5,482; 4,751 self-administered and 731 proxy-administered) and had complete data. In all cohorts, the research procedures were in accordance with the ethical standards of the responsible institutional or regional committees on human experimentation, and all participants provided informed consent. Combinations of Diseases All cohorts collected self-reported information on history of cancer, MI, stroke, and diabetes as part of the baseline questionnaires. Although clinical diagnosis by a physician is considered the standard criterion for ascertaining disease endpoints, large-scale epidemiological studies often rely on self-reported diagnoses. We were able to address the accuracy of our data by comparing the agreement between self-reported disease status to physicians diagnosis in a subset of the CSHA population (n = 1,068) who underwent an additional clinical examination at baseline. Moderate-to-good agreement was observed with κ’s ranging from 0.35 (heart disease) to 0.87 (diabetes). These findings are in line with previous research showing that self-reported diagnosis of the conditions under study compare relatively well to clinical diagnoses (16–18). As well, to assess the generalizability of the sample, disease prevalence estimates from the included studies were compared against representative national survey data (also using self-reports to define morbidity). Prevalence rates were in the same order of magnitude in CSHA and the 1991 General Social Survey (19), and the CHANCES cohorts and first European Health Interview Survey (2006–2008). Participants were classified into mutually exclusive groups of chronic disease combinations at baseline (i.e. single, dyads, and triads of conditions). The four conditions under study resulted in 16 unique disease combinations, including the reference category with none of these four conditions. The independent effect of conditions with less than 15 participants was not reported due to low power. Mortality Follow-Up EPIC-Elderly, ESTHER, and Tromsø were followed up for vital status by linkage to region, state, or country-wide death registries with coverage of ≥98.5%. The median follow-up period ranged from 12 years (ESTHER and EPIC), to 16 years (Tromsø). CSHA participants were followed up for vital status after 10 years, and date of death was recorded for individuals who were deceased. Missing dates of death for CSHA decedents were derived as previously described (20). Less than 3% of CSHA participants were lost to follow-up. Statistical Analysis Statistical analyses were performed separately for each study cohort (treating EPIC-Elderly as one cohort), for all CHANCES cohorts combined, and CHANCES and CSHA cohorts combined by pooling individual-level data. The aging effect of chronic disease combinations on mortality was estimated using rate advancement periods (RAP). The derivation of the RAP has been described in detail by Brenner and colleagues (21). The RAP was derived from a multivariable Cox proportional hazards models of the form: hazard (t, exposure) = b1 × disease combination + b2 × age + covariates. From this model, a point estimate of the RAP is obtained as b1/b2 (i.e. the ratio of the regression coefficients for the specific disease combination [numerator] and age at baseline [denominator]). The RAP specifies the period by which the rate of death is advanced among subjects with a specific chronic disease combination relative to the reference group in the absence of competing risks. In other words, it specifies the loss of years in terms of mortality risk associated with having a specific combination of chronic conditions. The RAP thus quantifies the impact of a specific combination of conditions on mortality relative to the effect of aging. For example, a RAP of 10 years for a specific exposure (E) means that means that subjects with that exposure (E+) have the same risk of dying as someone without that exposure (E−) who is 10 years their senior. Confidence intervals (CIs) for the RAPs were calculated using Fieller’s theorem. Sensitivity analyses were performed to test the robustness of results using a frailty model that included study as a random effect. Our primary analyses calculated hazard ratios (HRs) adjusted for age and sex only. We also explored the extent to which other demographic and lifestyle factors explained the associations between disease combinations and mortality. Variables that changed any RAP estimate by 10% or more were included as confounders in the multivariable analyses resulting in a final model adjusted for tobacco smoking (never, ever) and education (<high school, high school, >high school). CHANCES cohorts included information on anthropometry and blood pressure allowing us to investigate the effect of body mass index (<18.5, 18.5–24.9, 25.0–29.9, ≥30 kg/m2), and hypertension (diastolic ≥90 mmHg and/or systolic ≤140 mmHg) on the RAP estimates, whereas CSHA enabled us to examine the explanatory effect of disability status at baseline (combined presence or absence of self-reported limitations in seven activities of daily living). Possible effect modification by age (stratified by median age of 70 years), sex, and the previously mentioned covariates were tested using interaction terms in the Cox regression model. Because the denominator of the RAP estimates (i.e. age at baseline) is similar for each RAP derived from the same model, the difference in RAP estimates result from a difference in the regression estimate of the disease combinations. We evaluated if any two RAPs differed from each other by calculating Z-scores comparing the parameter estimates of unique combinations of conditions. Results from all pairwise comparisons are depicted using a heatmap-method. All analyses were performed using SAS V9.4. Results In total, 22,692 participants aged 65 years and older were included in this study (Table 2), with 8,708 (38%) deaths recorded during follow-up (Table 3). The average follow-up duration was between 7 and 11 years. Participants from EPIC-Elderly and ESTHER were younger than those from the Tromsø and CSHA cohorts. All studies included more women than men (overall 62% women), and 47% of participants were ever smokers. The highest level of education varied across studies; 26% of CSHA participants had more than a high school degree compared to 5.7% in the CHANCES studies. At enrolment, 1,928 (8.5%) participants reported having cancer, 1,730 (7.6%) MI, 1,062 (4.7%) stroke, and 2,408 (10.6%) diabetes, respectively. Disease rates were generally higher in the older cohorts (CSHA, Tromsø) compared to younger cohorts (EPIC-Elderly and ESTHER). The majority of participants had none (73.3%) or only one (22.6%) of the four chronic conditions under study. Table 2. Main Characteristics of the Study Participants by Cohort All studies combined CSHA CHANCES cohorts EPIC-Elderly Tromsø ESTHER Combined N 22,692 5,482 9,771 4,211 3,228 17,210 Age (y) (mean (SD)) 71.3 (5.9) 76.5 (7.0) 68.7 (3.2) 72.8 (5.8) 68.8 (2.9) 69.7 (4.3)  65–74 (N, %) 17,514 (77.2) 2,337 (42.6) 9,240 (94.6) 2,759 (65.5) 3,178 (98.5) 15,177 (88.2)  75–84 4,165 (18.4) 2,305 (42.1) 530 (5.4) 1,280 (30.4) 50 (1.6) 1,860 (10.8)  85+ 1,013 (4.5) 840 (15.3) 1 (0.0) 172 (4.1) 0 (0.0) 173 (1.0) Sex (N, %)  Men 8,639 (38.1) 3,133 (57.2) 2,964 (30.3) 1,795 (42.6) 1,531 (47.4) 6,290 (36.6)  Women 14,053 (61.9) 2,349 (42.9) 6,807 (69.7) 2,416 (57.4) 1,697 (52.6) 10,920 (63.5) Education (N, %)  Less than high school 14,344 (63.2) 1,757 (32.1) 7,290 (74.6) 2,850 (67.8) 2,447 (75.8) 12,587 (73.1)  High school 5,942 (26.2) 2,292 (41.8) 1,977 (20.2) 1,039 (24.7) 634 (19.6) 3,650 (21.2)  More than high school 2,406 (10.6) 1,433 (26.1) 504 (5.2) 322 (7.7) 147 (4.6) 973 (5.7) Smoking status (N, %)  Ever smoker 10,661 (47.0) 2,792 (50.9) 3,807 (39.0) 2,604 (61.8) 1,458 (45.2) 7,869 (45.7)  Never smoker 12,031 (53.0) 2,690 (49.1) 5,964 (61.0) 1,607 (38.2) 1,770 (54.8) 9,341 (54.3) Average duration of mortality follow-up (y) 10.02 (±3.9) 7.43 (±3.0) 10.83 (±3.4) 11.09 (±5.2) 10.58 (±2.6) 10.84 (±3.8) Self-reported history of chronic conditions (N, %)  History of cancer 1,928 (8.5) 747 (13.6) 482 (4.9) 399 (9.5) 300 (9.3) 1,181 (6.9)  History of MI 1,730 (7.6) 625 (11.4) 420 (4.3) 449 (10.7) 236 (7.3) 1,105 (6.4)  History of stroke 1,062 (4.7) 384 (7.0) 278 (2.9) 239 (5.7) 161 (5.0) 678 (3.9)  History of diabetes 2,408 (10.6) 535 (9.8) 1,180 (12.1) 252 (6.0) 411 (13.7) 1,873 (10.9) Disease counts (N, %)  No disease 16,621 (73.3) 3,615 (65.9) 7,660 (78.4) 3,059 (72.6) 2,287 (70.9) 13,006 (75.6)  1 disease 5,116 (22.6) 1,491 (27.2) 1,881 (19.3) 979 (23.3) 765 (23.7) 3,625 (21.1)  2 diseases 858 (3.8) 332 (6.1) 211 (2.2) 160 (3.8) 155 (4.8) 526 (3.1)  3 diseases 92 (0.4) 40 (0.7) 19 (0.2) 12 (0.3) 21 (0.7) 52 (0.3)  4 diseases 5 (0.0) 4 (0.1) 0 (0.0) 1 (0.0) 0 (0.0) 1 (0.1) All studies combined CSHA CHANCES cohorts EPIC-Elderly Tromsø ESTHER Combined N 22,692 5,482 9,771 4,211 3,228 17,210 Age (y) (mean (SD)) 71.3 (5.9) 76.5 (7.0) 68.7 (3.2) 72.8 (5.8) 68.8 (2.9) 69.7 (4.3)  65–74 (N, %) 17,514 (77.2) 2,337 (42.6) 9,240 (94.6) 2,759 (65.5) 3,178 (98.5) 15,177 (88.2)  75–84 4,165 (18.4) 2,305 (42.1) 530 (5.4) 1,280 (30.4) 50 (1.6) 1,860 (10.8)  85+ 1,013 (4.5) 840 (15.3) 1 (0.0) 172 (4.1) 0 (0.0) 173 (1.0) Sex (N, %)  Men 8,639 (38.1) 3,133 (57.2) 2,964 (30.3) 1,795 (42.6) 1,531 (47.4) 6,290 (36.6)  Women 14,053 (61.9) 2,349 (42.9) 6,807 (69.7) 2,416 (57.4) 1,697 (52.6) 10,920 (63.5) Education (N, %)  Less than high school 14,344 (63.2) 1,757 (32.1) 7,290 (74.6) 2,850 (67.8) 2,447 (75.8) 12,587 (73.1)  High school 5,942 (26.2) 2,292 (41.8) 1,977 (20.2) 1,039 (24.7) 634 (19.6) 3,650 (21.2)  More than high school 2,406 (10.6) 1,433 (26.1) 504 (5.2) 322 (7.7) 147 (4.6) 973 (5.7) Smoking status (N, %)  Ever smoker 10,661 (47.0) 2,792 (50.9) 3,807 (39.0) 2,604 (61.8) 1,458 (45.2) 7,869 (45.7)  Never smoker 12,031 (53.0) 2,690 (49.1) 5,964 (61.0) 1,607 (38.2) 1,770 (54.8) 9,341 (54.3) Average duration of mortality follow-up (y) 10.02 (±3.9) 7.43 (±3.0) 10.83 (±3.4) 11.09 (±5.2) 10.58 (±2.6) 10.84 (±3.8) Self-reported history of chronic conditions (N, %)  History of cancer 1,928 (8.5) 747 (13.6) 482 (4.9) 399 (9.5) 300 (9.3) 1,181 (6.9)  History of MI 1,730 (7.6) 625 (11.4) 420 (4.3) 449 (10.7) 236 (7.3) 1,105 (6.4)  History of stroke 1,062 (4.7) 384 (7.0) 278 (2.9) 239 (5.7) 161 (5.0) 678 (3.9)  History of diabetes 2,408 (10.6) 535 (9.8) 1,180 (12.1) 252 (6.0) 411 (13.7) 1,873 (10.9) Disease counts (N, %)  No disease 16,621 (73.3) 3,615 (65.9) 7,660 (78.4) 3,059 (72.6) 2,287 (70.9) 13,006 (75.6)  1 disease 5,116 (22.6) 1,491 (27.2) 1,881 (19.3) 979 (23.3) 765 (23.7) 3,625 (21.1)  2 diseases 858 (3.8) 332 (6.1) 211 (2.2) 160 (3.8) 155 (4.8) 526 (3.1)  3 diseases 92 (0.4) 40 (0.7) 19 (0.2) 12 (0.3) 21 (0.7) 52 (0.3)  4 diseases 5 (0.0) 4 (0.1) 0 (0.0) 1 (0.0) 0 (0.0) 1 (0.1) Note: CSHA = Canadian Study on Health and Aging; CHANCE = Consortium on Health and Ageing: Network of Cohorts in Europe and United States; ESTHER = Epidemiological Study on the Chances of Prevention, Early Recognition and Optimised Treatment of Chronic Diseases in the Older Population; EPIC = European Prospective Investigation Into Cancer and Nutrition; MI = myocardial infarction. View Large Table 2. Main Characteristics of the Study Participants by Cohort All studies combined CSHA CHANCES cohorts EPIC-Elderly Tromsø ESTHER Combined N 22,692 5,482 9,771 4,211 3,228 17,210 Age (y) (mean (SD)) 71.3 (5.9) 76.5 (7.0) 68.7 (3.2) 72.8 (5.8) 68.8 (2.9) 69.7 (4.3)  65–74 (N, %) 17,514 (77.2) 2,337 (42.6) 9,240 (94.6) 2,759 (65.5) 3,178 (98.5) 15,177 (88.2)  75–84 4,165 (18.4) 2,305 (42.1) 530 (5.4) 1,280 (30.4) 50 (1.6) 1,860 (10.8)  85+ 1,013 (4.5) 840 (15.3) 1 (0.0) 172 (4.1) 0 (0.0) 173 (1.0) Sex (N, %)  Men 8,639 (38.1) 3,133 (57.2) 2,964 (30.3) 1,795 (42.6) 1,531 (47.4) 6,290 (36.6)  Women 14,053 (61.9) 2,349 (42.9) 6,807 (69.7) 2,416 (57.4) 1,697 (52.6) 10,920 (63.5) Education (N, %)  Less than high school 14,344 (63.2) 1,757 (32.1) 7,290 (74.6) 2,850 (67.8) 2,447 (75.8) 12,587 (73.1)  High school 5,942 (26.2) 2,292 (41.8) 1,977 (20.2) 1,039 (24.7) 634 (19.6) 3,650 (21.2)  More than high school 2,406 (10.6) 1,433 (26.1) 504 (5.2) 322 (7.7) 147 (4.6) 973 (5.7) Smoking status (N, %)  Ever smoker 10,661 (47.0) 2,792 (50.9) 3,807 (39.0) 2,604 (61.8) 1,458 (45.2) 7,869 (45.7)  Never smoker 12,031 (53.0) 2,690 (49.1) 5,964 (61.0) 1,607 (38.2) 1,770 (54.8) 9,341 (54.3) Average duration of mortality follow-up (y) 10.02 (±3.9) 7.43 (±3.0) 10.83 (±3.4) 11.09 (±5.2) 10.58 (±2.6) 10.84 (±3.8) Self-reported history of chronic conditions (N, %)  History of cancer 1,928 (8.5) 747 (13.6) 482 (4.9) 399 (9.5) 300 (9.3) 1,181 (6.9)  History of MI 1,730 (7.6) 625 (11.4) 420 (4.3) 449 (10.7) 236 (7.3) 1,105 (6.4)  History of stroke 1,062 (4.7) 384 (7.0) 278 (2.9) 239 (5.7) 161 (5.0) 678 (3.9)  History of diabetes 2,408 (10.6) 535 (9.8) 1,180 (12.1) 252 (6.0) 411 (13.7) 1,873 (10.9) Disease counts (N, %)  No disease 16,621 (73.3) 3,615 (65.9) 7,660 (78.4) 3,059 (72.6) 2,287 (70.9) 13,006 (75.6)  1 disease 5,116 (22.6) 1,491 (27.2) 1,881 (19.3) 979 (23.3) 765 (23.7) 3,625 (21.1)  2 diseases 858 (3.8) 332 (6.1) 211 (2.2) 160 (3.8) 155 (4.8) 526 (3.1)  3 diseases 92 (0.4) 40 (0.7) 19 (0.2) 12 (0.3) 21 (0.7) 52 (0.3)  4 diseases 5 (0.0) 4 (0.1) 0 (0.0) 1 (0.0) 0 (0.0) 1 (0.1) All studies combined CSHA CHANCES cohorts EPIC-Elderly Tromsø ESTHER Combined N 22,692 5,482 9,771 4,211 3,228 17,210 Age (y) (mean (SD)) 71.3 (5.9) 76.5 (7.0) 68.7 (3.2) 72.8 (5.8) 68.8 (2.9) 69.7 (4.3)  65–74 (N, %) 17,514 (77.2) 2,337 (42.6) 9,240 (94.6) 2,759 (65.5) 3,178 (98.5) 15,177 (88.2)  75–84 4,165 (18.4) 2,305 (42.1) 530 (5.4) 1,280 (30.4) 50 (1.6) 1,860 (10.8)  85+ 1,013 (4.5) 840 (15.3) 1 (0.0) 172 (4.1) 0 (0.0) 173 (1.0) Sex (N, %)  Men 8,639 (38.1) 3,133 (57.2) 2,964 (30.3) 1,795 (42.6) 1,531 (47.4) 6,290 (36.6)  Women 14,053 (61.9) 2,349 (42.9) 6,807 (69.7) 2,416 (57.4) 1,697 (52.6) 10,920 (63.5) Education (N, %)  Less than high school 14,344 (63.2) 1,757 (32.1) 7,290 (74.6) 2,850 (67.8) 2,447 (75.8) 12,587 (73.1)  High school 5,942 (26.2) 2,292 (41.8) 1,977 (20.2) 1,039 (24.7) 634 (19.6) 3,650 (21.2)  More than high school 2,406 (10.6) 1,433 (26.1) 504 (5.2) 322 (7.7) 147 (4.6) 973 (5.7) Smoking status (N, %)  Ever smoker 10,661 (47.0) 2,792 (50.9) 3,807 (39.0) 2,604 (61.8) 1,458 (45.2) 7,869 (45.7)  Never smoker 12,031 (53.0) 2,690 (49.1) 5,964 (61.0) 1,607 (38.2) 1,770 (54.8) 9,341 (54.3) Average duration of mortality follow-up (y) 10.02 (±3.9) 7.43 (±3.0) 10.83 (±3.4) 11.09 (±5.2) 10.58 (±2.6) 10.84 (±3.8) Self-reported history of chronic conditions (N, %)  History of cancer 1,928 (8.5) 747 (13.6) 482 (4.9) 399 (9.5) 300 (9.3) 1,181 (6.9)  History of MI 1,730 (7.6) 625 (11.4) 420 (4.3) 449 (10.7) 236 (7.3) 1,105 (6.4)  History of stroke 1,062 (4.7) 384 (7.0) 278 (2.9) 239 (5.7) 161 (5.0) 678 (3.9)  History of diabetes 2,408 (10.6) 535 (9.8) 1,180 (12.1) 252 (6.0) 411 (13.7) 1,873 (10.9) Disease counts (N, %)  No disease 16,621 (73.3) 3,615 (65.9) 7,660 (78.4) 3,059 (72.6) 2,287 (70.9) 13,006 (75.6)  1 disease 5,116 (22.6) 1,491 (27.2) 1,881 (19.3) 979 (23.3) 765 (23.7) 3,625 (21.1)  2 diseases 858 (3.8) 332 (6.1) 211 (2.2) 160 (3.8) 155 (4.8) 526 (3.1)  3 diseases 92 (0.4) 40 (0.7) 19 (0.2) 12 (0.3) 21 (0.7) 52 (0.3)  4 diseases 5 (0.0) 4 (0.1) 0 (0.0) 1 (0.0) 0 (0.0) 1 (0.1) Note: CSHA = Canadian Study on Health and Aging; CHANCE = Consortium on Health and Ageing: Network of Cohorts in Europe and United States; ESTHER = Epidemiological Study on the Chances of Prevention, Early Recognition and Optimised Treatment of Chronic Diseases in the Older Population; EPIC = European Prospective Investigation Into Cancer and Nutrition; MI = myocardial infarction. View Large Table 3. Hazard Rate (HR) and Rate Advancement Period (RAP) Estimates and Corresponding 95% CI for Overall Mortality Associated With Specific Chronic Disease Combinations Disease Status at Baseline N N of deaths Age and Sex Age, sex, smoking, educationa HR (95% CI) RAP (95% CI) HR (95% CI) RAP (95% CI) Referenceb 16,621 5,480 1 (reference) (reference) 1 (reference) (reference) Cancer 1,496 728 1.38 (1.28–1.49) 2.97 (2.23–3.72) 1.37 (1.27–1.48) 2.83 (2.10–3.56) MI 1,169 688 1.57 (1.45–1.70) 4.14 (3.37–4.92) 1.53 (1.42–1.66) 3.84 (3.09–4.61) Stroke 621 366 1.59 (1.43–1.77) 4.24 (3.23–5.27) 1.56 (1.41–1.74) 4.02 (3.03–5.03) Diabetes 1,830 785 1.77 (1.64–1.91) 5.26 (4.53–6.00) 1.76 (1.64–1.90) 5.11 (4.39–5.83) Cancer and MI 141 99 2.10 (1.72–2.56) 6.79 (4.91–8.70) 2.11 (1.73–2.57) 6.70 (4.85–8.56) Cancer and stroke 78 61 2.42 (1.88–3.12) 8.12 (5.73–10.53) 2.34 (1.82–3.02) 7.66 (5.32–10.02) Cancer and diabetes 157 77 2.01 (1.60–2.52) 6.40 (4.28–8.54) 1.96 (1.56–2.45) 6.04 (3.97–8.13) MI and stroke 141 104 2.62 (2.15–3.18) 8.84 (6.99–10.71) 2.46 (2.02–2.99) 8.08 (6.27–9.91) MI and diabetes 195 146 3.29 (2.79–3.88) 10.93 (9.35–12.55) 3.18 (2.69–3.75) 10.39 (8.84–11.97) Stroke and diabetes 146 94 2.54 (2.07–3.11) 8.54 (6.61–10.50) 2.48 (2.02–3.05) 8.18 (6.29–10.09) Cancer and MI and stroke 17 13 3.13 (1.81–5.40) 10.48 (5.36–15.63) 3.40 (1.97–5.86) 11.00 (5.98–16.04) Cancer and MI and diabetes 21 18 4.25 (2.68–6.76) 13.30 (8.94–17.69) 4.22 (2.66–6.71) 12.95 (8.69–17.26) MI and stroke and diabetes 41 37 5.06 (3.66–6.99) 14.88 (11.82–17.99) 4.74 (3.43–6.56) 14.00 (11.00–17.04) Disease Status at Baseline N N of deaths Age and Sex Age, sex, smoking, educationa HR (95% CI) RAP (95% CI) HR (95% CI) RAP (95% CI) Referenceb 16,621 5,480 1 (reference) (reference) 1 (reference) (reference) Cancer 1,496 728 1.38 (1.28–1.49) 2.97 (2.23–3.72) 1.37 (1.27–1.48) 2.83 (2.10–3.56) MI 1,169 688 1.57 (1.45–1.70) 4.14 (3.37–4.92) 1.53 (1.42–1.66) 3.84 (3.09–4.61) Stroke 621 366 1.59 (1.43–1.77) 4.24 (3.23–5.27) 1.56 (1.41–1.74) 4.02 (3.03–5.03) Diabetes 1,830 785 1.77 (1.64–1.91) 5.26 (4.53–6.00) 1.76 (1.64–1.90) 5.11 (4.39–5.83) Cancer and MI 141 99 2.10 (1.72–2.56) 6.79 (4.91–8.70) 2.11 (1.73–2.57) 6.70 (4.85–8.56) Cancer and stroke 78 61 2.42 (1.88–3.12) 8.12 (5.73–10.53) 2.34 (1.82–3.02) 7.66 (5.32–10.02) Cancer and diabetes 157 77 2.01 (1.60–2.52) 6.40 (4.28–8.54) 1.96 (1.56–2.45) 6.04 (3.97–8.13) MI and stroke 141 104 2.62 (2.15–3.18) 8.84 (6.99–10.71) 2.46 (2.02–2.99) 8.08 (6.27–9.91) MI and diabetes 195 146 3.29 (2.79–3.88) 10.93 (9.35–12.55) 3.18 (2.69–3.75) 10.39 (8.84–11.97) Stroke and diabetes 146 94 2.54 (2.07–3.11) 8.54 (6.61–10.50) 2.48 (2.02–3.05) 8.18 (6.29–10.09) Cancer and MI and stroke 17 13 3.13 (1.81–5.40) 10.48 (5.36–15.63) 3.40 (1.97–5.86) 11.00 (5.98–16.04) Cancer and MI and diabetes 21 18 4.25 (2.68–6.76) 13.30 (8.94–17.69) 4.22 (2.66–6.71) 12.95 (8.69–17.26) MI and stroke and diabetes 41 37 5.06 (3.66–6.99) 14.88 (11.82–17.99) 4.74 (3.43–6.56) 14.00 (11.00–17.04) Notes: RAP estimates express the impact of a given exposure (i.e. combination of chronic conditions) on the risk of death, by determining the time (in years) by which the mortality rate is anticipated for study participants exposed compared to nonexposed, that is, individuals without any of the four investigated chronic conditions. All studies combined (EPIC-Elderly, Tromsø, ESTHER, and CSHA)c. CI = confidence interval; CHANCE = Consortium on Health and Ageing: Network of Cohorts in Europe and United States; CSHA = Canadian Study on Health and Aging; ESTHER = Epidemiological Study on the Chances of Prevention, Early Recognition and Optimised Treatment of Chronic Diseases in the Older Population; EPIC = European Prospective Investigation Into Cancer and Nutrition; MI = myocardial infarction. aAdjusted for sex, smoking status (ever/never), level of education (less than high school/high school/more than high school). bReference group had none of the four conditions under study. cStudy was included as a stratification factor in the analyses to account for baseline differences among the cohorts. View Large Table 3. Hazard Rate (HR) and Rate Advancement Period (RAP) Estimates and Corresponding 95% CI for Overall Mortality Associated With Specific Chronic Disease Combinations Disease Status at Baseline N N of deaths Age and Sex Age, sex, smoking, educationa HR (95% CI) RAP (95% CI) HR (95% CI) RAP (95% CI) Referenceb 16,621 5,480 1 (reference) (reference) 1 (reference) (reference) Cancer 1,496 728 1.38 (1.28–1.49) 2.97 (2.23–3.72) 1.37 (1.27–1.48) 2.83 (2.10–3.56) MI 1,169 688 1.57 (1.45–1.70) 4.14 (3.37–4.92) 1.53 (1.42–1.66) 3.84 (3.09–4.61) Stroke 621 366 1.59 (1.43–1.77) 4.24 (3.23–5.27) 1.56 (1.41–1.74) 4.02 (3.03–5.03) Diabetes 1,830 785 1.77 (1.64–1.91) 5.26 (4.53–6.00) 1.76 (1.64–1.90) 5.11 (4.39–5.83) Cancer and MI 141 99 2.10 (1.72–2.56) 6.79 (4.91–8.70) 2.11 (1.73–2.57) 6.70 (4.85–8.56) Cancer and stroke 78 61 2.42 (1.88–3.12) 8.12 (5.73–10.53) 2.34 (1.82–3.02) 7.66 (5.32–10.02) Cancer and diabetes 157 77 2.01 (1.60–2.52) 6.40 (4.28–8.54) 1.96 (1.56–2.45) 6.04 (3.97–8.13) MI and stroke 141 104 2.62 (2.15–3.18) 8.84 (6.99–10.71) 2.46 (2.02–2.99) 8.08 (6.27–9.91) MI and diabetes 195 146 3.29 (2.79–3.88) 10.93 (9.35–12.55) 3.18 (2.69–3.75) 10.39 (8.84–11.97) Stroke and diabetes 146 94 2.54 (2.07–3.11) 8.54 (6.61–10.50) 2.48 (2.02–3.05) 8.18 (6.29–10.09) Cancer and MI and stroke 17 13 3.13 (1.81–5.40) 10.48 (5.36–15.63) 3.40 (1.97–5.86) 11.00 (5.98–16.04) Cancer and MI and diabetes 21 18 4.25 (2.68–6.76) 13.30 (8.94–17.69) 4.22 (2.66–6.71) 12.95 (8.69–17.26) MI and stroke and diabetes 41 37 5.06 (3.66–6.99) 14.88 (11.82–17.99) 4.74 (3.43–6.56) 14.00 (11.00–17.04) Disease Status at Baseline N N of deaths Age and Sex Age, sex, smoking, educationa HR (95% CI) RAP (95% CI) HR (95% CI) RAP (95% CI) Referenceb 16,621 5,480 1 (reference) (reference) 1 (reference) (reference) Cancer 1,496 728 1.38 (1.28–1.49) 2.97 (2.23–3.72) 1.37 (1.27–1.48) 2.83 (2.10–3.56) MI 1,169 688 1.57 (1.45–1.70) 4.14 (3.37–4.92) 1.53 (1.42–1.66) 3.84 (3.09–4.61) Stroke 621 366 1.59 (1.43–1.77) 4.24 (3.23–5.27) 1.56 (1.41–1.74) 4.02 (3.03–5.03) Diabetes 1,830 785 1.77 (1.64–1.91) 5.26 (4.53–6.00) 1.76 (1.64–1.90) 5.11 (4.39–5.83) Cancer and MI 141 99 2.10 (1.72–2.56) 6.79 (4.91–8.70) 2.11 (1.73–2.57) 6.70 (4.85–8.56) Cancer and stroke 78 61 2.42 (1.88–3.12) 8.12 (5.73–10.53) 2.34 (1.82–3.02) 7.66 (5.32–10.02) Cancer and diabetes 157 77 2.01 (1.60–2.52) 6.40 (4.28–8.54) 1.96 (1.56–2.45) 6.04 (3.97–8.13) MI and stroke 141 104 2.62 (2.15–3.18) 8.84 (6.99–10.71) 2.46 (2.02–2.99) 8.08 (6.27–9.91) MI and diabetes 195 146 3.29 (2.79–3.88) 10.93 (9.35–12.55) 3.18 (2.69–3.75) 10.39 (8.84–11.97) Stroke and diabetes 146 94 2.54 (2.07–3.11) 8.54 (6.61–10.50) 2.48 (2.02–3.05) 8.18 (6.29–10.09) Cancer and MI and stroke 17 13 3.13 (1.81–5.40) 10.48 (5.36–15.63) 3.40 (1.97–5.86) 11.00 (5.98–16.04) Cancer and MI and diabetes 21 18 4.25 (2.68–6.76) 13.30 (8.94–17.69) 4.22 (2.66–6.71) 12.95 (8.69–17.26) MI and stroke and diabetes 41 37 5.06 (3.66–6.99) 14.88 (11.82–17.99) 4.74 (3.43–6.56) 14.00 (11.00–17.04) Notes: RAP estimates express the impact of a given exposure (i.e. combination of chronic conditions) on the risk of death, by determining the time (in years) by which the mortality rate is anticipated for study participants exposed compared to nonexposed, that is, individuals without any of the four investigated chronic conditions. All studies combined (EPIC-Elderly, Tromsø, ESTHER, and CSHA)c. CI = confidence interval; CHANCE = Consortium on Health and Ageing: Network of Cohorts in Europe and United States; CSHA = Canadian Study on Health and Aging; ESTHER = Epidemiological Study on the Chances of Prevention, Early Recognition and Optimised Treatment of Chronic Diseases in the Older Population; EPIC = European Prospective Investigation Into Cancer and Nutrition; MI = myocardial infarction. aAdjusted for sex, smoking status (ever/never), level of education (less than high school/high school/more than high school). bReference group had none of the four conditions under study. cStudy was included as a stratification factor in the analyses to account for baseline differences among the cohorts. View Large Diabetes was the most common condition to occur both in isolation and in combination with other chronic diseases (Table 3). Large differences in prevalence were observed among the specific disease combinations with the same number of conditions. All single diseases and disease combinations were associated with a statistically significant increased mortality rate compared to the reference category when analyzing all studies combined ranging from 2.97 (2.23, 3.72) for cancer alone to 14.88 (11.82, 17.99) for MI, stroke, and diabetes (Table 3). Separate analyses in the CSHA and CHANCES cohorts revealed the same ranking of conditions (Supplementary Table 1). Among the single-disease groups, diabetes was the condition associated with the largest age- and sex-adjusted HR (1.77; 95% CI, 1.64–1.91) and advanced the rate of dying by the most years (5.26 years; 95% CI, 4.53–6.00) compared to the reference group. A history of cancer alone had the smallest impact on mortality and advanced the rate of dying by 2.97 years (95% CI, 2.23–3.72). Among those identified with two conditions, there was a more than 4 year difference in the lowest to highest RAP (6.40 years for cancer and diabetes to 10.93 years for MI and diabetes; p < 0.0001). All HRs and RAPs attenuated slightly after additional adjustment for smoking status and level of education, but CIs did not cross one or zero, respectively (Table 3). Across all cohorts, effect modification by age was observed (p < 0.0001) with RAP estimates being larger in younger participants (<70 years) compared to older participants (≥70 years) (Supplementary Table 2). No interaction with sex, level of education, or smoking status was observed. Irrespective of disease status, men had the same mortality rate as women who are 5 years their junior (RAP 4.96, 95% CI, 4.53–5.39; results not shown). Similar HRs and RAPs were observed in analyses using a frailty model with random effect for study, suggesting that the hazard function was similar across the four cohorts (data not shown). Additional adjustment for BMI and hypertension in the CHANCES cohorts did not change the results notably (results not shown). Supplementary Table 3 shows the HR and RAP with and without adjustment for ADL disability, and adjustment for the combined presence of 11 other chronic conditions in the CSHA cohort (participants with missing data on disability and diseases were excluded from these analyses). Figure 1 rank orders the RAP of all single conditions and combinations of two and three conditions in order of magnitude (based on the confounder adjusted model) and depicts whether any two RAP estimates were statistically different. Pairwise comparison of two RAPs is shown in each cell and the statistical significance of each comparison is indicated by the shading; a vertical pattern fill meaning two RAPs are significantly different at p <0.05, a light gray shading indicating differences at p <0.01, and dark gray at p <0.001. For example, the RAP of having cancer alone (2.83 [2.10, 3.56] years) was lower than the RAP associated with any other single or combination of condition(s), and the RAP of diabetes alone (5.11 [4.39, 5.83]) was significantly higher than that of MI (4.02 [3.03, 5.03]). The RAPs of all combinations of two conditions were significantly higher than those for single conditions, with two exceptions. As well, the RAP for the combination of MI and diabetes (10.39 [8.84, 11.97]) was in the same order of magnitude as the RAP of most combinations of three diseases and significantly higher than the RAP of all combinations of two diseases that include cancer (ranging from 6.04 [3.97, 8.13] to 7.66 [5.32, 10.02]). Figure 1. View largeDownload slide Map comparing RAP estimates of single conditions, dyads, and tryads of conditions. Map was established using all studies combined. Pairwise comparison of two RAPs is shown in each cell. The following color scheme was applied: white for p >0.05, vertical pattern fill when p <0.05, shade of gray when p <0.01, and dark gray for p <0.001. All RAP estimates were derived from a Cox proportional Hazard model adjusted for age and sex. Participants who did not have any of four chronic conditions at baseline formed the reference group. Figure 1. View largeDownload slide Map comparing RAP estimates of single conditions, dyads, and tryads of conditions. Map was established using all studies combined. Pairwise comparison of two RAPs is shown in each cell. The following color scheme was applied: white for p >0.05, vertical pattern fill when p <0.05, shade of gray when p <0.01, and dark gray for p <0.001. All RAP estimates were derived from a Cox proportional Hazard model adjusted for age and sex. Participants who did not have any of four chronic conditions at baseline formed the reference group. Discussion Our prospective analyses of four cohort studies from Europe and Canada estimated the average time period by which mortality rates are accelerated among people aged 65 years and older with different combinations of cancer and cardiometabolic conditions (MI, stroke, and diabetes). Given the increasing prevalence and mortality burden associated with coexisting chronic conditions, our two main findings have important clinical and public health implications. First, we observed that certain disease (combinations) had a mortality burden that was similar or even higher compared to combinations with larger disease counts. This suggests that simple disease counts may not reflect the full underlying complexity of how multiple chronic conditions impact survival (results for disease counts are shown in Supplementary Table 4). Positioning these findings in the context of previous research is limited by the paucity of studies examining the impact of combinations of chronic conditions on mortality. A recent systematic review on the combined effect of cardiometabolic conditions on mortality identified only a handful of studies (9), which were not restricted to older populations. This review was part of a larger publication reporting only small differences in mortality rate among clusters of MI, stroke, and diabetes with the same number of conditions. In contrast to this study, we applied an aging lens and restricted our sample to adults 65 years and older. Multimorbidity prevalence and patterns have shown to change during the life course (22,23), and results from single-disease studies suggest that combinations of conditions likely impact mortality differently in middle-aged and senior adults (24–26). Two other reports previously presented data on unique combinations of conditions and survival in older populations (27,28), but these were not specifically focussed on cancer and cardiometabolic diseases and failed to specify the impact of disease combination on the progression of mortality. Second, in contrast to traditional measures of effect, the RAP expresses the time dimension by which mortality is impacted by disease status, capturing the ‘aging effect’ of having a multimorbidity. As a result, the RAP can help effectively communicate the impact of having chronic condition(s) on mortality, as well as the benefits of adopting preventive strategies to avoid the accumulation of conditions. Where burden of disease studies traditionally assign the highest priority ranking to the leading causes of death (i.e. cancer and cardiovascular disease in developed countries), we show that diabetes has the largest impact on accelerating people’s mortality rate. Unfortunately, an increase in misinterpretation of the RAP has been seen in literature over recent years where the RAP is incorrectly explained as the difference in mean survival time, or the time by which a survival curve is shifted between exposed and unexposed (29). In addition, although previous studies estimating RAP do not explore potential effect modification between age and exposures, we observed a strong interaction between baseline age and disease combinations on mortality rates, suggesting that the impact of multimorbidity on mortality RAPs decreases with advancing age. As such, our RAPs provide an average RAP, with even larger RAP for younger seniors with multiple conditions. As well, sex had a strong independent effect on the RAP above and beyond the presence of the four diseases under study; the RAP of male sex was comparable to that associated with having diabetes. When studying the effect of complex disease combinations, traditional models of interaction provide limited interpretation, which is why we opted to report the main effect of mutually exclusive disease combinations. We chose to focus our analyses on four key diseases that are among the most prevalent, costly, and preventable causes of death in high-income countries making them priority targets for public health actions. However, we could have potentially introduced bias into our estimates if certain excluded conditions more frequently coexist with one of the key conditions under study. Analyses of the CSHA data that included 11 additional chronic conditions did not change the results of our study. We used information about disease status at baseline and were not able to account for additional disease accumulation during follow-up in our analyses. We also had no information about the duration, temporality, and severity of the conditions. Although we found self-reported disease status to have moderate-to-good agreement with clinical assessment in the CSHA, and disease rates to be comparable to population prevalence across studies, misclassification of disease status may have occurred; likely resulting in an underestimation of disease occurrence leading to attenuated effect estimates. The mortality rate in our older population may be modified by the duration of disease, where conditions developed early in life may have a different effect on survival than those developed later. We restricted our analyses to seniors because these have the largest multimorbidity prevalence. Our RAP estimates may not be generalizable to younger populations as also indicated by the interaction observed with baseline age. CSHA participants with complete data were younger, higher educated, and more physically active compared those with incomplete data, suggesting that our complete case analysis may have resulted in a healthier sample. Although the CSHA cohort was on average higher educated than the CHANCES studies, similar severity ranking of conditions was observed across cohorts. We did not have information on the type of cancer reported at baseline and our sample is likely overrepresented by patients diagnosed with cancers with higher long-term survival rates. In addition, changes in the definition of MI, published in 2000, have led to a substantial increase in diagnosis. Although the studies included in the current analysis collected information on disease status before these changes took effect, research suggests that patients diagnosed with MI by this new definition have a higher long-term mortality (30), potentially resulting in an underestimation of the RAP associated with MI in our analyses. Our findings highlight that the care of older people with multiple chronic conditions requires a targeted approach that takes into account the presence of unique combinations of conditions. Understanding how these conditions accelerate the rate of dying alone and in combination, can help aid decision making. Supplementary Material Supplementary data is available at The Journals of Gerontology, Series A: Biological Sciences and Medical Sciences online. Funding The CHANCES project is coordinated by the Hellenic Health Foundation, Greece and received funding by the FP7 framework programme of DG-RESEARCH in the European Commission (grant agreement no. HEALTH-F3-2010–242244). The CSHA was funded by the Seniors Independence Research Programme, administered by the National Health and Development Programme of Health and Welfare Canada. Neither funder had any role in data collection, data analysis, data interpretation, or writing of the report. Lauren Griffith is supported by a Canadian Institutes of Health Research New Investigator’s Award, and the McLaughlin Foundation Professorship in Population and Public Health. Parminder Raina holds a Tier 1 Canada Research Chair in Geroscience and the Raymond and Margaret Labarge Chair in Research and Knowledge Application for Optimal Aging. Anne Gilsing is supported by a Canadian Institute of Health Research Post-doctoral Fellowship and a Michael G. DeGroote Fellowship Award from McMaster University. Conflict of Interest None reported. References 1. GBD 2013 Mortality and Causes of Death Collaborators . Global, regional, and national age-sex specific all-cause and cause-specific mortality for 240 causes of death, 1990–2013: a systematic analysis for the Global Burden of Disease Study 2013 . Lancet . 2015 ; 385 : 117 – 71 . doi: 10.1016/S0140-6736(14)61682-2 CrossRef Search ADS PubMed 2. United Nations Department of Economic and Social Affairs Population Division . World Population Ageing 2013 . New York, NY: United Nations Secretariat; 2013 . Report No.: ST/ESA/SER.A/348. 3. United Nations Department of Economic and Social Affairs Population Division . World Population Prospects: The 2015 Revision, Key Findings and Advance Tables . New York, NY: United Nations Secretariat; 2015 . Report No.: Working Paper No. ESA/P/WP.241. 4. World Healh Organization . Global Status Report on Noncommunicable Diseases 2014 . Geneva, Switzerland : World Health Organization ; 2014 . ISBN: 978 92 4 156485 4 5. Fortin M , Stewart M , Poitras ME , Almirall J , Maddocks H . A systematic review of prevalence studies on multimorbidity: toward a more uniform methodology . Ann Fam Med . 2012 ; 10 : 142 – 151 . doi: 10.1370/afm.1337 Google Scholar CrossRef Search ADS PubMed 6. Pefoyo AJ , Bronskill SE , Gruneir A et al. The increasing burden and complexity of multimorbidity . BMC Public Health . 2015 ; 15 : 415 . doi: 10.1186/s12889-015-1733-2 Google Scholar CrossRef Search ADS PubMed 7. Marengoni A , Angleman S , Melis R et al. Aging with multimorbidity: a systematic review of the literature . Ageing Res Rev . 2011 ; 10 : 430 – 439 . doi: 10.1016/j.arr.2011.03.003 Google Scholar CrossRef Search ADS PubMed 8. Fabbri E , Zoli M , Gonzalez-Freire M , Salive ME , Studenski SA , Ferrucci L . Aging and multimorbidity: new tasks, priorities, and frontiers for integrated gerontological and clinical research . J Am Med Dir Assoc . 2015 ; 16 : 640 – 647 . doi: 10.1016/j.jamda.2015.03.013 Google Scholar CrossRef Search ADS PubMed 9. Di Angelantonio E , Kaptoge S , Wormser D et al. Association of cardiometabolic multimorbidity with mortality . JAMA . 2015 ; 314 : 52 – 60 . doi: 10.1001/jama.2015.7008 Google Scholar CrossRef Search ADS PubMed 10. Boffetta P , Bobak M , Borsch-Supan A et al. The consortium on health and ageing: network of Cohorts in Europe and the United States (CHANCES) project–design, population and data harmonization of a large-scale, international study . Eur J Epidemiol . 2014 ; 29 : 929 – 936 . doi: 10.1007/s10654-014-9977-1 Google Scholar CrossRef Search ADS PubMed 11. Kuulasmaa K , Palosaari T . Contributors from Partners of the Consortium on Health and Ageing: Network of Cohorts in Europe and the United States (CHANCES). CHANCES cohort descriptions, assessment of the availability and quality of data, and definitions of variables. MORGAM Project e-publications [Internet] . 2015 ; 6 URN: NBN: fi-fe201501151161. http://www.thl.fi/publications/morgam/chances_d9/index.html. Accessed October, 2017. 12. Riboli E , Kaaks R . The EPIC project: rationale and study design. European prospective investigation into cancer and nutrition . Int J Epidemiol 1997 ; 26 ( suppl 1 ): S6 – 14 . doi: 10.1079/PHN2002394 Google Scholar CrossRef Search ADS PubMed 13. Raum E , Rothenbacher D , Löw M , Stegmaier C , Ziegler H , Brenner H . Changes of cardiovascular risk factors and their implications in subsequent birth cohorts of older adults in Germany: a life course approach . Eur J Cardiovasc Prev Rehabil . 2007 ; 14 : 809 – 814 . doi: 10.1097/HJR.0b013e3282eeb308 Google Scholar CrossRef Search ADS PubMed 14. Jacobsen BK , Eggen AE , Mathiesen EB , Wilsgaard T , Njølstad I . Cohort profile: the Tromso study . Int J Epidemiol . 2012 ; 41 : 961 – 967 . doi: 10.1093/ije/dyr049 Google Scholar CrossRef Search ADS PubMed 15. McDowell I , Hill G , Lindsay J . An overview of the Canadian Study of Health and Aging . Int Psychogeriatr 2001 ; 13 ( supp 1 ): 7 – 18 . doi: 10.1017/S1041610202007949 Google Scholar CrossRef Search ADS PubMed 16. Engstad T , Bonaa KH , Viitanen M . Validity of self-reported stroke: the Tromso Study . Stroke . 2000 ; 31 : 1602 – 1607 . doi:10.1161/01.STR.31.7.1602 Google Scholar CrossRef Search ADS PubMed 17. Okura Y , Urban LH , Mahoney DW , Jacobsen SJ , Rodeheffer RJ . Agreement between self-report questionnaires and medical record data was substantial for diabetes, hypertension, myocardial infarction and stroke but not for heart failure . J Clin Epidemiol . 2004 ; 57 : 1096 – 1103 . doi: 10.1016/ j.jclinepi.2004.04.005 Google Scholar CrossRef Search ADS PubMed 18. van den Akker M , van Steenkiste B , Krutwagen E , Metsemakers JF . Disease or no disease? Disagreement on diagnoses between self-reports and medical records of adult patients . Eur J Gen Pract . 2015 ; 21 : 45 – 51 . doi: 10.3109/13814788.2014.907266 Google Scholar CrossRef Search ADS PubMed 19. The Canadian Study on Health and Aging Working Group . Disability and frailty among elderly Canadians: a comparison of six surveys . Int Psychogeriatr 2001 ; 13 ( supp 1 ): 159 – 67 . doi: 10.1017/S1041610202008104 CrossRef Search ADS PubMed 20. Dubois MF , Hebert R . Imputation of missing dates of death or institutionalization for time-to-event analyses in the Canadian Study of Health and Aging . Int Psychogeriatr 2001 ; 13 ( supp 1 ): 91 – 7 . doi: 10.1017/S1041610202008025 Google Scholar CrossRef Search ADS PubMed 21. Brenner H , Gefeller O , Greenland S . Risk and rate advancement periods as measures of exposure impact on the occurrence of chronic diseases . Epidemiology . 1993 ; 4 : 229 – 236 . PMID: 8512987. Google Scholar CrossRef Search ADS PubMed 22. Violan C , Foguet-Boreu Q , Flores-Mateo G et al. Prevalence, determinants and patterns of multimorbidity in primary care: a systematic review of observational studies . PLoS One . 2014 ; 9 : e102149 . doi: 10.1371/journal.pone.0102149 Google Scholar CrossRef Search ADS PubMed 23. Prados-Torres A , Calderón-Larrañaga A , Hancco-Saavedra J , Poblador-Plou B , van den Akker M . Multimorbidity patterns: a systematic review . J Clin Epidemiol . 2014 ; 67 : 254 – 266 . doi: 10.1016/j.jclinepi.2013.09.021 Google Scholar CrossRef Search ADS PubMed 24. Brønnum-Hansen H , Davidsen M , Thorvaldsen P ; Danish MONICA Study Group . Long-term survival and causes of death after stroke . Stroke . 2001 ; 32 : 2131 – 2136 . doi:10.1161/hs0901.094253 Google Scholar CrossRef Search ADS PubMed 25. Smolina K , Wright FL , Rayner M , Goldacre MJ . Long-term survival and recurrence after acute myocardial infarction in England, 2004 to 2010 . Circ Cardiovasc Qual Outcomes 2012 ; 5 : 532 – 540 . doi: 10.1161/CIRCOUTCOMES.111.964700 Google Scholar CrossRef Search ADS PubMed 26. De Angelis AR , Sant M , Coleman MP et al. Cancer survival in Europe 1999–2007 by country and age: results of EUROCARE--5-a population-based study . Lancet Oncol 2014 ; 15 : 23 – 34 . doi: 10.1016/S1470-2045(13)70546-1 Google Scholar CrossRef Search ADS PubMed 27. Caughey GE , Ramsay EN , Vitry AI et al. Comorbid chronic diseases, discordant impact on mortality in older people: a 14-year longitudinal population study . J Epidemiol Community Health . 2010 ; 64 : 1036 – 1042 . doi: 10.1136/jech.2009.088260 Google Scholar CrossRef Search ADS PubMed 28. Lee TA , Shields AE , Vogeli C et al. Mortality rate in veterans with multiple chronic conditions . J Gen Intern Med 2007 ; 22 ( suppl 3 ): 403 – 7 . doi: 10.1007/s11606-007-0277-2 Google Scholar CrossRef Search ADS PubMed 29. Discacciati A , Bellavia A , Orsini N , Greenland S . On the interpretation of risk and rate advancement periods . Int J Epidemiol . 2016 ; 45 : 278 – 284 . doi: 10.1093/ije/dyv320 Google Scholar CrossRef Search ADS PubMed 30. Hochholzer W , Buettner HJ , Trenk D et al. New definition of myocardial infarction: impact on long-term mortality . Am J Med . 2008 ; 121 : 399 – 405 . doi: 10.1016/j.amjmed.2008.01.033 Google Scholar CrossRef Search ADS PubMed © The Author(s) 2018. Published by Oxford University Press on behalf of The Gerontological Society of America. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png The Journals of Gerontology Series A: Biomedical Sciences and Medical Sciences Oxford University Press

Loading next page...
 
/lp/ou_press/the-combined-effect-of-cancer-and-cardiometabolic-conditions-on-the-aLjtfMSk1f
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of The Gerontological Society of America. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.
ISSN
1079-5006
eISSN
1758-535X
D.O.I.
10.1093/gerona/gly053
Publisher site
See Article on Publisher Site

Abstract

Abstract Background The number of older people living with cancer and cardiometabolic conditions is increasing, but little is known about how specific combinations of these conditions impact mortality. Methods A total of 22,692 participants aged 65 years and older from four international cohorts were followed-up for mortality for an average of 10 years (8,596 deaths). Data were harmonized across cohorts and mutually exclusive groups of disease combinations were created for cancer, myocardial infarction (MI), stroke, and diabetes at baseline. Cox proportional hazards models for all-cause mortality were used to estimate the age- and sex-adjusted hazard ratio and rate advancement period (RAP) (in years). Results At baseline, 23.6% (n = 5,116) of participants reported having one condition and 4.2% (n = 955) had two or more conditions. Data from all studies combined showed that the RAP increased with each additional condition. Diabetes advanced the rate of dying by the most years (5.26 years; 95% confidence interval [CI], 4.53–6.00), but the effect of any single condition was smaller than the effect of disease combinations. Some combinations had a significantly greater impact on the period by which the rate of death was advanced than others with the same number of conditions, for example, 10.9 years (95% CI, 9.4–12.6) for MI and diabetes versus 6.4 years (95% CI, 4.3–8.5) for cancer and diabetes. Conclusions Combinations of cancer and cardiometabolic conditions accelerate mortality rates in older adults differently. Although most studies investigating mortality associated with multimorbidity used disease counts, these provide little guidance for managing complex patients as they age. Multimorbidity, Epidemiology, Risk factor, Aging The global population of people aged 60 years and older is increasing in almost all regions of the world and is projected to double in size over the next 35 years, reaching nearly 2.1 billion by 2050 (1–3). This transition is accompanied by an increase in prevalence of many chronic diseases (1,4), with most older adults having two or more conditions coexisting at the same time (5,6). Of these, cancer, ischemic heart disease, stroke, and diabetes were identified as the main drivers of mortality in developed countries by the Global Burden of Disease Study 2013, accounting for almost 25 million deaths worldwide (1). The accumulation of chronic conditions serves as an important indicator for the progressive loss of resilience and functional independence (7) and has been suggested as an early marker of accelerated biological aging (8). Furthermore, having two or more chronic conditions has been shown to be an independent predictor of mortality (7) and is associated with marked reductions in life expectancy (9). Researchers, however, seldom examine the impact of specific combinations of chronic conditions on survival, even though results from single-disease studies suggest that specific disease combinations are likely to have a much stronger association with mortality than others. Therefore, better understanding the consequences of specific combinations of chronic conditions on mortality is a critical first step toward optimizing the public health and health care needs of older adults. The purpose of this study is to determine how specific combinations of the main drivers of mortality, including cancer, myocardial infarction (MI), stroke, and diabetes, accelerate mortality rates in older adults. Methods Study Design and Participants Our study uses longitudinal individual-level harmonized data from four cohort studies, three from the Consortium on Health and Ageing: Network of Cohorts in Europe and United States (CHANCES) (10), and one from Canada, the Canadian Study on Health and Aging (CSHA). The cohorts’ key design characteristics are summarized in Table 1. Table 1. Characteristics of the Four Cohort Studies for the Analyses on Cancer and Cardiometabolic Multimorbidity and All-Cause Mortality Studya Countryb Period of enrollment Mortality follow-up Participants 65+ y, N Complete mortality FU, N Complete chronic disease data, N Complete confounder data, Nc EPIC-Elderly (13) (selected centers) DK, GR, NL, ES, SE 1992–2000 1992–2011 10,309 10,079 9,956 9,771 ESTHER (14) DE 2000–2002 2000–2013 3,845 3,842 3,421 3,228 Tromsø Study (15) NO 1994–1995 1994–2010 4,286 4,286 4,251 4,211 CSHA (16) CA 1991 1991–2001 9,008 8,743 5,627d 5,482 Studya Countryb Period of enrollment Mortality follow-up Participants 65+ y, N Complete mortality FU, N Complete chronic disease data, N Complete confounder data, Nc EPIC-Elderly (13) (selected centers) DK, GR, NL, ES, SE 1992–2000 1992–2011 10,309 10,079 9,956 9,771 ESTHER (14) DE 2000–2002 2000–2013 3,845 3,842 3,421 3,228 Tromsø Study (15) NO 1994–1995 1994–2010 4,286 4,286 4,251 4,211 CSHA (16) CA 1991 1991–2001 9,008 8,743 5,627d 5,482 Note: aEPIC: European Prospective Investigation Into Cancer and Nutrition; ESTHER: Epidemiological Study on the Chances of Prevention, Early Recognition and Optimised Treatment of Chronic Diseases in the Older Population; CSHA: Canadian Study on Health and Aging. bCanada (CA); Denmark (DK); Germany (DE); Greece (GR); Netherlands (NL); Norway (NO); Spain (ES); Sweden (SE). cData on smoking (ever smoker, never smoker) and level of education (less than high school, high school, more than high school). dOne thousand four hundred and thirty-four CSHA participants did not complete the risk factor questionnaire and another 1,682 participants did not have complete exposure data (i.e. history of the four chronic conditions under study). View Large Table 1. Characteristics of the Four Cohort Studies for the Analyses on Cancer and Cardiometabolic Multimorbidity and All-Cause Mortality Studya Countryb Period of enrollment Mortality follow-up Participants 65+ y, N Complete mortality FU, N Complete chronic disease data, N Complete confounder data, Nc EPIC-Elderly (13) (selected centers) DK, GR, NL, ES, SE 1992–2000 1992–2011 10,309 10,079 9,956 9,771 ESTHER (14) DE 2000–2002 2000–2013 3,845 3,842 3,421 3,228 Tromsø Study (15) NO 1994–1995 1994–2010 4,286 4,286 4,251 4,211 CSHA (16) CA 1991 1991–2001 9,008 8,743 5,627d 5,482 Studya Countryb Period of enrollment Mortality follow-up Participants 65+ y, N Complete mortality FU, N Complete chronic disease data, N Complete confounder data, Nc EPIC-Elderly (13) (selected centers) DK, GR, NL, ES, SE 1992–2000 1992–2011 10,309 10,079 9,956 9,771 ESTHER (14) DE 2000–2002 2000–2013 3,845 3,842 3,421 3,228 Tromsø Study (15) NO 1994–1995 1994–2010 4,286 4,286 4,251 4,211 CSHA (16) CA 1991 1991–2001 9,008 8,743 5,627d 5,482 Note: aEPIC: European Prospective Investigation Into Cancer and Nutrition; ESTHER: Epidemiological Study on the Chances of Prevention, Early Recognition and Optimised Treatment of Chronic Diseases in the Older Population; CSHA: Canadian Study on Health and Aging. bCanada (CA); Denmark (DK); Germany (DE); Greece (GR); Netherlands (NL); Norway (NO); Spain (ES); Sweden (SE). cData on smoking (ever smoker, never smoker) and level of education (less than high school, high school, more than high school). dOne thousand four hundred and thirty-four CSHA participants did not complete the risk factor questionnaire and another 1,682 participants did not have complete exposure data (i.e. history of the four chronic conditions under study). View Large The CHANCES Consortium The CHANCES consortium is a large collaborative project which harmonized data from on-going prospective cohort studies in Europe and the United States to study aging-related health characteristics and determinants of healthy aging (10). Variables were harmonized using predetermined standardized procedures (11). Three cohorts were included in the present analyses: the European Prospective Investigation Into Cancer and Nutrition–Elderly (EPIC-Elderly) Study (12) from Spain, the Netherlands, Greece, Sweden, and Denmark; the Epidemiological Study on the Chances of Prevention, Early Recognition and Optimised Treatment of Chronic Diseases in the Older Population (ESTHER) Study (13) from Germany; and the Tromsø Study (14) from Norway. The current analysis includes all individuals 65 years and older who completed a self-administered questionnaire on lifestyle characteristics, chronic disease status, and potential chronic disease risk factors at baseline (Table 1) (n = 17,210; 9,771 [EPIC-Elderly], 3,228 [ESTHER], and 4,211 [Tromsø]). The CSHA The CSHA is a national, population-based study of cognitive impairment and other aspects of health in Canadian adults aged 65 years and older (15). The CSHA consists of 10,263 participants in community (n = 9,008) and institutional settings (n = 1,255) who were recruited from the 10 Canadian provinces. In the first wave of the CSHA in 1991, face-to-face interviews were conducted with all community participants to screen for dementia. Participants who were cognitively normal were asked to complete a self-administered risk factor questionnaire on demographic characteristics, lifestyle, and chronic disease status. The risk factor questionnaire was completed by a proxy for participants found to be cognitively impaired. CSHA variables were harmonized following CHANCES procedures. The current analysis includes community-dwelling individuals who completed a risk factor questionnaire (n = 5,482; 4,751 self-administered and 731 proxy-administered) and had complete data. In all cohorts, the research procedures were in accordance with the ethical standards of the responsible institutional or regional committees on human experimentation, and all participants provided informed consent. Combinations of Diseases All cohorts collected self-reported information on history of cancer, MI, stroke, and diabetes as part of the baseline questionnaires. Although clinical diagnosis by a physician is considered the standard criterion for ascertaining disease endpoints, large-scale epidemiological studies often rely on self-reported diagnoses. We were able to address the accuracy of our data by comparing the agreement between self-reported disease status to physicians diagnosis in a subset of the CSHA population (n = 1,068) who underwent an additional clinical examination at baseline. Moderate-to-good agreement was observed with κ’s ranging from 0.35 (heart disease) to 0.87 (diabetes). These findings are in line with previous research showing that self-reported diagnosis of the conditions under study compare relatively well to clinical diagnoses (16–18). As well, to assess the generalizability of the sample, disease prevalence estimates from the included studies were compared against representative national survey data (also using self-reports to define morbidity). Prevalence rates were in the same order of magnitude in CSHA and the 1991 General Social Survey (19), and the CHANCES cohorts and first European Health Interview Survey (2006–2008). Participants were classified into mutually exclusive groups of chronic disease combinations at baseline (i.e. single, dyads, and triads of conditions). The four conditions under study resulted in 16 unique disease combinations, including the reference category with none of these four conditions. The independent effect of conditions with less than 15 participants was not reported due to low power. Mortality Follow-Up EPIC-Elderly, ESTHER, and Tromsø were followed up for vital status by linkage to region, state, or country-wide death registries with coverage of ≥98.5%. The median follow-up period ranged from 12 years (ESTHER and EPIC), to 16 years (Tromsø). CSHA participants were followed up for vital status after 10 years, and date of death was recorded for individuals who were deceased. Missing dates of death for CSHA decedents were derived as previously described (20). Less than 3% of CSHA participants were lost to follow-up. Statistical Analysis Statistical analyses were performed separately for each study cohort (treating EPIC-Elderly as one cohort), for all CHANCES cohorts combined, and CHANCES and CSHA cohorts combined by pooling individual-level data. The aging effect of chronic disease combinations on mortality was estimated using rate advancement periods (RAP). The derivation of the RAP has been described in detail by Brenner and colleagues (21). The RAP was derived from a multivariable Cox proportional hazards models of the form: hazard (t, exposure) = b1 × disease combination + b2 × age + covariates. From this model, a point estimate of the RAP is obtained as b1/b2 (i.e. the ratio of the regression coefficients for the specific disease combination [numerator] and age at baseline [denominator]). The RAP specifies the period by which the rate of death is advanced among subjects with a specific chronic disease combination relative to the reference group in the absence of competing risks. In other words, it specifies the loss of years in terms of mortality risk associated with having a specific combination of chronic conditions. The RAP thus quantifies the impact of a specific combination of conditions on mortality relative to the effect of aging. For example, a RAP of 10 years for a specific exposure (E) means that means that subjects with that exposure (E+) have the same risk of dying as someone without that exposure (E−) who is 10 years their senior. Confidence intervals (CIs) for the RAPs were calculated using Fieller’s theorem. Sensitivity analyses were performed to test the robustness of results using a frailty model that included study as a random effect. Our primary analyses calculated hazard ratios (HRs) adjusted for age and sex only. We also explored the extent to which other demographic and lifestyle factors explained the associations between disease combinations and mortality. Variables that changed any RAP estimate by 10% or more were included as confounders in the multivariable analyses resulting in a final model adjusted for tobacco smoking (never, ever) and education (<high school, high school, >high school). CHANCES cohorts included information on anthropometry and blood pressure allowing us to investigate the effect of body mass index (<18.5, 18.5–24.9, 25.0–29.9, ≥30 kg/m2), and hypertension (diastolic ≥90 mmHg and/or systolic ≤140 mmHg) on the RAP estimates, whereas CSHA enabled us to examine the explanatory effect of disability status at baseline (combined presence or absence of self-reported limitations in seven activities of daily living). Possible effect modification by age (stratified by median age of 70 years), sex, and the previously mentioned covariates were tested using interaction terms in the Cox regression model. Because the denominator of the RAP estimates (i.e. age at baseline) is similar for each RAP derived from the same model, the difference in RAP estimates result from a difference in the regression estimate of the disease combinations. We evaluated if any two RAPs differed from each other by calculating Z-scores comparing the parameter estimates of unique combinations of conditions. Results from all pairwise comparisons are depicted using a heatmap-method. All analyses were performed using SAS V9.4. Results In total, 22,692 participants aged 65 years and older were included in this study (Table 2), with 8,708 (38%) deaths recorded during follow-up (Table 3). The average follow-up duration was between 7 and 11 years. Participants from EPIC-Elderly and ESTHER were younger than those from the Tromsø and CSHA cohorts. All studies included more women than men (overall 62% women), and 47% of participants were ever smokers. The highest level of education varied across studies; 26% of CSHA participants had more than a high school degree compared to 5.7% in the CHANCES studies. At enrolment, 1,928 (8.5%) participants reported having cancer, 1,730 (7.6%) MI, 1,062 (4.7%) stroke, and 2,408 (10.6%) diabetes, respectively. Disease rates were generally higher in the older cohorts (CSHA, Tromsø) compared to younger cohorts (EPIC-Elderly and ESTHER). The majority of participants had none (73.3%) or only one (22.6%) of the four chronic conditions under study. Table 2. Main Characteristics of the Study Participants by Cohort All studies combined CSHA CHANCES cohorts EPIC-Elderly Tromsø ESTHER Combined N 22,692 5,482 9,771 4,211 3,228 17,210 Age (y) (mean (SD)) 71.3 (5.9) 76.5 (7.0) 68.7 (3.2) 72.8 (5.8) 68.8 (2.9) 69.7 (4.3)  65–74 (N, %) 17,514 (77.2) 2,337 (42.6) 9,240 (94.6) 2,759 (65.5) 3,178 (98.5) 15,177 (88.2)  75–84 4,165 (18.4) 2,305 (42.1) 530 (5.4) 1,280 (30.4) 50 (1.6) 1,860 (10.8)  85+ 1,013 (4.5) 840 (15.3) 1 (0.0) 172 (4.1) 0 (0.0) 173 (1.0) Sex (N, %)  Men 8,639 (38.1) 3,133 (57.2) 2,964 (30.3) 1,795 (42.6) 1,531 (47.4) 6,290 (36.6)  Women 14,053 (61.9) 2,349 (42.9) 6,807 (69.7) 2,416 (57.4) 1,697 (52.6) 10,920 (63.5) Education (N, %)  Less than high school 14,344 (63.2) 1,757 (32.1) 7,290 (74.6) 2,850 (67.8) 2,447 (75.8) 12,587 (73.1)  High school 5,942 (26.2) 2,292 (41.8) 1,977 (20.2) 1,039 (24.7) 634 (19.6) 3,650 (21.2)  More than high school 2,406 (10.6) 1,433 (26.1) 504 (5.2) 322 (7.7) 147 (4.6) 973 (5.7) Smoking status (N, %)  Ever smoker 10,661 (47.0) 2,792 (50.9) 3,807 (39.0) 2,604 (61.8) 1,458 (45.2) 7,869 (45.7)  Never smoker 12,031 (53.0) 2,690 (49.1) 5,964 (61.0) 1,607 (38.2) 1,770 (54.8) 9,341 (54.3) Average duration of mortality follow-up (y) 10.02 (±3.9) 7.43 (±3.0) 10.83 (±3.4) 11.09 (±5.2) 10.58 (±2.6) 10.84 (±3.8) Self-reported history of chronic conditions (N, %)  History of cancer 1,928 (8.5) 747 (13.6) 482 (4.9) 399 (9.5) 300 (9.3) 1,181 (6.9)  History of MI 1,730 (7.6) 625 (11.4) 420 (4.3) 449 (10.7) 236 (7.3) 1,105 (6.4)  History of stroke 1,062 (4.7) 384 (7.0) 278 (2.9) 239 (5.7) 161 (5.0) 678 (3.9)  History of diabetes 2,408 (10.6) 535 (9.8) 1,180 (12.1) 252 (6.0) 411 (13.7) 1,873 (10.9) Disease counts (N, %)  No disease 16,621 (73.3) 3,615 (65.9) 7,660 (78.4) 3,059 (72.6) 2,287 (70.9) 13,006 (75.6)  1 disease 5,116 (22.6) 1,491 (27.2) 1,881 (19.3) 979 (23.3) 765 (23.7) 3,625 (21.1)  2 diseases 858 (3.8) 332 (6.1) 211 (2.2) 160 (3.8) 155 (4.8) 526 (3.1)  3 diseases 92 (0.4) 40 (0.7) 19 (0.2) 12 (0.3) 21 (0.7) 52 (0.3)  4 diseases 5 (0.0) 4 (0.1) 0 (0.0) 1 (0.0) 0 (0.0) 1 (0.1) All studies combined CSHA CHANCES cohorts EPIC-Elderly Tromsø ESTHER Combined N 22,692 5,482 9,771 4,211 3,228 17,210 Age (y) (mean (SD)) 71.3 (5.9) 76.5 (7.0) 68.7 (3.2) 72.8 (5.8) 68.8 (2.9) 69.7 (4.3)  65–74 (N, %) 17,514 (77.2) 2,337 (42.6) 9,240 (94.6) 2,759 (65.5) 3,178 (98.5) 15,177 (88.2)  75–84 4,165 (18.4) 2,305 (42.1) 530 (5.4) 1,280 (30.4) 50 (1.6) 1,860 (10.8)  85+ 1,013 (4.5) 840 (15.3) 1 (0.0) 172 (4.1) 0 (0.0) 173 (1.0) Sex (N, %)  Men 8,639 (38.1) 3,133 (57.2) 2,964 (30.3) 1,795 (42.6) 1,531 (47.4) 6,290 (36.6)  Women 14,053 (61.9) 2,349 (42.9) 6,807 (69.7) 2,416 (57.4) 1,697 (52.6) 10,920 (63.5) Education (N, %)  Less than high school 14,344 (63.2) 1,757 (32.1) 7,290 (74.6) 2,850 (67.8) 2,447 (75.8) 12,587 (73.1)  High school 5,942 (26.2) 2,292 (41.8) 1,977 (20.2) 1,039 (24.7) 634 (19.6) 3,650 (21.2)  More than high school 2,406 (10.6) 1,433 (26.1) 504 (5.2) 322 (7.7) 147 (4.6) 973 (5.7) Smoking status (N, %)  Ever smoker 10,661 (47.0) 2,792 (50.9) 3,807 (39.0) 2,604 (61.8) 1,458 (45.2) 7,869 (45.7)  Never smoker 12,031 (53.0) 2,690 (49.1) 5,964 (61.0) 1,607 (38.2) 1,770 (54.8) 9,341 (54.3) Average duration of mortality follow-up (y) 10.02 (±3.9) 7.43 (±3.0) 10.83 (±3.4) 11.09 (±5.2) 10.58 (±2.6) 10.84 (±3.8) Self-reported history of chronic conditions (N, %)  History of cancer 1,928 (8.5) 747 (13.6) 482 (4.9) 399 (9.5) 300 (9.3) 1,181 (6.9)  History of MI 1,730 (7.6) 625 (11.4) 420 (4.3) 449 (10.7) 236 (7.3) 1,105 (6.4)  History of stroke 1,062 (4.7) 384 (7.0) 278 (2.9) 239 (5.7) 161 (5.0) 678 (3.9)  History of diabetes 2,408 (10.6) 535 (9.8) 1,180 (12.1) 252 (6.0) 411 (13.7) 1,873 (10.9) Disease counts (N, %)  No disease 16,621 (73.3) 3,615 (65.9) 7,660 (78.4) 3,059 (72.6) 2,287 (70.9) 13,006 (75.6)  1 disease 5,116 (22.6) 1,491 (27.2) 1,881 (19.3) 979 (23.3) 765 (23.7) 3,625 (21.1)  2 diseases 858 (3.8) 332 (6.1) 211 (2.2) 160 (3.8) 155 (4.8) 526 (3.1)  3 diseases 92 (0.4) 40 (0.7) 19 (0.2) 12 (0.3) 21 (0.7) 52 (0.3)  4 diseases 5 (0.0) 4 (0.1) 0 (0.0) 1 (0.0) 0 (0.0) 1 (0.1) Note: CSHA = Canadian Study on Health and Aging; CHANCE = Consortium on Health and Ageing: Network of Cohorts in Europe and United States; ESTHER = Epidemiological Study on the Chances of Prevention, Early Recognition and Optimised Treatment of Chronic Diseases in the Older Population; EPIC = European Prospective Investigation Into Cancer and Nutrition; MI = myocardial infarction. View Large Table 2. Main Characteristics of the Study Participants by Cohort All studies combined CSHA CHANCES cohorts EPIC-Elderly Tromsø ESTHER Combined N 22,692 5,482 9,771 4,211 3,228 17,210 Age (y) (mean (SD)) 71.3 (5.9) 76.5 (7.0) 68.7 (3.2) 72.8 (5.8) 68.8 (2.9) 69.7 (4.3)  65–74 (N, %) 17,514 (77.2) 2,337 (42.6) 9,240 (94.6) 2,759 (65.5) 3,178 (98.5) 15,177 (88.2)  75–84 4,165 (18.4) 2,305 (42.1) 530 (5.4) 1,280 (30.4) 50 (1.6) 1,860 (10.8)  85+ 1,013 (4.5) 840 (15.3) 1 (0.0) 172 (4.1) 0 (0.0) 173 (1.0) Sex (N, %)  Men 8,639 (38.1) 3,133 (57.2) 2,964 (30.3) 1,795 (42.6) 1,531 (47.4) 6,290 (36.6)  Women 14,053 (61.9) 2,349 (42.9) 6,807 (69.7) 2,416 (57.4) 1,697 (52.6) 10,920 (63.5) Education (N, %)  Less than high school 14,344 (63.2) 1,757 (32.1) 7,290 (74.6) 2,850 (67.8) 2,447 (75.8) 12,587 (73.1)  High school 5,942 (26.2) 2,292 (41.8) 1,977 (20.2) 1,039 (24.7) 634 (19.6) 3,650 (21.2)  More than high school 2,406 (10.6) 1,433 (26.1) 504 (5.2) 322 (7.7) 147 (4.6) 973 (5.7) Smoking status (N, %)  Ever smoker 10,661 (47.0) 2,792 (50.9) 3,807 (39.0) 2,604 (61.8) 1,458 (45.2) 7,869 (45.7)  Never smoker 12,031 (53.0) 2,690 (49.1) 5,964 (61.0) 1,607 (38.2) 1,770 (54.8) 9,341 (54.3) Average duration of mortality follow-up (y) 10.02 (±3.9) 7.43 (±3.0) 10.83 (±3.4) 11.09 (±5.2) 10.58 (±2.6) 10.84 (±3.8) Self-reported history of chronic conditions (N, %)  History of cancer 1,928 (8.5) 747 (13.6) 482 (4.9) 399 (9.5) 300 (9.3) 1,181 (6.9)  History of MI 1,730 (7.6) 625 (11.4) 420 (4.3) 449 (10.7) 236 (7.3) 1,105 (6.4)  History of stroke 1,062 (4.7) 384 (7.0) 278 (2.9) 239 (5.7) 161 (5.0) 678 (3.9)  History of diabetes 2,408 (10.6) 535 (9.8) 1,180 (12.1) 252 (6.0) 411 (13.7) 1,873 (10.9) Disease counts (N, %)  No disease 16,621 (73.3) 3,615 (65.9) 7,660 (78.4) 3,059 (72.6) 2,287 (70.9) 13,006 (75.6)  1 disease 5,116 (22.6) 1,491 (27.2) 1,881 (19.3) 979 (23.3) 765 (23.7) 3,625 (21.1)  2 diseases 858 (3.8) 332 (6.1) 211 (2.2) 160 (3.8) 155 (4.8) 526 (3.1)  3 diseases 92 (0.4) 40 (0.7) 19 (0.2) 12 (0.3) 21 (0.7) 52 (0.3)  4 diseases 5 (0.0) 4 (0.1) 0 (0.0) 1 (0.0) 0 (0.0) 1 (0.1) All studies combined CSHA CHANCES cohorts EPIC-Elderly Tromsø ESTHER Combined N 22,692 5,482 9,771 4,211 3,228 17,210 Age (y) (mean (SD)) 71.3 (5.9) 76.5 (7.0) 68.7 (3.2) 72.8 (5.8) 68.8 (2.9) 69.7 (4.3)  65–74 (N, %) 17,514 (77.2) 2,337 (42.6) 9,240 (94.6) 2,759 (65.5) 3,178 (98.5) 15,177 (88.2)  75–84 4,165 (18.4) 2,305 (42.1) 530 (5.4) 1,280 (30.4) 50 (1.6) 1,860 (10.8)  85+ 1,013 (4.5) 840 (15.3) 1 (0.0) 172 (4.1) 0 (0.0) 173 (1.0) Sex (N, %)  Men 8,639 (38.1) 3,133 (57.2) 2,964 (30.3) 1,795 (42.6) 1,531 (47.4) 6,290 (36.6)  Women 14,053 (61.9) 2,349 (42.9) 6,807 (69.7) 2,416 (57.4) 1,697 (52.6) 10,920 (63.5) Education (N, %)  Less than high school 14,344 (63.2) 1,757 (32.1) 7,290 (74.6) 2,850 (67.8) 2,447 (75.8) 12,587 (73.1)  High school 5,942 (26.2) 2,292 (41.8) 1,977 (20.2) 1,039 (24.7) 634 (19.6) 3,650 (21.2)  More than high school 2,406 (10.6) 1,433 (26.1) 504 (5.2) 322 (7.7) 147 (4.6) 973 (5.7) Smoking status (N, %)  Ever smoker 10,661 (47.0) 2,792 (50.9) 3,807 (39.0) 2,604 (61.8) 1,458 (45.2) 7,869 (45.7)  Never smoker 12,031 (53.0) 2,690 (49.1) 5,964 (61.0) 1,607 (38.2) 1,770 (54.8) 9,341 (54.3) Average duration of mortality follow-up (y) 10.02 (±3.9) 7.43 (±3.0) 10.83 (±3.4) 11.09 (±5.2) 10.58 (±2.6) 10.84 (±3.8) Self-reported history of chronic conditions (N, %)  History of cancer 1,928 (8.5) 747 (13.6) 482 (4.9) 399 (9.5) 300 (9.3) 1,181 (6.9)  History of MI 1,730 (7.6) 625 (11.4) 420 (4.3) 449 (10.7) 236 (7.3) 1,105 (6.4)  History of stroke 1,062 (4.7) 384 (7.0) 278 (2.9) 239 (5.7) 161 (5.0) 678 (3.9)  History of diabetes 2,408 (10.6) 535 (9.8) 1,180 (12.1) 252 (6.0) 411 (13.7) 1,873 (10.9) Disease counts (N, %)  No disease 16,621 (73.3) 3,615 (65.9) 7,660 (78.4) 3,059 (72.6) 2,287 (70.9) 13,006 (75.6)  1 disease 5,116 (22.6) 1,491 (27.2) 1,881 (19.3) 979 (23.3) 765 (23.7) 3,625 (21.1)  2 diseases 858 (3.8) 332 (6.1) 211 (2.2) 160 (3.8) 155 (4.8) 526 (3.1)  3 diseases 92 (0.4) 40 (0.7) 19 (0.2) 12 (0.3) 21 (0.7) 52 (0.3)  4 diseases 5 (0.0) 4 (0.1) 0 (0.0) 1 (0.0) 0 (0.0) 1 (0.1) Note: CSHA = Canadian Study on Health and Aging; CHANCE = Consortium on Health and Ageing: Network of Cohorts in Europe and United States; ESTHER = Epidemiological Study on the Chances of Prevention, Early Recognition and Optimised Treatment of Chronic Diseases in the Older Population; EPIC = European Prospective Investigation Into Cancer and Nutrition; MI = myocardial infarction. View Large Table 3. Hazard Rate (HR) and Rate Advancement Period (RAP) Estimates and Corresponding 95% CI for Overall Mortality Associated With Specific Chronic Disease Combinations Disease Status at Baseline N N of deaths Age and Sex Age, sex, smoking, educationa HR (95% CI) RAP (95% CI) HR (95% CI) RAP (95% CI) Referenceb 16,621 5,480 1 (reference) (reference) 1 (reference) (reference) Cancer 1,496 728 1.38 (1.28–1.49) 2.97 (2.23–3.72) 1.37 (1.27–1.48) 2.83 (2.10–3.56) MI 1,169 688 1.57 (1.45–1.70) 4.14 (3.37–4.92) 1.53 (1.42–1.66) 3.84 (3.09–4.61) Stroke 621 366 1.59 (1.43–1.77) 4.24 (3.23–5.27) 1.56 (1.41–1.74) 4.02 (3.03–5.03) Diabetes 1,830 785 1.77 (1.64–1.91) 5.26 (4.53–6.00) 1.76 (1.64–1.90) 5.11 (4.39–5.83) Cancer and MI 141 99 2.10 (1.72–2.56) 6.79 (4.91–8.70) 2.11 (1.73–2.57) 6.70 (4.85–8.56) Cancer and stroke 78 61 2.42 (1.88–3.12) 8.12 (5.73–10.53) 2.34 (1.82–3.02) 7.66 (5.32–10.02) Cancer and diabetes 157 77 2.01 (1.60–2.52) 6.40 (4.28–8.54) 1.96 (1.56–2.45) 6.04 (3.97–8.13) MI and stroke 141 104 2.62 (2.15–3.18) 8.84 (6.99–10.71) 2.46 (2.02–2.99) 8.08 (6.27–9.91) MI and diabetes 195 146 3.29 (2.79–3.88) 10.93 (9.35–12.55) 3.18 (2.69–3.75) 10.39 (8.84–11.97) Stroke and diabetes 146 94 2.54 (2.07–3.11) 8.54 (6.61–10.50) 2.48 (2.02–3.05) 8.18 (6.29–10.09) Cancer and MI and stroke 17 13 3.13 (1.81–5.40) 10.48 (5.36–15.63) 3.40 (1.97–5.86) 11.00 (5.98–16.04) Cancer and MI and diabetes 21 18 4.25 (2.68–6.76) 13.30 (8.94–17.69) 4.22 (2.66–6.71) 12.95 (8.69–17.26) MI and stroke and diabetes 41 37 5.06 (3.66–6.99) 14.88 (11.82–17.99) 4.74 (3.43–6.56) 14.00 (11.00–17.04) Disease Status at Baseline N N of deaths Age and Sex Age, sex, smoking, educationa HR (95% CI) RAP (95% CI) HR (95% CI) RAP (95% CI) Referenceb 16,621 5,480 1 (reference) (reference) 1 (reference) (reference) Cancer 1,496 728 1.38 (1.28–1.49) 2.97 (2.23–3.72) 1.37 (1.27–1.48) 2.83 (2.10–3.56) MI 1,169 688 1.57 (1.45–1.70) 4.14 (3.37–4.92) 1.53 (1.42–1.66) 3.84 (3.09–4.61) Stroke 621 366 1.59 (1.43–1.77) 4.24 (3.23–5.27) 1.56 (1.41–1.74) 4.02 (3.03–5.03) Diabetes 1,830 785 1.77 (1.64–1.91) 5.26 (4.53–6.00) 1.76 (1.64–1.90) 5.11 (4.39–5.83) Cancer and MI 141 99 2.10 (1.72–2.56) 6.79 (4.91–8.70) 2.11 (1.73–2.57) 6.70 (4.85–8.56) Cancer and stroke 78 61 2.42 (1.88–3.12) 8.12 (5.73–10.53) 2.34 (1.82–3.02) 7.66 (5.32–10.02) Cancer and diabetes 157 77 2.01 (1.60–2.52) 6.40 (4.28–8.54) 1.96 (1.56–2.45) 6.04 (3.97–8.13) MI and stroke 141 104 2.62 (2.15–3.18) 8.84 (6.99–10.71) 2.46 (2.02–2.99) 8.08 (6.27–9.91) MI and diabetes 195 146 3.29 (2.79–3.88) 10.93 (9.35–12.55) 3.18 (2.69–3.75) 10.39 (8.84–11.97) Stroke and diabetes 146 94 2.54 (2.07–3.11) 8.54 (6.61–10.50) 2.48 (2.02–3.05) 8.18 (6.29–10.09) Cancer and MI and stroke 17 13 3.13 (1.81–5.40) 10.48 (5.36–15.63) 3.40 (1.97–5.86) 11.00 (5.98–16.04) Cancer and MI and diabetes 21 18 4.25 (2.68–6.76) 13.30 (8.94–17.69) 4.22 (2.66–6.71) 12.95 (8.69–17.26) MI and stroke and diabetes 41 37 5.06 (3.66–6.99) 14.88 (11.82–17.99) 4.74 (3.43–6.56) 14.00 (11.00–17.04) Notes: RAP estimates express the impact of a given exposure (i.e. combination of chronic conditions) on the risk of death, by determining the time (in years) by which the mortality rate is anticipated for study participants exposed compared to nonexposed, that is, individuals without any of the four investigated chronic conditions. All studies combined (EPIC-Elderly, Tromsø, ESTHER, and CSHA)c. CI = confidence interval; CHANCE = Consortium on Health and Ageing: Network of Cohorts in Europe and United States; CSHA = Canadian Study on Health and Aging; ESTHER = Epidemiological Study on the Chances of Prevention, Early Recognition and Optimised Treatment of Chronic Diseases in the Older Population; EPIC = European Prospective Investigation Into Cancer and Nutrition; MI = myocardial infarction. aAdjusted for sex, smoking status (ever/never), level of education (less than high school/high school/more than high school). bReference group had none of the four conditions under study. cStudy was included as a stratification factor in the analyses to account for baseline differences among the cohorts. View Large Table 3. Hazard Rate (HR) and Rate Advancement Period (RAP) Estimates and Corresponding 95% CI for Overall Mortality Associated With Specific Chronic Disease Combinations Disease Status at Baseline N N of deaths Age and Sex Age, sex, smoking, educationa HR (95% CI) RAP (95% CI) HR (95% CI) RAP (95% CI) Referenceb 16,621 5,480 1 (reference) (reference) 1 (reference) (reference) Cancer 1,496 728 1.38 (1.28–1.49) 2.97 (2.23–3.72) 1.37 (1.27–1.48) 2.83 (2.10–3.56) MI 1,169 688 1.57 (1.45–1.70) 4.14 (3.37–4.92) 1.53 (1.42–1.66) 3.84 (3.09–4.61) Stroke 621 366 1.59 (1.43–1.77) 4.24 (3.23–5.27) 1.56 (1.41–1.74) 4.02 (3.03–5.03) Diabetes 1,830 785 1.77 (1.64–1.91) 5.26 (4.53–6.00) 1.76 (1.64–1.90) 5.11 (4.39–5.83) Cancer and MI 141 99 2.10 (1.72–2.56) 6.79 (4.91–8.70) 2.11 (1.73–2.57) 6.70 (4.85–8.56) Cancer and stroke 78 61 2.42 (1.88–3.12) 8.12 (5.73–10.53) 2.34 (1.82–3.02) 7.66 (5.32–10.02) Cancer and diabetes 157 77 2.01 (1.60–2.52) 6.40 (4.28–8.54) 1.96 (1.56–2.45) 6.04 (3.97–8.13) MI and stroke 141 104 2.62 (2.15–3.18) 8.84 (6.99–10.71) 2.46 (2.02–2.99) 8.08 (6.27–9.91) MI and diabetes 195 146 3.29 (2.79–3.88) 10.93 (9.35–12.55) 3.18 (2.69–3.75) 10.39 (8.84–11.97) Stroke and diabetes 146 94 2.54 (2.07–3.11) 8.54 (6.61–10.50) 2.48 (2.02–3.05) 8.18 (6.29–10.09) Cancer and MI and stroke 17 13 3.13 (1.81–5.40) 10.48 (5.36–15.63) 3.40 (1.97–5.86) 11.00 (5.98–16.04) Cancer and MI and diabetes 21 18 4.25 (2.68–6.76) 13.30 (8.94–17.69) 4.22 (2.66–6.71) 12.95 (8.69–17.26) MI and stroke and diabetes 41 37 5.06 (3.66–6.99) 14.88 (11.82–17.99) 4.74 (3.43–6.56) 14.00 (11.00–17.04) Disease Status at Baseline N N of deaths Age and Sex Age, sex, smoking, educationa HR (95% CI) RAP (95% CI) HR (95% CI) RAP (95% CI) Referenceb 16,621 5,480 1 (reference) (reference) 1 (reference) (reference) Cancer 1,496 728 1.38 (1.28–1.49) 2.97 (2.23–3.72) 1.37 (1.27–1.48) 2.83 (2.10–3.56) MI 1,169 688 1.57 (1.45–1.70) 4.14 (3.37–4.92) 1.53 (1.42–1.66) 3.84 (3.09–4.61) Stroke 621 366 1.59 (1.43–1.77) 4.24 (3.23–5.27) 1.56 (1.41–1.74) 4.02 (3.03–5.03) Diabetes 1,830 785 1.77 (1.64–1.91) 5.26 (4.53–6.00) 1.76 (1.64–1.90) 5.11 (4.39–5.83) Cancer and MI 141 99 2.10 (1.72–2.56) 6.79 (4.91–8.70) 2.11 (1.73–2.57) 6.70 (4.85–8.56) Cancer and stroke 78 61 2.42 (1.88–3.12) 8.12 (5.73–10.53) 2.34 (1.82–3.02) 7.66 (5.32–10.02) Cancer and diabetes 157 77 2.01 (1.60–2.52) 6.40 (4.28–8.54) 1.96 (1.56–2.45) 6.04 (3.97–8.13) MI and stroke 141 104 2.62 (2.15–3.18) 8.84 (6.99–10.71) 2.46 (2.02–2.99) 8.08 (6.27–9.91) MI and diabetes 195 146 3.29 (2.79–3.88) 10.93 (9.35–12.55) 3.18 (2.69–3.75) 10.39 (8.84–11.97) Stroke and diabetes 146 94 2.54 (2.07–3.11) 8.54 (6.61–10.50) 2.48 (2.02–3.05) 8.18 (6.29–10.09) Cancer and MI and stroke 17 13 3.13 (1.81–5.40) 10.48 (5.36–15.63) 3.40 (1.97–5.86) 11.00 (5.98–16.04) Cancer and MI and diabetes 21 18 4.25 (2.68–6.76) 13.30 (8.94–17.69) 4.22 (2.66–6.71) 12.95 (8.69–17.26) MI and stroke and diabetes 41 37 5.06 (3.66–6.99) 14.88 (11.82–17.99) 4.74 (3.43–6.56) 14.00 (11.00–17.04) Notes: RAP estimates express the impact of a given exposure (i.e. combination of chronic conditions) on the risk of death, by determining the time (in years) by which the mortality rate is anticipated for study participants exposed compared to nonexposed, that is, individuals without any of the four investigated chronic conditions. All studies combined (EPIC-Elderly, Tromsø, ESTHER, and CSHA)c. CI = confidence interval; CHANCE = Consortium on Health and Ageing: Network of Cohorts in Europe and United States; CSHA = Canadian Study on Health and Aging; ESTHER = Epidemiological Study on the Chances of Prevention, Early Recognition and Optimised Treatment of Chronic Diseases in the Older Population; EPIC = European Prospective Investigation Into Cancer and Nutrition; MI = myocardial infarction. aAdjusted for sex, smoking status (ever/never), level of education (less than high school/high school/more than high school). bReference group had none of the four conditions under study. cStudy was included as a stratification factor in the analyses to account for baseline differences among the cohorts. View Large Diabetes was the most common condition to occur both in isolation and in combination with other chronic diseases (Table 3). Large differences in prevalence were observed among the specific disease combinations with the same number of conditions. All single diseases and disease combinations were associated with a statistically significant increased mortality rate compared to the reference category when analyzing all studies combined ranging from 2.97 (2.23, 3.72) for cancer alone to 14.88 (11.82, 17.99) for MI, stroke, and diabetes (Table 3). Separate analyses in the CSHA and CHANCES cohorts revealed the same ranking of conditions (Supplementary Table 1). Among the single-disease groups, diabetes was the condition associated with the largest age- and sex-adjusted HR (1.77; 95% CI, 1.64–1.91) and advanced the rate of dying by the most years (5.26 years; 95% CI, 4.53–6.00) compared to the reference group. A history of cancer alone had the smallest impact on mortality and advanced the rate of dying by 2.97 years (95% CI, 2.23–3.72). Among those identified with two conditions, there was a more than 4 year difference in the lowest to highest RAP (6.40 years for cancer and diabetes to 10.93 years for MI and diabetes; p < 0.0001). All HRs and RAPs attenuated slightly after additional adjustment for smoking status and level of education, but CIs did not cross one or zero, respectively (Table 3). Across all cohorts, effect modification by age was observed (p < 0.0001) with RAP estimates being larger in younger participants (<70 years) compared to older participants (≥70 years) (Supplementary Table 2). No interaction with sex, level of education, or smoking status was observed. Irrespective of disease status, men had the same mortality rate as women who are 5 years their junior (RAP 4.96, 95% CI, 4.53–5.39; results not shown). Similar HRs and RAPs were observed in analyses using a frailty model with random effect for study, suggesting that the hazard function was similar across the four cohorts (data not shown). Additional adjustment for BMI and hypertension in the CHANCES cohorts did not change the results notably (results not shown). Supplementary Table 3 shows the HR and RAP with and without adjustment for ADL disability, and adjustment for the combined presence of 11 other chronic conditions in the CSHA cohort (participants with missing data on disability and diseases were excluded from these analyses). Figure 1 rank orders the RAP of all single conditions and combinations of two and three conditions in order of magnitude (based on the confounder adjusted model) and depicts whether any two RAP estimates were statistically different. Pairwise comparison of two RAPs is shown in each cell and the statistical significance of each comparison is indicated by the shading; a vertical pattern fill meaning two RAPs are significantly different at p <0.05, a light gray shading indicating differences at p <0.01, and dark gray at p <0.001. For example, the RAP of having cancer alone (2.83 [2.10, 3.56] years) was lower than the RAP associated with any other single or combination of condition(s), and the RAP of diabetes alone (5.11 [4.39, 5.83]) was significantly higher than that of MI (4.02 [3.03, 5.03]). The RAPs of all combinations of two conditions were significantly higher than those for single conditions, with two exceptions. As well, the RAP for the combination of MI and diabetes (10.39 [8.84, 11.97]) was in the same order of magnitude as the RAP of most combinations of three diseases and significantly higher than the RAP of all combinations of two diseases that include cancer (ranging from 6.04 [3.97, 8.13] to 7.66 [5.32, 10.02]). Figure 1. View largeDownload slide Map comparing RAP estimates of single conditions, dyads, and tryads of conditions. Map was established using all studies combined. Pairwise comparison of two RAPs is shown in each cell. The following color scheme was applied: white for p >0.05, vertical pattern fill when p <0.05, shade of gray when p <0.01, and dark gray for p <0.001. All RAP estimates were derived from a Cox proportional Hazard model adjusted for age and sex. Participants who did not have any of four chronic conditions at baseline formed the reference group. Figure 1. View largeDownload slide Map comparing RAP estimates of single conditions, dyads, and tryads of conditions. Map was established using all studies combined. Pairwise comparison of two RAPs is shown in each cell. The following color scheme was applied: white for p >0.05, vertical pattern fill when p <0.05, shade of gray when p <0.01, and dark gray for p <0.001. All RAP estimates were derived from a Cox proportional Hazard model adjusted for age and sex. Participants who did not have any of four chronic conditions at baseline formed the reference group. Discussion Our prospective analyses of four cohort studies from Europe and Canada estimated the average time period by which mortality rates are accelerated among people aged 65 years and older with different combinations of cancer and cardiometabolic conditions (MI, stroke, and diabetes). Given the increasing prevalence and mortality burden associated with coexisting chronic conditions, our two main findings have important clinical and public health implications. First, we observed that certain disease (combinations) had a mortality burden that was similar or even higher compared to combinations with larger disease counts. This suggests that simple disease counts may not reflect the full underlying complexity of how multiple chronic conditions impact survival (results for disease counts are shown in Supplementary Table 4). Positioning these findings in the context of previous research is limited by the paucity of studies examining the impact of combinations of chronic conditions on mortality. A recent systematic review on the combined effect of cardiometabolic conditions on mortality identified only a handful of studies (9), which were not restricted to older populations. This review was part of a larger publication reporting only small differences in mortality rate among clusters of MI, stroke, and diabetes with the same number of conditions. In contrast to this study, we applied an aging lens and restricted our sample to adults 65 years and older. Multimorbidity prevalence and patterns have shown to change during the life course (22,23), and results from single-disease studies suggest that combinations of conditions likely impact mortality differently in middle-aged and senior adults (24–26). Two other reports previously presented data on unique combinations of conditions and survival in older populations (27,28), but these were not specifically focussed on cancer and cardiometabolic diseases and failed to specify the impact of disease combination on the progression of mortality. Second, in contrast to traditional measures of effect, the RAP expresses the time dimension by which mortality is impacted by disease status, capturing the ‘aging effect’ of having a multimorbidity. As a result, the RAP can help effectively communicate the impact of having chronic condition(s) on mortality, as well as the benefits of adopting preventive strategies to avoid the accumulation of conditions. Where burden of disease studies traditionally assign the highest priority ranking to the leading causes of death (i.e. cancer and cardiovascular disease in developed countries), we show that diabetes has the largest impact on accelerating people’s mortality rate. Unfortunately, an increase in misinterpretation of the RAP has been seen in literature over recent years where the RAP is incorrectly explained as the difference in mean survival time, or the time by which a survival curve is shifted between exposed and unexposed (29). In addition, although previous studies estimating RAP do not explore potential effect modification between age and exposures, we observed a strong interaction between baseline age and disease combinations on mortality rates, suggesting that the impact of multimorbidity on mortality RAPs decreases with advancing age. As such, our RAPs provide an average RAP, with even larger RAP for younger seniors with multiple conditions. As well, sex had a strong independent effect on the RAP above and beyond the presence of the four diseases under study; the RAP of male sex was comparable to that associated with having diabetes. When studying the effect of complex disease combinations, traditional models of interaction provide limited interpretation, which is why we opted to report the main effect of mutually exclusive disease combinations. We chose to focus our analyses on four key diseases that are among the most prevalent, costly, and preventable causes of death in high-income countries making them priority targets for public health actions. However, we could have potentially introduced bias into our estimates if certain excluded conditions more frequently coexist with one of the key conditions under study. Analyses of the CSHA data that included 11 additional chronic conditions did not change the results of our study. We used information about disease status at baseline and were not able to account for additional disease accumulation during follow-up in our analyses. We also had no information about the duration, temporality, and severity of the conditions. Although we found self-reported disease status to have moderate-to-good agreement with clinical assessment in the CSHA, and disease rates to be comparable to population prevalence across studies, misclassification of disease status may have occurred; likely resulting in an underestimation of disease occurrence leading to attenuated effect estimates. The mortality rate in our older population may be modified by the duration of disease, where conditions developed early in life may have a different effect on survival than those developed later. We restricted our analyses to seniors because these have the largest multimorbidity prevalence. Our RAP estimates may not be generalizable to younger populations as also indicated by the interaction observed with baseline age. CSHA participants with complete data were younger, higher educated, and more physically active compared those with incomplete data, suggesting that our complete case analysis may have resulted in a healthier sample. Although the CSHA cohort was on average higher educated than the CHANCES studies, similar severity ranking of conditions was observed across cohorts. We did not have information on the type of cancer reported at baseline and our sample is likely overrepresented by patients diagnosed with cancers with higher long-term survival rates. In addition, changes in the definition of MI, published in 2000, have led to a substantial increase in diagnosis. Although the studies included in the current analysis collected information on disease status before these changes took effect, research suggests that patients diagnosed with MI by this new definition have a higher long-term mortality (30), potentially resulting in an underestimation of the RAP associated with MI in our analyses. Our findings highlight that the care of older people with multiple chronic conditions requires a targeted approach that takes into account the presence of unique combinations of conditions. Understanding how these conditions accelerate the rate of dying alone and in combination, can help aid decision making. Supplementary Material Supplementary data is available at The Journals of Gerontology, Series A: Biological Sciences and Medical Sciences online. Funding The CHANCES project is coordinated by the Hellenic Health Foundation, Greece and received funding by the FP7 framework programme of DG-RESEARCH in the European Commission (grant agreement no. HEALTH-F3-2010–242244). The CSHA was funded by the Seniors Independence Research Programme, administered by the National Health and Development Programme of Health and Welfare Canada. Neither funder had any role in data collection, data analysis, data interpretation, or writing of the report. Lauren Griffith is supported by a Canadian Institutes of Health Research New Investigator’s Award, and the McLaughlin Foundation Professorship in Population and Public Health. Parminder Raina holds a Tier 1 Canada Research Chair in Geroscience and the Raymond and Margaret Labarge Chair in Research and Knowledge Application for Optimal Aging. Anne Gilsing is supported by a Canadian Institute of Health Research Post-doctoral Fellowship and a Michael G. DeGroote Fellowship Award from McMaster University. Conflict of Interest None reported. References 1. GBD 2013 Mortality and Causes of Death Collaborators . Global, regional, and national age-sex specific all-cause and cause-specific mortality for 240 causes of death, 1990–2013: a systematic analysis for the Global Burden of Disease Study 2013 . Lancet . 2015 ; 385 : 117 – 71 . doi: 10.1016/S0140-6736(14)61682-2 CrossRef Search ADS PubMed 2. United Nations Department of Economic and Social Affairs Population Division . World Population Ageing 2013 . New York, NY: United Nations Secretariat; 2013 . Report No.: ST/ESA/SER.A/348. 3. United Nations Department of Economic and Social Affairs Population Division . World Population Prospects: The 2015 Revision, Key Findings and Advance Tables . New York, NY: United Nations Secretariat; 2015 . Report No.: Working Paper No. ESA/P/WP.241. 4. World Healh Organization . Global Status Report on Noncommunicable Diseases 2014 . Geneva, Switzerland : World Health Organization ; 2014 . ISBN: 978 92 4 156485 4 5. Fortin M , Stewart M , Poitras ME , Almirall J , Maddocks H . A systematic review of prevalence studies on multimorbidity: toward a more uniform methodology . Ann Fam Med . 2012 ; 10 : 142 – 151 . doi: 10.1370/afm.1337 Google Scholar CrossRef Search ADS PubMed 6. Pefoyo AJ , Bronskill SE , Gruneir A et al. The increasing burden and complexity of multimorbidity . BMC Public Health . 2015 ; 15 : 415 . doi: 10.1186/s12889-015-1733-2 Google Scholar CrossRef Search ADS PubMed 7. Marengoni A , Angleman S , Melis R et al. Aging with multimorbidity: a systematic review of the literature . Ageing Res Rev . 2011 ; 10 : 430 – 439 . doi: 10.1016/j.arr.2011.03.003 Google Scholar CrossRef Search ADS PubMed 8. Fabbri E , Zoli M , Gonzalez-Freire M , Salive ME , Studenski SA , Ferrucci L . Aging and multimorbidity: new tasks, priorities, and frontiers for integrated gerontological and clinical research . J Am Med Dir Assoc . 2015 ; 16 : 640 – 647 . doi: 10.1016/j.jamda.2015.03.013 Google Scholar CrossRef Search ADS PubMed 9. Di Angelantonio E , Kaptoge S , Wormser D et al. Association of cardiometabolic multimorbidity with mortality . JAMA . 2015 ; 314 : 52 – 60 . doi: 10.1001/jama.2015.7008 Google Scholar CrossRef Search ADS PubMed 10. Boffetta P , Bobak M , Borsch-Supan A et al. The consortium on health and ageing: network of Cohorts in Europe and the United States (CHANCES) project–design, population and data harmonization of a large-scale, international study . Eur J Epidemiol . 2014 ; 29 : 929 – 936 . doi: 10.1007/s10654-014-9977-1 Google Scholar CrossRef Search ADS PubMed 11. Kuulasmaa K , Palosaari T . Contributors from Partners of the Consortium on Health and Ageing: Network of Cohorts in Europe and the United States (CHANCES). CHANCES cohort descriptions, assessment of the availability and quality of data, and definitions of variables. MORGAM Project e-publications [Internet] . 2015 ; 6 URN: NBN: fi-fe201501151161. http://www.thl.fi/publications/morgam/chances_d9/index.html. Accessed October, 2017. 12. Riboli E , Kaaks R . The EPIC project: rationale and study design. European prospective investigation into cancer and nutrition . Int J Epidemiol 1997 ; 26 ( suppl 1 ): S6 – 14 . doi: 10.1079/PHN2002394 Google Scholar CrossRef Search ADS PubMed 13. Raum E , Rothenbacher D , Löw M , Stegmaier C , Ziegler H , Brenner H . Changes of cardiovascular risk factors and their implications in subsequent birth cohorts of older adults in Germany: a life course approach . Eur J Cardiovasc Prev Rehabil . 2007 ; 14 : 809 – 814 . doi: 10.1097/HJR.0b013e3282eeb308 Google Scholar CrossRef Search ADS PubMed 14. Jacobsen BK , Eggen AE , Mathiesen EB , Wilsgaard T , Njølstad I . Cohort profile: the Tromso study . Int J Epidemiol . 2012 ; 41 : 961 – 967 . doi: 10.1093/ije/dyr049 Google Scholar CrossRef Search ADS PubMed 15. McDowell I , Hill G , Lindsay J . An overview of the Canadian Study of Health and Aging . Int Psychogeriatr 2001 ; 13 ( supp 1 ): 7 – 18 . doi: 10.1017/S1041610202007949 Google Scholar CrossRef Search ADS PubMed 16. Engstad T , Bonaa KH , Viitanen M . Validity of self-reported stroke: the Tromso Study . Stroke . 2000 ; 31 : 1602 – 1607 . doi:10.1161/01.STR.31.7.1602 Google Scholar CrossRef Search ADS PubMed 17. Okura Y , Urban LH , Mahoney DW , Jacobsen SJ , Rodeheffer RJ . Agreement between self-report questionnaires and medical record data was substantial for diabetes, hypertension, myocardial infarction and stroke but not for heart failure . J Clin Epidemiol . 2004 ; 57 : 1096 – 1103 . doi: 10.1016/ j.jclinepi.2004.04.005 Google Scholar CrossRef Search ADS PubMed 18. van den Akker M , van Steenkiste B , Krutwagen E , Metsemakers JF . Disease or no disease? Disagreement on diagnoses between self-reports and medical records of adult patients . Eur J Gen Pract . 2015 ; 21 : 45 – 51 . doi: 10.3109/13814788.2014.907266 Google Scholar CrossRef Search ADS PubMed 19. The Canadian Study on Health and Aging Working Group . Disability and frailty among elderly Canadians: a comparison of six surveys . Int Psychogeriatr 2001 ; 13 ( supp 1 ): 159 – 67 . doi: 10.1017/S1041610202008104 CrossRef Search ADS PubMed 20. Dubois MF , Hebert R . Imputation of missing dates of death or institutionalization for time-to-event analyses in the Canadian Study of Health and Aging . Int Psychogeriatr 2001 ; 13 ( supp 1 ): 91 – 7 . doi: 10.1017/S1041610202008025 Google Scholar CrossRef Search ADS PubMed 21. Brenner H , Gefeller O , Greenland S . Risk and rate advancement periods as measures of exposure impact on the occurrence of chronic diseases . Epidemiology . 1993 ; 4 : 229 – 236 . PMID: 8512987. Google Scholar CrossRef Search ADS PubMed 22. Violan C , Foguet-Boreu Q , Flores-Mateo G et al. Prevalence, determinants and patterns of multimorbidity in primary care: a systematic review of observational studies . PLoS One . 2014 ; 9 : e102149 . doi: 10.1371/journal.pone.0102149 Google Scholar CrossRef Search ADS PubMed 23. Prados-Torres A , Calderón-Larrañaga A , Hancco-Saavedra J , Poblador-Plou B , van den Akker M . Multimorbidity patterns: a systematic review . J Clin Epidemiol . 2014 ; 67 : 254 – 266 . doi: 10.1016/j.jclinepi.2013.09.021 Google Scholar CrossRef Search ADS PubMed 24. Brønnum-Hansen H , Davidsen M , Thorvaldsen P ; Danish MONICA Study Group . Long-term survival and causes of death after stroke . Stroke . 2001 ; 32 : 2131 – 2136 . doi:10.1161/hs0901.094253 Google Scholar CrossRef Search ADS PubMed 25. Smolina K , Wright FL , Rayner M , Goldacre MJ . Long-term survival and recurrence after acute myocardial infarction in England, 2004 to 2010 . Circ Cardiovasc Qual Outcomes 2012 ; 5 : 532 – 540 . doi: 10.1161/CIRCOUTCOMES.111.964700 Google Scholar CrossRef Search ADS PubMed 26. De Angelis AR , Sant M , Coleman MP et al. Cancer survival in Europe 1999–2007 by country and age: results of EUROCARE--5-a population-based study . Lancet Oncol 2014 ; 15 : 23 – 34 . doi: 10.1016/S1470-2045(13)70546-1 Google Scholar CrossRef Search ADS PubMed 27. Caughey GE , Ramsay EN , Vitry AI et al. Comorbid chronic diseases, discordant impact on mortality in older people: a 14-year longitudinal population study . J Epidemiol Community Health . 2010 ; 64 : 1036 – 1042 . doi: 10.1136/jech.2009.088260 Google Scholar CrossRef Search ADS PubMed 28. Lee TA , Shields AE , Vogeli C et al. Mortality rate in veterans with multiple chronic conditions . J Gen Intern Med 2007 ; 22 ( suppl 3 ): 403 – 7 . doi: 10.1007/s11606-007-0277-2 Google Scholar CrossRef Search ADS PubMed 29. Discacciati A , Bellavia A , Orsini N , Greenland S . On the interpretation of risk and rate advancement periods . Int J Epidemiol . 2016 ; 45 : 278 – 284 . doi: 10.1093/ije/dyv320 Google Scholar CrossRef Search ADS PubMed 30. Hochholzer W , Buettner HJ , Trenk D et al. New definition of myocardial infarction: impact on long-term mortality . Am J Med . 2008 ; 121 : 399 – 405 . doi: 10.1016/j.amjmed.2008.01.033 Google Scholar CrossRef Search ADS PubMed © The Author(s) 2018. Published by Oxford University Press on behalf of The Gerontological Society of America. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

The Journals of Gerontology Series A: Biomedical Sciences and Medical SciencesOxford University Press

Published: Mar 19, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off