- Email: [email protected]
Multilevel socioeconomic status and incidence of frailty post myocardial infarction
International Journal of Cardiology 170 (2014) 338–343
Contents lists available at ScienceDirect
International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard
Multilevel socioeconomic status and incidence of frailty post myocardial infarction☆,☆☆,★ Vicki Myers a, Yaacov Drory b, Uri Goldbourt a, Yariv Gerber a,⁎ a b
Department of Epidemiology and Preventive Medicine, School of Public Health, Sackler Faculty of Medicine, Tel Aviv University, Israel Department of Rehabilitation, Sackler Medical School, Tel Aviv University, Israel
a r t i c l e
i n f o
Article history: Received 7 February 2013 Received in revised form 2 September 2013 Accepted 2 November 2013 Available online 12 November 2013 Keywords: Myocardial infarction Frailty Socioeconomic status Risk discrimination Neighborhood socioeconomic status
a b s t r a c t Background: Frailty predicts mortality and hospitalizations in post-myocardial infarction (MI) patients. Socioeconomic status (SES) demonstrates a clear relationship with post-MI outcomes and is also associated with community frailty; however this relationship has yet to be evaluated in post-MI patients. We investigated the predictive value of socioeconomic factors in the development of post-MI frailty. Methods: A cohort of 1151 post-MI patients was followed up from initial hospitalization in 1992–1993 for 10– 13 years. Individual and neighborhood SES measures were assessed at baseline and frailty was assessed during follow-up via an index of deficit accumulation. Logistic regression models and discrimination indices enabled determination of the predictive value of socioeconomic factors over basic clinical variables in classifying frailty risk. Results: During follow-up, 399 patients (35%) developed frailty. Individual and neighborhood SES were significantly and independently associated with the risk of developing frailty. Low income patients had more than twice the risk of becoming frail compared with those with high income [odds ratio (OR), 2.29, 95% CI 1.41– 3.73]; while being in the lower vs. upper neighborhood SES tertile was associated with a 60% increased odds (OR, 1.60, 95% CI 1.03–2.49). Inclusion of multilevel SES yielded substantial gains in c-statistic (0.70 to 0.76), net reclassification improvement (21.4%) and integrated discrimination improvement (6.4%) over basic clinical factors (all p b 0.001), indicating increased predictive value and gains in sensitivity and specificity. Conclusions: Individual and neighborhood socioeconomic factors influence the development of frailty post-MI, and contribute to risk discrimination in this population. © 2013 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Frailty is a syndrome of increasing vulnerability and decreasing resistance to stressors resulting from a generalized decline in multiple physiological systems [1]. This emerging construct has been shown to predict both mortality [2–4] and disability [5]. Frailty has been demonstrated as clinically relevant in post-myocardial infarction (MI) patients, predicting mortality and healthcare use [6–8]. Despite variations in assessment techniques, frailty is consistently associated with poorer outcomes and greater mortality risk, as well as being an indicator of institutionalization and healthcare use [6,9,10]. However, the mechanisms underlying this phenomenon have yet to be elucidated. Socioeconomic differences in both individual and neighborhood domains are well-known to affect health outcomes and specifically ☆ Institution at which research performed: Tel Aviv University. ☆☆ See Acknowledgments for a list of participating medical centers and investigators. ★ Funding: This research was supported by Grant Award No. SGA 1204 from the Environment and Health Fund, Israel. ⁎ Corresponding author at: Department of Epidemiology and Preventive Medicine, School of Public Health, Sackler Faculty of Medicine, Tel Aviv University, Ramat Aviv, Tel Aviv 69978, Israel. Tel.: +972 3 6409867; fax: +972 3 6409868. E-mail address: [email protected] (Y. Gerber). 0167-5273/$ – see front matter © 2013 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ijcard.2013.11.009
cardiovascular prognosis, with income and occupation inversely related to cardiovascular mortality [11] and post-MI outcomes [12–15]. Socioeconomic status (SES) has further been associated with frailty in cross-sectional analyses. In an elderly Chinese sample, higher frailty score was associated with lower educational level and insufficient finances [16]. In the Women's Health and Aging Study, low-income women had twice the odds of frailty compared to high-income women, independent of age, and low education was associated with a threefold increased risk of frailty [17]. Beyond general population samples, the influence of SES on the development of frailty in MI survivors has not been investigated. Since post-MI patients constitute a highrisk population [18,19], with corresponding high levels of frailty [20], it is important to consider the antecedents of frailty in this group. This study investigated the predictive value of multilevel SES in the incidence of frailty in a cohort of post-MI patients followed up for 13 years. 2. Methods 2.1. Sample A cohort of first MI patients aged ≤65 years was followed up from initial hospitalization in 1992–1993 in one of the eight hospitals in central Israel, through 2002–2005, in the Israel Study of First Acute Myocardial Infarction. Of initial 1626 consecutive patients
V. Myers et al. / International Journal of Cardiology 170 (2014) 338–343 admitted, 81 died during initial hospitalization, a further 330 died before the end of follow-up 10–13 years later, and 24 were lost to follow-up. Of the remaining 1191 patients, 40 (3.4%) were classed as frail at baseline and excluded, leaving 1151 patients in the final sample. Clinical, sociodemographic and psychosocial data were collected from medical records and structured interviews approximately one week after the index hospitalization, and subsequently 10 to 13 years after MI. All aspects of the study were approved by the appropriate Institutional Ethics Committees. 2.2. Frailty index The Rockwood index of accumulation of deficits [21,22] was adapted to develop a frailty index comprising 40 items (Appendix 1), and patients were assessed 10–13 years after MI. A similar index was previously validated in the same cohort [8]. Variables included perceived health, comorbidity, functional limitations, weight loss, physical activity and psychological factors. Dichotomous items were coded as 0 if the deficit was absent and 1 if it was present; while ordinal or continuous variables were graded into a score between 0 and 1 (0 representing no impairment, 0.5 for minor impairment and 1 for major impairment). Scores were then summed and divided by the total number of variables to give a frailty index between 0 and 1, with 1 representing the greatest frailty. According to Rockwood et al., weighting of items is unnecessary since major disease usually goes hand in hand with other comorbidities, and the frailty index focuses on the accumulation of multiple problems, rather than their nature [23]. A threshold of ≥0.25 was used to define frailty, based on values used in previous research [2,24], a figure achieved by comparing a phenotypic definition of frailty with a frailty index, and modeling density distributions of deficits [24]. Patients were classified into two categories, defined as frail if they scored ≥0.25, and non-frail if they scored b0.25. 2.3. Frailty index variables 2.3.1. Perceived health Self-rated health was assessed on a 5-point scale from excellent to poor. Patients rated their own health currently and compared to 5 years previously. 2.3.2. Functional limitations Participants rated their functional fitness; the extent to which their health limited certain activities, including activities of daily living, housework and social activities; pain; energy; and how often they get sick, as part of the Hebrew version of the Medical Outcomes Study Short-Form 36 (SF-36). The reliability, validity, and responsiveness of the SF-36 are well documented both in general population and in patients with coronary artery disease [25,26]. 2.3.3. Comorbidity The presence of comorbid conditions, including diabetes, stroke, chronic obstructive pulmonary disorder (COPD), cancer, arthritis, Alzheimer's disease, Parkinson's disease, kidney disease, liver disease, ulcer and chronic disability, was determined via medical records and personal interviews. 2.3.4. Clinical variables Obesity was defined as body mass index ≥30 kg/m2. Current weight and weight loss since baseline were recorded. Leisure time physical activity was assessed by self-report including average frequency and duration of walking, cycling, swimming, gardening, going to the gym and team sports. These responses were summarized by a senior cardiologist into 3 groups: inactive, irregularly active and regularly active patients. Further details on physical activity were previously published [27,28]. Regular physical activity was defined as at least three 30 minute sessions per week, according to published guidelines [29]. 2.3.5. Psychological variables Patients completed the Mental Health Inventory, which assesses emotional distress and well-being, including questions on feelings of depression, anxiety, and loneliness. This questionnaire has shown validity and reliability in the Hebrew version [30]. 2.4. Socioeconomic status (SES) data Individual SES data were self-reported at study entry and included the following measures: family income relative to the national average (categorized as below average, average or above average), education (years of schooling), and pre-MI employment status (full-time or part-time vs. none) [13,31]. Neighborhood SES was estimated through an index developed by the Israel Central Bureau of Statistics, which summarizes socioeconomic measures from the 1995 National Census data, allowing the classification of small geographic units into SES categories, on a 20-point scale [32]. Tertiles of the scale distribution were used for analysis. 2.5. Additional variables Cardiovascular risk factors, MI characteristics and disease severity indices were recorded at the index hospitalization. Cigarette use was classified into current smoking vs. never or past smoking. Hypertension and hypercholesterolemia were defined according to standard criteria based on clinical and laboratory data. Comorbidity was defined based on the Charlson index as ≥1 point, indicating at least moderate comorbidity [33]. MI characteristics and severity indicators included infarct type
339
(Q wave or non-Q wave), Killip class N1, admission to intensive care unit (ICU), early revascularization, and pharmacotherapy use [β-blockers, angiotensin converting enzyme (ACE) inhibitors, and aspirin]. 2.6. Statistical analysis Differences in baseline characteristics by outcome status at the end of follow-up (not frail, frail, and deceased) were assessed with analysis of variance (ANOVA) for continuous variables and chi-square tests for categorical variables. Multivariable logistic regression models were constructed to assess the odds ratios (OR) and 95% confidence intervals (CI) for frailty associated with different sets of covariates. The demographic plus clinical model (“basic model”) included age, gender, hypertension, smoking, hypercholesterolemia, obesity, Q-wave MI, Killip class, comorbidity, self-rated health, admission to ICU and early revascularization, as well as β-blockers, ACE inhibitors, and aspirin on discharge. This model was repeated with the addition of SES variables (“extended model”), including education, family income, pre-MI employment, and neighborhood SES. These potential predictors were chosen based on previous literature and data availability. In a sensitivity analysis, left ventricular ejection fraction (available in 55% of the cohort) was additionally adjusted for in this subgroup. A linear regression was additionally performed with frailty score modeled as a continuous variable. Furthermore, in order to address selection bias introduced by exclusion of the 330 patients who had died before follow-up, so that their frailty score could not be evaluated, the Diehr method was applied [34,35]. Predicted probabilities of frailty at the end of follow-up were calculated for each model and categorized as less than 20%, 20% to 40%, and greater than 40%, corresponding to low, intermediate, and high risk, respectively. The incremental discriminatory value of the extended model over the basic model was examined by several methods. The integrated discrimination improvement (IDI) and net reclassification improvement (NRI) indices proposed by Pencina and colleagues [36] were calculated. In general, IDI measures the improvement in average sensitivity associated with the more advanced model, and subtracts any decrease in average specificity. NRI, in contrast, measures the extent to which individuals with and without the event of interest are appropriately reclassified into clinically accepted higher or lower risk categories with the extended versus basic model [31]. In addition, receiveroperating characteristic (ROC) curves were plotted for the basic and extended models. The c-statistic, a measure of area under the ROC curve, was calculated for each model and compared by methods previously described [37]. Missing values for family income (17%) and neighborhood SES (7%), as well as for variables comprising the frailty index, which ranged from 0 to 16% (mean, 5%), were imputed using multiple imputation methodology. The Markov Chain Monte Carlo (MCMC) method was used for this purpose (with the assumption of an arbitrary pattern of missing data). Five datasets were created, with missing values replaced by imputed values based on models incorporating demographic, socioeconomic, psychosocial, anthropometric, and clinical variables. The results of these datasets were then combined using Rubin's rules [38]. A complete-case analysis was subsequently performed, in order to ensure the robustness of the multiple imputation results. For this analysis, frailty scores were estimated based on non-missing items only, using the same approach described above (“Frailty index”). Analyses were performed using SAS 9.2 (SAS Institute Inc, Cary, NC).
3. Results The most frequent deficits were non-insulin-dependent diabetes mellitus (43%), physical inactivity (36%), lack of energy (36%), work limitations (32%), limitation climbing stairs (28%) and self-rated health deterioration (25%). Rare deficits included Parkinson's (3%), Alzheimer's (2%), liver disease (2%), insulin-dependent diabetes (b1%) and weight loss (b 1%) (Appendix 1). By the end of follow-up, 10–13 years after index MI, 399 (34.7%) patients were defined as frail, having a frailty index score ≥0.25. Patients who developed frailty were older at baseline, more likely to be female and to suffer from hypertension and obesity, and to have had a more severe MI (as defined by Killip class) than those who did not develop frailty, as well as being less likely to have been admitted to the ICU, to have undergone revascularization or to have been prescribed β-blockers or aspirin on discharge (Table 1). Patients defined as frail at follow-up further exhibited a poorer socioeconomic profile, with fewer years of education, lower family income, lower neighborhood SES and more unemployment at baseline (all p b 0.01). Subjects who died during follow-up were older at baseline than survivors and more likely to have hypertension, obesity, Killip class N 1 and to present with comorbidities. They also displayed a poorer socioeconomic profile. Logistic regression provided OR for the likelihood of developing frailty post-MI. In the basic model, factors associated with increased risk of frailty development were age, female gender, smoking, obesity and poor self-rated health (Table 2). With the addition of SES measures in the extended model, the same factors remained significant predictors,
340
V. Myers et al. / International Journal of Cardiology 170 (2014) 338–343
Table 1 Baseline characteristics by outcome category 10–13 years post MI. Characteristics
n Age, mean (SD), y Male Clinical variables Hypertension Current smoking Hypercholesterolemia Obesity Q-wave MI Killip N1 Comorbidity Self-rated health Poor Average Good Management and pharmacotherapy Admission to ICU Early revascularization β-blockers ACE inhibitors Aspirin Socioeconomic variables Education, mean (SD), y Family income Below average Average Above average Pre-MI employment Neighborhood SES Lower tertile Middle tertile Upper tertile
Overall
Frailty status at end of follow-up
Table 2 Logistic regression models for frailty development 10–13 years after first MI. Died during follow-up
Not developed
Developed
1481 54 (8) 82
752 52 (8) 88
399 55 (8) 75
330 56 (7)⁎ 75⁎
38 53 36 18 75 21 38
31 51 37 14 73 13 28
40 53 38 24 76 21 38
50⁎ 56 33 21⁎ 76 38⁎ 62⁎
18 23 60
8 20 72
21 25 55
26 39
79 24 36 21 82
83 29 42 16 86
77 23 35 22 81
71⁎ 12⁎ 23⁎ 32⁎ 75⁎
11.1 (4.2)
12.2 (3.8)
10.4 (4.1)
9.6 (4.1)⁎
44 32 24 76
31 35 35 89
53 35 12 72
63⁎ 23 14 53⁎
30 39 31
23 38 39
34 42 24
42⁎ 36 22
Data are presented as percentage unless otherwise specified. ⁎ p ≤ 0.01.
except for smoking. Income, education and neighborhood SES were all significantly associated with frailty development — patients with below average income had more than twice the odds of becoming frail compared with those declaring above average income. Reclassification tables comparing the basic and extended models are presented in Table 3. In total, 24% (n = 273) of events (developing frailty) and non-events (not developing frailty) were appropriately reclassified and 13% (n = 147) were inappropriately reclassified with the extended model. Addition of multivariable SES measures yielded substantial gains in predictive value, improving the c-statistic from 0.70 in the basic model to 0.76 in the extended model (p b 0.001). IDI was 6.4% (p b 0.001) and NRI was 21.4% (p b 0.001); 11.8% for nonevents only and 9.6% for events only. Fig. 1 illustrates the improvement in area under the ROC curve by the extended model incorporating SES measures. A series of sensitivity analyses were conducted. In a complete-case analysis (calculating the frailty score based on available items only), a similar or slightly stronger association was observed between SES measures and frailty; OR = 2.54 (95% CI: 1.66–3.90) for below average vs. above average income; OR = 0.93 (95% CI: 0.90–0.97) per additional year of education; and OR = 1.66 (95% CI: 1.12–2.44) for lower vs. upper neighborhood SES tertile. A linear regression was additionally conducted, avoiding dichotomization of the frailty index, which supported the results of the logistic regression (p b 0.001 for income, education and neighborhood SES in the adjusted model). Subsequently, an analysis was carried out in a subsample with available ejection fraction data (638 participants; 203 outcome events). The multivariableadjusted OR (95% CI) for frailty, accounting also for baseline ejection fraction, was 2.03 (1.18–3.49) for below average vs. above average
Basic model
Age, 1 year Female Hypertension Current smoking Hypercholesterolemia Obesity Q-wave MI Killip N1 Comorbidity Self-rated health Poor Average Good Admission to ICU Early revascularization β-blockers ACE inhibitors Aspirin Education, 1 year Family income Below average Average Above average Pre-MI employment Neighborhood SES Lower tertile Middle tertile Upper tertile
Extended model
OR
95% CI
OR
95% CI
1.05 2.24 1.08 1.47 0.92 1.98 1.25 1.19 1.17
1.03–1.07 1.53–3.28 0.80–1.47 1.10–1.96 0.69–1.21 1.41–2.79 0.91–1.73 0.75–1.92 0.80–1.71
1.04 1.67 1.26 1.26 0.92 2.05 1.27 1.22 1.12
1.01–1.06 1.10–2.55 0.92–1.74 0.93–1.71 0.68–1.23 1.43–2.93 0.91–1.78 0.74–2.00 0.75–1.66
2.85 1.38 1 (ref.) 0.74 0.78 0.88 1.29 0.92
1.90–4.29 0.99–1.93 – 0.52–1.04 0.58–1.07 0.65–1.19 0.91–1.84 0.64–1.32
2.63 1.36 1 (ref.) 0.84 0.95 0.97 1.15 0.90 0.93
1.73–4.00 0.96–1.93 – 0.58–1.20 0.68–1.32 0.71–1.34 0.79–1.67 0.61–1.32 0.89–0.97
2.29 1.94 1 (ref.) 0.68
1.41–3.73 1.26–2.99 – 0.46–1.03
1.60 1.48 1 (ref.)
1.03–2.49 1.02–2.15 –
Basic model: age, gender, hypertension, current smoking, hypercholesterolemia, obesity, Q-wave MI, Killip N1, comorbidity, self-rated health, admission to ICU, early revascularization, β-blockers, ACE inhibitors and aspirin. Extended model: basic model plus SES measures (family income, education, pre-MI employment, neighborhood SES).
income, 0.88 (0.83–0.94) per additional year of education, and 1.40 (0.88–2.25) for lower vs. upper neighborhood SES tertile. The adjusted OR (95% CI) for frailty associated with ejection fraction were 1.53 (0.96–2.44) for 35–49% and 2.11 (0.91–4.89) for b 35% compared with the ≥50% group (p for trend = 0.03). Finally, applying the Diehr method [34,35], which considers as frail patients who had died before frailty assessment, the OR of developing frailty were similar at 1.89 (95% CI: 1.22–2.91) for below average vs. above average income; 0.93 (95% CI: 0.90–0.97) per additional year of education; and 1.56 (95% CI: 1.08– 2.24) for lower vs. upper neighborhood SES tertile. Altogether, these sensitivity analyses yielded results similar to the main analyses, thereby attesting to their robustness. Table 3 Model performance statistics: Cross tabulation of predicted risk categories by basic and extended models. Extended model
Reclassified as higher risk
Reclassified as lower risk
188 388 169 745
46 59 NA 105
NA 142 51 193
39 148 208 395
22 58 NA 80
NA 12 30 42
Basic model
b20%
20–40%
N40%
Total
Non-events b20% 20–40% N40% Total
142 142 9 293
46 187 42 275
0 59 118 177
17 12 3 32
22 78 27 127
0 58 178 236
Events b20% 20–40% N40% Total
NA, not applicable. Basic model: age, gender, hypertension, current smoking, hypercholesterolemia, obesity, Q-wave MI, Killip N1, comorbidity, self-rated health, admission to ICU, early revascularization, β-blockers, ACE inhibitors and aspirin. Extended model: basic model plus SES measures (family income, education, pre-MI employment, neighborhood SES).
V. Myers et al. / International Journal of Cardiology 170 (2014) 338–343
Fig. 1. ROC curves for extended versus basic models. “N” (new) represents the extended model's curve (solid red line) and “O” (old) represents the basic model's curve (dotted blue line).
4. Discussion Among first MI patients, low SES was disproportionately represented among those who developed frailty during 10–13 years of follow-up. Multilevel SES, including both individual and neighborhood measures, substantially improved risk prediction for frailty development, providing gains in both sensitivity and specificity. To date, previous studies reported an association between SES and frailty only in community samples [16,17], and without including neighborhood SES. The current study demonstrates the strong relationship between frailty and SES in MI survivors and further highlights the predictive importance of SES in risk assessment. Beyond individual SES, neighborhood SES – found to be associated with post-MI survival in previous research [15,39] – was also predictive of frailty development. This supports the findings from the English Longitudinal Study of Aging, which reported neighborhood deprivation to be associated with increased frailty index score [40], and extends these findings to MI patients. On a wider scale, socioeconomic differences between countries have been suggested to be responsible for disparities in frailty prevalence [41]. In general, disease severity at study entry was positively associated with the development of frailty, as would be expected. However, admission to the intensive care unit was inversely related to frailty. This may be indicative of differences in access to care, or may suggest that those surviving the ICU were the most robust, in fact, patients who died during follow-up had the lowest rate of admission to the ICU. Frail patients were also less likely to have undergone surgical procedures or to have been prescribed certain medications. Socioeconomic differences in treatment have been previously reported. Indeed low-SES MI patients were less likely to be admitted to the ICU in one study [13]. Low SES is a plausible risk factor for post-MI frailty since socioeconomic factors have shown a consistent relationship with adverse health outcomes and poorer prognosis. To cite just one of numerous examples, in the FINMONICA study, low-income men had more than twice the rate of pre-hospital coronary death compared to high-income men, and in those surviving MI, the 12 month mortality rate was significantly higher in low income patients [12]. These discrepancies were in part due to low income patients presenting with greater delay. SES exerts its influence on health outcomes in numerous ways, from reduced access to care in poorer neighborhoods (e.g. [42]) to low health literacy in disadvantaged communities (e.g. [43]), and increased prevalence of risk factors such as smoking, inactivity and unhealthy diet (e.g. [44]). Although the mechanisms underlying frailty have yet to be clearly defined, previous
341
research has demonstrated associations with certain mechanisms which are also associated with SES. For example, inflammatory markers such as IL-6 and fibrinogen are raised in lower socioeconomic groups [45] as well as being associated with frailty prevalence [46]. Furthermore, female gender is a consistent predictor of frailty [4,47,48], and women often have lower SES than men. The current study features several strengths, particularly its longitudinal nature, where previous studies used cross-sectional designs [16,17]. While causality cannot be claimed from an observational study, measurement of SES several years prior to assessment of frailty does indicate the direction of the relationship, suggesting that SES may affect future frailty status. The frailty index was developed based on a similar index which was previously demonstrated to predict mortality and hospitalizations in the same post-MI cohort [8]. Although frailty indices differ in the number and type of items included, a number of studies found consistent results, citing similar rate of accumulation of deficits and sub-maximal limit [16,22,23,49,50]. The index comprises a broad range of variables, representing multiple systems, and the majority of which are readily available from medical records. This multisystem assessment fulfills the criteria proposed for frailty assessment [1,23]. While many studies used a phenotypic assessment of frailty, some research has shown the accumulation of deficits method to be a more accurate predictor of mortality [51]. Previous research has demonstrated these two types of frailty measure to be highly correlated in cardiac patients [52]. The rate of frailty detected in the current study is comparable to that reported in other studies of patients with cardiovascular disease [20]. Furthermore, the use of the IDI and NRI measures of model performance provides additional support for the prognostic significance of SES in the development of frailty. Limitations of the study include the relatively young sample in comparison to much frailty research which focuses on over 65s. This discrepancy may have led to a lower frailty yield, although cardiac patients are generally frailer than community patients. Selection bias may have resulted from deaths occurring prior to the 10–13 year follow-up assessment, since these cases were most likely the frailest participants; however this was addressed using the Diehr method. Some important prognostic indicators were not included in the analysis due to missing data, for example ejection fraction was not routinely measured at study entry, with approximately 45% missing. However, our sensitivity analysis showed similar results in a subsample with available ejection fraction data. Additionally, some items included in the frailty index had some missing values, necessitating the use of multiple imputations. Finally, the historical age of the cohort, recruited 20 years ago, while allowing prospective investigation, reduces the generalizability of results, since more modern MI cohorts are defined on newer criteria including cardiac biomarkers. 5. Implications Despite an abundance of theories, it remains uncertain exactly how SES affects post-MI outcome. This study identifies frailty as a potential intermediate factor between SES and long-term prognosis after MI. The key prognostic role of frailty and its trajectories in mortality risk and hospital admissions post-MI has been demonstrated in this cohort in a recent publication [8]. The findings presented herein, demonstrating a substantial increase in frailty risk among lower-SES survivors of an MI, may thus have important clinical and public health implications. Frailty prevention initiatives after MI need to be considered, particularly among high-risk groups such as the elderly, women, and low-SES individuals. Participation in cardiac rehabilitation – which significantly improves prognosis [53] – should be strongly encouraged in high-risk individuals, and services provided in disadvantaged areas in order to address socioeconomic inequalities. In the study cohort, less than 20% of patients reported having attended, and non-attendees were less educated and more likely to be female. Furthermore, deficits such as functional limitations and locomotor disturbances, psychosocial vulnerability factors, and cognitive disorders, which are frequently overlooked in the
342
V. Myers et al. / International Journal of Cardiology 170 (2014) 338–343
traditional disease-specific model of cardiac care applied to MI patients [54], should be assessed and treated [55].
6. Conclusion This study demonstrated significant socioeconomic differences in the incidence of post-MI frailty. Individual and neighborhood-level SES assessed at the time of MI improved the accuracy of predicting which patients would develop frailty 13 years later. Ability to predict development of frailty at the time of MI would allow preventive intervention and forecasting of healthcare use.
Funding This work was supported by the Environment and Health Fund, Israel [Grant Award No. SGA 1204]. The funding source played no role in the design, execution, data analysis and interpretation or writing of the study.
Acknowledgments The following investigators and institutions took part in the Israel Study Group on First Acute Myocardial Infarction: Yaacov Drory, MD, Principal Investigator, Department of Rehabilitation, Sackler Medical School, Tel Aviv University, Tel Aviv; Yeheskiel Kishon, MD, Michael Kriwisky, MD, and Yoseph Rosenman, MD, Wolfson Medical Center, Holon; Uri Goldbourt, PhD, Hanoch Hod, MD, Eliezer Kaplinsky, MD, and Michael Eldar, MD, Sheba Medical Center, Tel Hashomer; Itzhak Shapira, MD, Amos Pines, MD, Margalit Drory, MSW, Arie Roth, MD, Shlomo Laniado, MD, and Gad Keren, MD, Tel-Aviv Sourasky Medical Center, Tel-Aviv; Daniel David, MD, Morton Leibowitz, MD, and Hana Pausner, MD, Meir Medical Center, Kfar Sava; Zvi Schlesinger, MD, and Zvi Vered, MD, Assaf Harofeh Medical Center, Zerifin; Alexander Battler, MD, Alejandro Solodky, MD, and Samuel Sclarovsky, MD, Beilinson Medical Center, Petach Tikvah; Izhar Zehavi, MD, and Rachel MaromKlibansky, MD, Hasharon Medical Center, Petah Tikvah; and Ron Leor, MD, Laniado Medical Center, Netanya. The authors are also indebted to Zalman Kaufman, MSc, for assistance with the Geographic Information System analysis.
Appendix 1. Frailty index variables assessed 10–13 years post-MI No.
Variable
Grading
% subjects scoring 1
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Self-rated health Self-rated health deterioration Intensive exercise limitations Stairs limitations Walk limitations Housework limitations Activities of daily living limitations Work limitations Social activities limitations Physical function (NYHA) Insulin dependent diabetes mellitus Non-insulin-dependent diabetes mellitus Peripheral vascular disease Stroke Parkinson COPD Cancer Kidney disease Arthritis Eye disease Liver disease Ulcer Carotid disease Digestion disease Psychological/mental disease Thyroid gland disease Blood disease Gall bladder disease Chronic disability Obesity Leisure time physical activity Weight loss Gets sick easily Self-reported functional fitness Tension/anxiety Alzheimer's Sadness/depression Loneliness Energetic Physical pain
Poor/not so good = 1; fair = 0.5; good/excellent = 0 Worse: reduced ≥2 points = 1; worse: reduced 1 point = 0.5; same/better = 0 Major limitation = 1; minor limitation = 0.5; no limitation = 0 Major limitation = 1; minor limitation = 0.5; no limitation = 0 Major limitation = 1; minor limitation = 0.5; no limitation = 0 Major limitation = 1; minor limitation = 0.5; no limitation = 0 Major limitation = 1; minor limitation = 0.5; no limitation = 0 Major limitation = 1; minor limitation = 0.5; no limitation = 0 Major limitation = 1; minor limitation = 0.5; no limitation = 0 Major limitation = 1; minor limitation = 0.5; no limitation = 0 Present = 1; absent = 0 Present = 1; absent = 0 Present = 1; absent = 0 Present = 1; absent = 0 Present = 1; absent = 0 Present = 1; absent = 0 Present = 1; absent = 0 Present = 1; absent = 0 Present = 1; absent = 0 Present = 1; absent = 0 Present = 1; absent = 0 Present = 1; absent = 0 Present = 1; absent = 0 Present = 1; absent = 0 Present = 1; absent = 0 Present = 1; absent = 0 Present = 1; absent = 0 Present = 1; absent = 0 Present = 1; absent = 0 BMI ≥ 30 = 1; BMI b 30 = 0 Inactive = 1; irregularly active = 0.5; regularly active = 0 N20% reduction = 1; 10%–20% reduction = 0.5; b10% loss = 0 Agree = 1; somewhat = 0.5; disagree = 0 1 = poor; 0.5 = average; 0 = great 1 = all the time; 0.5 = sometimes; 0 = never Present = 1; absent = 0 1 = all the time; 0.5 = sometimes; 0 = never 1 = all the time; 0.5 = sometimes; 0 = never 1 = never; 0.5 = sometimes; 0 = always 1 = very bad; 0.5 = some; 0 = not at all
23 25 16 28 10 23 3 32 16 10 b1 43 15 14 3 14 8 20 12 19 2 18 6 5 16 9 5 13 4 22 36 b1 19 23 10 2 9 9 36 13
References [1] Ferrucci L, Guralnik JM, Studenski S, Fried LP, Cutler Jr GB, Walston JD. Designing randomized, controlled trials aimed at preventing or delaying functional decline and disability in frail, older persons: a consensus report. J Am Geriatr Soc 2004;52:625–34.
[2] Song X, Mitnitski A, Rockwood K. Prevalence and 10-year outcomes of frailty in older adults in relation to deficit accumulation. J Am Geriatr Soc 2010;58:681–7. [3] Garcia-Gonzalez JJ, Garcia-Pena C, Franco-Marina F, Gutierrez-Robledo LM. A frailty index to predict the mortality risk in a population of senior Mexican adults. BMC Geriatr 2009;9:47.
V. Myers et al. / International Journal of Cardiology 170 (2014) 338–343 [4] Rockwood K, Howlett SE, MacKnight C, et al. Prevalence, attributes, and outcomes of fitness and frailty in community-dwelling older adults: report from the Canadian study of health and aging. J Gerontol A Biol Sci Med Sci 2004;59:1310–7. [5] Freiheit EA, Hogan DB, Eliasziw M, et al. Development of a frailty index for patients with coronary artery disease. J Am Geriatr Soc 2010;58:1526–31. [6] Ekerstad N, Swahn E, Janzon M, et al. Frailty is independently associated with shortterm outcomes for elderly patients with non-ST-segment elevation myocardial infarction. Circulation 2011;124:2397–404. [7] Singh M, Alexander K, Roger VL, et al. Frailty and its potential relevance to cardiovascular care. Mayo Clin Proc 2008;83:1146–53. [8] Myers V, Drory Y, Gerber Y. Clinical relevance of frailty trajectory post-myocardial infarction. Eur J Prev Cardiol 2012 [in press]. [9] Rockwood K, Mitnitski A, Song X, Steen B, Skoog I. Long-term risks of death and institutionalization of elderly people in relation to deficit accumulation at age 70. J Am Geriatr Soc 2006;54:975–9. [10] Jones D, Song X, Mitnitski A, Rockwood K. Evaluation of a frailty index based on a comprehensive geriatric assessment in a population based study of elderly Canadians. Aging Clin Exp Res 2005;17:465–71. [11] Mackenbach JP, Cavelaars AE, Kunst AE, Groenhof F. Socioeconomic inequalities in cardiovascular disease mortality: An international study. Eur Heart J 2000;21:1141–51. [12] Salomaa V, Miettinen H, Niemela M, et al. Relation of socioeconomic position to the case fatality, prognosis and treatment of myocardial infarction events: the FINMONICA MI register study. J Epidemiol Community Health 2001;55:475–82. [13] Gerber Y, Goldbourt U, Drory Y. Interaction between income and education in predicting long-term survival after acute myocardial infarction. Eur J Cardiovasc Prev Rehabil 2008;15:526–32. [14] Alter DA, Chong A, Austin PC, et al. Socioeconomic status and mortality after acute myocardial infarction. Ann Intern Med 2006;144:82–93. [15] Gerber Y, Benyamini Y, Goldbourt U, Drory Y. Neighbourhood socioeconomic context and long-term survival after myocardial infarction. Circulation 2010;121:375–83. [16] Woo J, Goggins W, Sham A, Ho SC. Social determinants of frailty. Gerontology 2005;51:402–8. [17] Szanton SL, Seplaki CL, Thorpe Jr RJ, Allen JK, Fried LP. Socioeconomic status is associated with frailty: the Women's Health and Aging Studies. J Epidemiol Community Health 2010;64:63–7. [18] Roger VL, Jacobsen SJ, Weston SA, et al. Trends in the incidence and survival of patients with hospitalized myocardial infarction, Olmsted County, Minnesota, 1979 to 1994. Ann Intern Med 2002;136:341–8. [19] 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;5:532–40. [20] Afilalo J. Frailty in patients with cardiovascular disease: why, when, and how to measure. Curr Cardiovasc Risk Rep 2011;5:467–72. [21] Rockwood K, Song X, Mitnitski A. Changes in relative fitness and frailty across the adult lifespan: evidence from the Canadian National Population Health Survey. CMAJ 2011;183:E487–94. [22] Mitnitski AB, Mogilner AJ, Rockwood K. Accumulation of deficits as a proxy measure of aging. Sci World J 2001;1:323–36. [23] Searle SD, Mitnitski A, Gahbauer EA, Gill TM, Rockwood K. A standard procedure for creating a frailty index. BMC Geriatr 2008;8:24. [24] Rockwood K, Andrew M, Mitnitski A. A comparison of two approaches to measuring frailty in elderly people. J Gerontol A Biol Sci Med Sci 2007;62:738–43. [25] Lewin-Epstein N, Sagiv-Schifter T, Shabtai EL, Shmueli A. Validation of the 36-item short-form Health Survey (Hebrew version) in the adult population of Israel. Med Care 1998;36:1361–70. [26] Cruz LN, Camey SA, Fleck MP, Polanczyk CA. World Health Organization quality of life instrument-brief and Short Form-36 in patients with coronary artery disease: do they measure similar quality of life concepts? Psychol Health Med 2009;14:619–28. [27] Gerber Y, Myers V, Goldbourt U, Benyamini Y, Scheinowitz M, Drory Y. Long-term trajectory of leisure time physical activity and survival after first myocardial infarction: a population-based cohort study. Eur J Epidemiol 2011;26:109–16. [28] Gerber Y, Koren-Morag N, Myers V, Benyamini Y, Goldbourt U, Drory Y. Long-term predictors of smoking cessation in a cohort of myocardial infarction survivors: a longitudinal study. Eur J Cardiovasc Prev Rehabil 2011;18:533–41. [29] Pate RR, Pratt M, Blair SN, et al. Physical activity and public health. A recommendation from the Centers for Disease Control and Prevention and the American College of Sports Medicine. JAMA 1995;273:402–7.
343
[30] Florian V, Drory Y. The Mental Health Inventory (MHI): psychometric properties and normative data in the Israeli population. Psychol Israel J Psychol 1990;2:26–35. [31] Molshatzki N, Drory Y, Myers V, et al. Role of socioeconomic status measures in long-term mortality risk prediction after myocardial infarction. Med Care 2011;49:673–8. [32] Burck L, Feinstein Y. Characterization and classification of geographical units by the socioeconomic level of the population. Series of 1995 census of population and housing publications. Jerusalem: Israel Central Bureau of Statistics; 2000. [33] Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987;40:373–83. [34] Diehr P, Patrick DL, Spertus JA, Kiefe CI, McDonell M, Fihn SD. Transforming selfrated health and the SF-36 scales to include death and improve interpretability. Med Care 2001;39:670–80. [35] Diehr P, Patrick DL. Trajectories of health for older adults over time: accounting fully for death. Ann Intern Med 2003;139:416–20. [36] Pencina MJ, D'Agostino Sr RB, D'Agostino Jr RB, Vasan RS. Evaluating the added predictive ability of a new marker: from area under the ROC curve to reclassification and beyond. Stat Med 2008;27:157–72 [discussion 207–12]. [37] Hanley JA, McNeil BJ. A method of comparing the areas under receiver operating characteristic curves derived from the same cases. Radiology 1983:148. [38] Rubin D. Multiple imputations for nonresponse in surveys. New York: John Wiley & Sons, Inc.; 1987. [39] Tonne C, Schwartz J, Mittelman M, Melly S, Suh H, Goldberg R. Long-term survival after acute myocardial infarction is lower in more deprived neighbourhoods. Circulation 2005;111:3063–70. [40] Lang IA, Hubbard RE, Andrew MK, Llewellyn DJ, Melzer D, Rockwood K. Neighborhood deprivation, individual socioeconomic status, and frailty in older adults. J Am Geriatr Soc 2009;57:1776–80. [41] Santos-Eggimann B, Cuenoud P, Spagnoli J, Junod J. Prevalence of frailty in middleaged and older community-dwelling Europeans living in 10 countries. J Gerontol A Biol Sci Med Sci 2009;64:675–81. [42] Alter DA, Naylor CD, Austin P, Tu JV. Effects of socioeconomic status on access to invasive cardiac procedures and on mortality after acute myocardial infarction. N Engl J Med 1999;341:1359–67. [43] Bostock S, Steptoe A. Association between low functional health literacy and mortality in older adults: longitudinal cohort study. BMJ 2012;344:e1602. [44] Hotchkiss JW, Davies C, Gray L, Bromley C, Capewell S, Leyland AH. Trends in adult cardiovascular disease risk factors and their socio-economic patterning in the Scottish population 1995–2008: cross-sectional surveys. BMJ Open 2011;1:e000176. [45] Wilson TW, Kaplan GA, Kauhanen J, et al. Association between plasma fibrinogen concentration and five socioeconomic indices in the Kuopio Ischemic Heart Disease Risk Factor Study. Am J Epidemiol 1993;137:292–300. [46] Leng SX, Xue QL, Tian J, Walston JD, Fried LP. Inflammation and frailty in older women. J Am Geriatr Soc 2007;55:864–71. [47] Puts MT, Lips P, Deeg DJ. Sex differences in the risk of frailty for mortality independent of disability and chronic diseases. J Am Geriatr Soc 2005;53:40–7. [48] Alvarado BE, Zunzunegui MV, Beland F, Bamvita JM. Life course social and health conditions linked to frailty in Latin American older men and women. J Gerontol A Biol Sci Med Sci 2008;63:1399–406. [49] Mitnitski A, Song X, Skoog I, et al. Relative fitness and frailty of elderly men and women in developed countries and their relationship with mortality. J Am Geriatr Soc 2005;53:2184–9. [50] Goggins WB, Woo J, Sham A, Ho SC. Frailty index as a measure of biological age in a Chinese population. J Gerontol A Biol Sci Med Sci 2005;60:1046–51. [51] Kulminski AM, Ukraintseva SV, Kulminskaya IV, Arbeev KG, Land K, Yashin AI. Cumulative deficits better characterize susceptibility to death in elderly people than phenotypic frailty: lessons from the Cardiovascular Health Study. J Am Geriatr Soc 2008;56:898–903. [52] Mcnallan SM, Chamberlain AM, Gerber Y, et al. Measuring frailty in heart failure: a community perspective. Am Heart J 2013;166(4):768–74. [53] Lawler PR, Filion KB, Eisenberg MJ. Efficacy of exercise-based cardiac rehabilitation post-myocardial infarction: a systematic review and meta-analysis of randomized controlled trials. Am Heart J 2011;162:571–84. [54] Tinetti ME, Bogardus Jr ST, Agostini JV. Potential pitfalls of disease-specific guidelines for patients with multiple conditions. N Engl J Med 2004;351:2870–4. [55] Michel JP, Newton JL, Kirkwood TB. Medical challenges of improving the quality of a longer life. JAMA 2008;299:688–90.
Contents lists available at ScienceDirect
International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard
Multilevel socioeconomic status and incidence of frailty post myocardial infarction☆,☆☆,★ Vicki Myers a, Yaacov Drory b, Uri Goldbourt a, Yariv Gerber a,⁎ a b
Department of Epidemiology and Preventive Medicine, School of Public Health, Sackler Faculty of Medicine, Tel Aviv University, Israel Department of Rehabilitation, Sackler Medical School, Tel Aviv University, Israel
a r t i c l e
i n f o
Article history: Received 7 February 2013 Received in revised form 2 September 2013 Accepted 2 November 2013 Available online 12 November 2013 Keywords: Myocardial infarction Frailty Socioeconomic status Risk discrimination Neighborhood socioeconomic status
a b s t r a c t Background: Frailty predicts mortality and hospitalizations in post-myocardial infarction (MI) patients. Socioeconomic status (SES) demonstrates a clear relationship with post-MI outcomes and is also associated with community frailty; however this relationship has yet to be evaluated in post-MI patients. We investigated the predictive value of socioeconomic factors in the development of post-MI frailty. Methods: A cohort of 1151 post-MI patients was followed up from initial hospitalization in 1992–1993 for 10– 13 years. Individual and neighborhood SES measures were assessed at baseline and frailty was assessed during follow-up via an index of deficit accumulation. Logistic regression models and discrimination indices enabled determination of the predictive value of socioeconomic factors over basic clinical variables in classifying frailty risk. Results: During follow-up, 399 patients (35%) developed frailty. Individual and neighborhood SES were significantly and independently associated with the risk of developing frailty. Low income patients had more than twice the risk of becoming frail compared with those with high income [odds ratio (OR), 2.29, 95% CI 1.41– 3.73]; while being in the lower vs. upper neighborhood SES tertile was associated with a 60% increased odds (OR, 1.60, 95% CI 1.03–2.49). Inclusion of multilevel SES yielded substantial gains in c-statistic (0.70 to 0.76), net reclassification improvement (21.4%) and integrated discrimination improvement (6.4%) over basic clinical factors (all p b 0.001), indicating increased predictive value and gains in sensitivity and specificity. Conclusions: Individual and neighborhood socioeconomic factors influence the development of frailty post-MI, and contribute to risk discrimination in this population. © 2013 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Frailty is a syndrome of increasing vulnerability and decreasing resistance to stressors resulting from a generalized decline in multiple physiological systems [1]. This emerging construct has been shown to predict both mortality [2–4] and disability [5]. Frailty has been demonstrated as clinically relevant in post-myocardial infarction (MI) patients, predicting mortality and healthcare use [6–8]. Despite variations in assessment techniques, frailty is consistently associated with poorer outcomes and greater mortality risk, as well as being an indicator of institutionalization and healthcare use [6,9,10]. However, the mechanisms underlying this phenomenon have yet to be elucidated. Socioeconomic differences in both individual and neighborhood domains are well-known to affect health outcomes and specifically ☆ Institution at which research performed: Tel Aviv University. ☆☆ See Acknowledgments for a list of participating medical centers and investigators. ★ Funding: This research was supported by Grant Award No. SGA 1204 from the Environment and Health Fund, Israel. ⁎ Corresponding author at: Department of Epidemiology and Preventive Medicine, School of Public Health, Sackler Faculty of Medicine, Tel Aviv University, Ramat Aviv, Tel Aviv 69978, Israel. Tel.: +972 3 6409867; fax: +972 3 6409868. E-mail address: [email protected] (Y. Gerber). 0167-5273/$ – see front matter © 2013 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ijcard.2013.11.009
cardiovascular prognosis, with income and occupation inversely related to cardiovascular mortality [11] and post-MI outcomes [12–15]. Socioeconomic status (SES) has further been associated with frailty in cross-sectional analyses. In an elderly Chinese sample, higher frailty score was associated with lower educational level and insufficient finances [16]. In the Women's Health and Aging Study, low-income women had twice the odds of frailty compared to high-income women, independent of age, and low education was associated with a threefold increased risk of frailty [17]. Beyond general population samples, the influence of SES on the development of frailty in MI survivors has not been investigated. Since post-MI patients constitute a highrisk population [18,19], with corresponding high levels of frailty [20], it is important to consider the antecedents of frailty in this group. This study investigated the predictive value of multilevel SES in the incidence of frailty in a cohort of post-MI patients followed up for 13 years. 2. Methods 2.1. Sample A cohort of first MI patients aged ≤65 years was followed up from initial hospitalization in 1992–1993 in one of the eight hospitals in central Israel, through 2002–2005, in the Israel Study of First Acute Myocardial Infarction. Of initial 1626 consecutive patients
V. Myers et al. / International Journal of Cardiology 170 (2014) 338–343 admitted, 81 died during initial hospitalization, a further 330 died before the end of follow-up 10–13 years later, and 24 were lost to follow-up. Of the remaining 1191 patients, 40 (3.4%) were classed as frail at baseline and excluded, leaving 1151 patients in the final sample. Clinical, sociodemographic and psychosocial data were collected from medical records and structured interviews approximately one week after the index hospitalization, and subsequently 10 to 13 years after MI. All aspects of the study were approved by the appropriate Institutional Ethics Committees. 2.2. Frailty index The Rockwood index of accumulation of deficits [21,22] was adapted to develop a frailty index comprising 40 items (Appendix 1), and patients were assessed 10–13 years after MI. A similar index was previously validated in the same cohort [8]. Variables included perceived health, comorbidity, functional limitations, weight loss, physical activity and psychological factors. Dichotomous items were coded as 0 if the deficit was absent and 1 if it was present; while ordinal or continuous variables were graded into a score between 0 and 1 (0 representing no impairment, 0.5 for minor impairment and 1 for major impairment). Scores were then summed and divided by the total number of variables to give a frailty index between 0 and 1, with 1 representing the greatest frailty. According to Rockwood et al., weighting of items is unnecessary since major disease usually goes hand in hand with other comorbidities, and the frailty index focuses on the accumulation of multiple problems, rather than their nature [23]. A threshold of ≥0.25 was used to define frailty, based on values used in previous research [2,24], a figure achieved by comparing a phenotypic definition of frailty with a frailty index, and modeling density distributions of deficits [24]. Patients were classified into two categories, defined as frail if they scored ≥0.25, and non-frail if they scored b0.25. 2.3. Frailty index variables 2.3.1. Perceived health Self-rated health was assessed on a 5-point scale from excellent to poor. Patients rated their own health currently and compared to 5 years previously. 2.3.2. Functional limitations Participants rated their functional fitness; the extent to which their health limited certain activities, including activities of daily living, housework and social activities; pain; energy; and how often they get sick, as part of the Hebrew version of the Medical Outcomes Study Short-Form 36 (SF-36). The reliability, validity, and responsiveness of the SF-36 are well documented both in general population and in patients with coronary artery disease [25,26]. 2.3.3. Comorbidity The presence of comorbid conditions, including diabetes, stroke, chronic obstructive pulmonary disorder (COPD), cancer, arthritis, Alzheimer's disease, Parkinson's disease, kidney disease, liver disease, ulcer and chronic disability, was determined via medical records and personal interviews. 2.3.4. Clinical variables Obesity was defined as body mass index ≥30 kg/m2. Current weight and weight loss since baseline were recorded. Leisure time physical activity was assessed by self-report including average frequency and duration of walking, cycling, swimming, gardening, going to the gym and team sports. These responses were summarized by a senior cardiologist into 3 groups: inactive, irregularly active and regularly active patients. Further details on physical activity were previously published [27,28]. Regular physical activity was defined as at least three 30 minute sessions per week, according to published guidelines [29]. 2.3.5. Psychological variables Patients completed the Mental Health Inventory, which assesses emotional distress and well-being, including questions on feelings of depression, anxiety, and loneliness. This questionnaire has shown validity and reliability in the Hebrew version [30]. 2.4. Socioeconomic status (SES) data Individual SES data were self-reported at study entry and included the following measures: family income relative to the national average (categorized as below average, average or above average), education (years of schooling), and pre-MI employment status (full-time or part-time vs. none) [13,31]. Neighborhood SES was estimated through an index developed by the Israel Central Bureau of Statistics, which summarizes socioeconomic measures from the 1995 National Census data, allowing the classification of small geographic units into SES categories, on a 20-point scale [32]. Tertiles of the scale distribution were used for analysis. 2.5. Additional variables Cardiovascular risk factors, MI characteristics and disease severity indices were recorded at the index hospitalization. Cigarette use was classified into current smoking vs. never or past smoking. Hypertension and hypercholesterolemia were defined according to standard criteria based on clinical and laboratory data. Comorbidity was defined based on the Charlson index as ≥1 point, indicating at least moderate comorbidity [33]. MI characteristics and severity indicators included infarct type
339
(Q wave or non-Q wave), Killip class N1, admission to intensive care unit (ICU), early revascularization, and pharmacotherapy use [β-blockers, angiotensin converting enzyme (ACE) inhibitors, and aspirin]. 2.6. Statistical analysis Differences in baseline characteristics by outcome status at the end of follow-up (not frail, frail, and deceased) were assessed with analysis of variance (ANOVA) for continuous variables and chi-square tests for categorical variables. Multivariable logistic regression models were constructed to assess the odds ratios (OR) and 95% confidence intervals (CI) for frailty associated with different sets of covariates. The demographic plus clinical model (“basic model”) included age, gender, hypertension, smoking, hypercholesterolemia, obesity, Q-wave MI, Killip class, comorbidity, self-rated health, admission to ICU and early revascularization, as well as β-blockers, ACE inhibitors, and aspirin on discharge. This model was repeated with the addition of SES variables (“extended model”), including education, family income, pre-MI employment, and neighborhood SES. These potential predictors were chosen based on previous literature and data availability. In a sensitivity analysis, left ventricular ejection fraction (available in 55% of the cohort) was additionally adjusted for in this subgroup. A linear regression was additionally performed with frailty score modeled as a continuous variable. Furthermore, in order to address selection bias introduced by exclusion of the 330 patients who had died before follow-up, so that their frailty score could not be evaluated, the Diehr method was applied [34,35]. Predicted probabilities of frailty at the end of follow-up were calculated for each model and categorized as less than 20%, 20% to 40%, and greater than 40%, corresponding to low, intermediate, and high risk, respectively. The incremental discriminatory value of the extended model over the basic model was examined by several methods. The integrated discrimination improvement (IDI) and net reclassification improvement (NRI) indices proposed by Pencina and colleagues [36] were calculated. In general, IDI measures the improvement in average sensitivity associated with the more advanced model, and subtracts any decrease in average specificity. NRI, in contrast, measures the extent to which individuals with and without the event of interest are appropriately reclassified into clinically accepted higher or lower risk categories with the extended versus basic model [31]. In addition, receiveroperating characteristic (ROC) curves were plotted for the basic and extended models. The c-statistic, a measure of area under the ROC curve, was calculated for each model and compared by methods previously described [37]. Missing values for family income (17%) and neighborhood SES (7%), as well as for variables comprising the frailty index, which ranged from 0 to 16% (mean, 5%), were imputed using multiple imputation methodology. The Markov Chain Monte Carlo (MCMC) method was used for this purpose (with the assumption of an arbitrary pattern of missing data). Five datasets were created, with missing values replaced by imputed values based on models incorporating demographic, socioeconomic, psychosocial, anthropometric, and clinical variables. The results of these datasets were then combined using Rubin's rules [38]. A complete-case analysis was subsequently performed, in order to ensure the robustness of the multiple imputation results. For this analysis, frailty scores were estimated based on non-missing items only, using the same approach described above (“Frailty index”). Analyses were performed using SAS 9.2 (SAS Institute Inc, Cary, NC).
3. Results The most frequent deficits were non-insulin-dependent diabetes mellitus (43%), physical inactivity (36%), lack of energy (36%), work limitations (32%), limitation climbing stairs (28%) and self-rated health deterioration (25%). Rare deficits included Parkinson's (3%), Alzheimer's (2%), liver disease (2%), insulin-dependent diabetes (b1%) and weight loss (b 1%) (Appendix 1). By the end of follow-up, 10–13 years after index MI, 399 (34.7%) patients were defined as frail, having a frailty index score ≥0.25. Patients who developed frailty were older at baseline, more likely to be female and to suffer from hypertension and obesity, and to have had a more severe MI (as defined by Killip class) than those who did not develop frailty, as well as being less likely to have been admitted to the ICU, to have undergone revascularization or to have been prescribed β-blockers or aspirin on discharge (Table 1). Patients defined as frail at follow-up further exhibited a poorer socioeconomic profile, with fewer years of education, lower family income, lower neighborhood SES and more unemployment at baseline (all p b 0.01). Subjects who died during follow-up were older at baseline than survivors and more likely to have hypertension, obesity, Killip class N 1 and to present with comorbidities. They also displayed a poorer socioeconomic profile. Logistic regression provided OR for the likelihood of developing frailty post-MI. In the basic model, factors associated with increased risk of frailty development were age, female gender, smoking, obesity and poor self-rated health (Table 2). With the addition of SES measures in the extended model, the same factors remained significant predictors,
340
V. Myers et al. / International Journal of Cardiology 170 (2014) 338–343
Table 1 Baseline characteristics by outcome category 10–13 years post MI. Characteristics
n Age, mean (SD), y Male Clinical variables Hypertension Current smoking Hypercholesterolemia Obesity Q-wave MI Killip N1 Comorbidity Self-rated health Poor Average Good Management and pharmacotherapy Admission to ICU Early revascularization β-blockers ACE inhibitors Aspirin Socioeconomic variables Education, mean (SD), y Family income Below average Average Above average Pre-MI employment Neighborhood SES Lower tertile Middle tertile Upper tertile
Overall
Frailty status at end of follow-up
Table 2 Logistic regression models for frailty development 10–13 years after first MI. Died during follow-up
Not developed
Developed
1481 54 (8) 82
752 52 (8) 88
399 55 (8) 75
330 56 (7)⁎ 75⁎
38 53 36 18 75 21 38
31 51 37 14 73 13 28
40 53 38 24 76 21 38
50⁎ 56 33 21⁎ 76 38⁎ 62⁎
18 23 60
8 20 72
21 25 55
26 39
79 24 36 21 82
83 29 42 16 86
77 23 35 22 81
71⁎ 12⁎ 23⁎ 32⁎ 75⁎
11.1 (4.2)
12.2 (3.8)
10.4 (4.1)
9.6 (4.1)⁎
44 32 24 76
31 35 35 89
53 35 12 72
63⁎ 23 14 53⁎
30 39 31
23 38 39
34 42 24
42⁎ 36 22
Data are presented as percentage unless otherwise specified. ⁎ p ≤ 0.01.
except for smoking. Income, education and neighborhood SES were all significantly associated with frailty development — patients with below average income had more than twice the odds of becoming frail compared with those declaring above average income. Reclassification tables comparing the basic and extended models are presented in Table 3. In total, 24% (n = 273) of events (developing frailty) and non-events (not developing frailty) were appropriately reclassified and 13% (n = 147) were inappropriately reclassified with the extended model. Addition of multivariable SES measures yielded substantial gains in predictive value, improving the c-statistic from 0.70 in the basic model to 0.76 in the extended model (p b 0.001). IDI was 6.4% (p b 0.001) and NRI was 21.4% (p b 0.001); 11.8% for nonevents only and 9.6% for events only. Fig. 1 illustrates the improvement in area under the ROC curve by the extended model incorporating SES measures. A series of sensitivity analyses were conducted. In a complete-case analysis (calculating the frailty score based on available items only), a similar or slightly stronger association was observed between SES measures and frailty; OR = 2.54 (95% CI: 1.66–3.90) for below average vs. above average income; OR = 0.93 (95% CI: 0.90–0.97) per additional year of education; and OR = 1.66 (95% CI: 1.12–2.44) for lower vs. upper neighborhood SES tertile. A linear regression was additionally conducted, avoiding dichotomization of the frailty index, which supported the results of the logistic regression (p b 0.001 for income, education and neighborhood SES in the adjusted model). Subsequently, an analysis was carried out in a subsample with available ejection fraction data (638 participants; 203 outcome events). The multivariableadjusted OR (95% CI) for frailty, accounting also for baseline ejection fraction, was 2.03 (1.18–3.49) for below average vs. above average
Basic model
Age, 1 year Female Hypertension Current smoking Hypercholesterolemia Obesity Q-wave MI Killip N1 Comorbidity Self-rated health Poor Average Good Admission to ICU Early revascularization β-blockers ACE inhibitors Aspirin Education, 1 year Family income Below average Average Above average Pre-MI employment Neighborhood SES Lower tertile Middle tertile Upper tertile
Extended model
OR
95% CI
OR
95% CI
1.05 2.24 1.08 1.47 0.92 1.98 1.25 1.19 1.17
1.03–1.07 1.53–3.28 0.80–1.47 1.10–1.96 0.69–1.21 1.41–2.79 0.91–1.73 0.75–1.92 0.80–1.71
1.04 1.67 1.26 1.26 0.92 2.05 1.27 1.22 1.12
1.01–1.06 1.10–2.55 0.92–1.74 0.93–1.71 0.68–1.23 1.43–2.93 0.91–1.78 0.74–2.00 0.75–1.66
2.85 1.38 1 (ref.) 0.74 0.78 0.88 1.29 0.92
1.90–4.29 0.99–1.93 – 0.52–1.04 0.58–1.07 0.65–1.19 0.91–1.84 0.64–1.32
2.63 1.36 1 (ref.) 0.84 0.95 0.97 1.15 0.90 0.93
1.73–4.00 0.96–1.93 – 0.58–1.20 0.68–1.32 0.71–1.34 0.79–1.67 0.61–1.32 0.89–0.97
2.29 1.94 1 (ref.) 0.68
1.41–3.73 1.26–2.99 – 0.46–1.03
1.60 1.48 1 (ref.)
1.03–2.49 1.02–2.15 –
Basic model: age, gender, hypertension, current smoking, hypercholesterolemia, obesity, Q-wave MI, Killip N1, comorbidity, self-rated health, admission to ICU, early revascularization, β-blockers, ACE inhibitors and aspirin. Extended model: basic model plus SES measures (family income, education, pre-MI employment, neighborhood SES).
income, 0.88 (0.83–0.94) per additional year of education, and 1.40 (0.88–2.25) for lower vs. upper neighborhood SES tertile. The adjusted OR (95% CI) for frailty associated with ejection fraction were 1.53 (0.96–2.44) for 35–49% and 2.11 (0.91–4.89) for b 35% compared with the ≥50% group (p for trend = 0.03). Finally, applying the Diehr method [34,35], which considers as frail patients who had died before frailty assessment, the OR of developing frailty were similar at 1.89 (95% CI: 1.22–2.91) for below average vs. above average income; 0.93 (95% CI: 0.90–0.97) per additional year of education; and 1.56 (95% CI: 1.08– 2.24) for lower vs. upper neighborhood SES tertile. Altogether, these sensitivity analyses yielded results similar to the main analyses, thereby attesting to their robustness. Table 3 Model performance statistics: Cross tabulation of predicted risk categories by basic and extended models. Extended model
Reclassified as higher risk
Reclassified as lower risk
188 388 169 745
46 59 NA 105
NA 142 51 193
39 148 208 395
22 58 NA 80
NA 12 30 42
Basic model
b20%
20–40%
N40%
Total
Non-events b20% 20–40% N40% Total
142 142 9 293
46 187 42 275
0 59 118 177
17 12 3 32
22 78 27 127
0 58 178 236
Events b20% 20–40% N40% Total
NA, not applicable. Basic model: age, gender, hypertension, current smoking, hypercholesterolemia, obesity, Q-wave MI, Killip N1, comorbidity, self-rated health, admission to ICU, early revascularization, β-blockers, ACE inhibitors and aspirin. Extended model: basic model plus SES measures (family income, education, pre-MI employment, neighborhood SES).
V. Myers et al. / International Journal of Cardiology 170 (2014) 338–343
Fig. 1. ROC curves for extended versus basic models. “N” (new) represents the extended model's curve (solid red line) and “O” (old) represents the basic model's curve (dotted blue line).
4. Discussion Among first MI patients, low SES was disproportionately represented among those who developed frailty during 10–13 years of follow-up. Multilevel SES, including both individual and neighborhood measures, substantially improved risk prediction for frailty development, providing gains in both sensitivity and specificity. To date, previous studies reported an association between SES and frailty only in community samples [16,17], and without including neighborhood SES. The current study demonstrates the strong relationship between frailty and SES in MI survivors and further highlights the predictive importance of SES in risk assessment. Beyond individual SES, neighborhood SES – found to be associated with post-MI survival in previous research [15,39] – was also predictive of frailty development. This supports the findings from the English Longitudinal Study of Aging, which reported neighborhood deprivation to be associated with increased frailty index score [40], and extends these findings to MI patients. On a wider scale, socioeconomic differences between countries have been suggested to be responsible for disparities in frailty prevalence [41]. In general, disease severity at study entry was positively associated with the development of frailty, as would be expected. However, admission to the intensive care unit was inversely related to frailty. This may be indicative of differences in access to care, or may suggest that those surviving the ICU were the most robust, in fact, patients who died during follow-up had the lowest rate of admission to the ICU. Frail patients were also less likely to have undergone surgical procedures or to have been prescribed certain medications. Socioeconomic differences in treatment have been previously reported. Indeed low-SES MI patients were less likely to be admitted to the ICU in one study [13]. Low SES is a plausible risk factor for post-MI frailty since socioeconomic factors have shown a consistent relationship with adverse health outcomes and poorer prognosis. To cite just one of numerous examples, in the FINMONICA study, low-income men had more than twice the rate of pre-hospital coronary death compared to high-income men, and in those surviving MI, the 12 month mortality rate was significantly higher in low income patients [12]. These discrepancies were in part due to low income patients presenting with greater delay. SES exerts its influence on health outcomes in numerous ways, from reduced access to care in poorer neighborhoods (e.g. [42]) to low health literacy in disadvantaged communities (e.g. [43]), and increased prevalence of risk factors such as smoking, inactivity and unhealthy diet (e.g. [44]). Although the mechanisms underlying frailty have yet to be clearly defined, previous
341
research has demonstrated associations with certain mechanisms which are also associated with SES. For example, inflammatory markers such as IL-6 and fibrinogen are raised in lower socioeconomic groups [45] as well as being associated with frailty prevalence [46]. Furthermore, female gender is a consistent predictor of frailty [4,47,48], and women often have lower SES than men. The current study features several strengths, particularly its longitudinal nature, where previous studies used cross-sectional designs [16,17]. While causality cannot be claimed from an observational study, measurement of SES several years prior to assessment of frailty does indicate the direction of the relationship, suggesting that SES may affect future frailty status. The frailty index was developed based on a similar index which was previously demonstrated to predict mortality and hospitalizations in the same post-MI cohort [8]. Although frailty indices differ in the number and type of items included, a number of studies found consistent results, citing similar rate of accumulation of deficits and sub-maximal limit [16,22,23,49,50]. The index comprises a broad range of variables, representing multiple systems, and the majority of which are readily available from medical records. This multisystem assessment fulfills the criteria proposed for frailty assessment [1,23]. While many studies used a phenotypic assessment of frailty, some research has shown the accumulation of deficits method to be a more accurate predictor of mortality [51]. Previous research has demonstrated these two types of frailty measure to be highly correlated in cardiac patients [52]. The rate of frailty detected in the current study is comparable to that reported in other studies of patients with cardiovascular disease [20]. Furthermore, the use of the IDI and NRI measures of model performance provides additional support for the prognostic significance of SES in the development of frailty. Limitations of the study include the relatively young sample in comparison to much frailty research which focuses on over 65s. This discrepancy may have led to a lower frailty yield, although cardiac patients are generally frailer than community patients. Selection bias may have resulted from deaths occurring prior to the 10–13 year follow-up assessment, since these cases were most likely the frailest participants; however this was addressed using the Diehr method. Some important prognostic indicators were not included in the analysis due to missing data, for example ejection fraction was not routinely measured at study entry, with approximately 45% missing. However, our sensitivity analysis showed similar results in a subsample with available ejection fraction data. Additionally, some items included in the frailty index had some missing values, necessitating the use of multiple imputations. Finally, the historical age of the cohort, recruited 20 years ago, while allowing prospective investigation, reduces the generalizability of results, since more modern MI cohorts are defined on newer criteria including cardiac biomarkers. 5. Implications Despite an abundance of theories, it remains uncertain exactly how SES affects post-MI outcome. This study identifies frailty as a potential intermediate factor between SES and long-term prognosis after MI. The key prognostic role of frailty and its trajectories in mortality risk and hospital admissions post-MI has been demonstrated in this cohort in a recent publication [8]. The findings presented herein, demonstrating a substantial increase in frailty risk among lower-SES survivors of an MI, may thus have important clinical and public health implications. Frailty prevention initiatives after MI need to be considered, particularly among high-risk groups such as the elderly, women, and low-SES individuals. Participation in cardiac rehabilitation – which significantly improves prognosis [53] – should be strongly encouraged in high-risk individuals, and services provided in disadvantaged areas in order to address socioeconomic inequalities. In the study cohort, less than 20% of patients reported having attended, and non-attendees were less educated and more likely to be female. Furthermore, deficits such as functional limitations and locomotor disturbances, psychosocial vulnerability factors, and cognitive disorders, which are frequently overlooked in the
342
V. Myers et al. / International Journal of Cardiology 170 (2014) 338–343
traditional disease-specific model of cardiac care applied to MI patients [54], should be assessed and treated [55].
6. Conclusion This study demonstrated significant socioeconomic differences in the incidence of post-MI frailty. Individual and neighborhood-level SES assessed at the time of MI improved the accuracy of predicting which patients would develop frailty 13 years later. Ability to predict development of frailty at the time of MI would allow preventive intervention and forecasting of healthcare use.
Funding This work was supported by the Environment and Health Fund, Israel [Grant Award No. SGA 1204]. The funding source played no role in the design, execution, data analysis and interpretation or writing of the study.
Acknowledgments The following investigators and institutions took part in the Israel Study Group on First Acute Myocardial Infarction: Yaacov Drory, MD, Principal Investigator, Department of Rehabilitation, Sackler Medical School, Tel Aviv University, Tel Aviv; Yeheskiel Kishon, MD, Michael Kriwisky, MD, and Yoseph Rosenman, MD, Wolfson Medical Center, Holon; Uri Goldbourt, PhD, Hanoch Hod, MD, Eliezer Kaplinsky, MD, and Michael Eldar, MD, Sheba Medical Center, Tel Hashomer; Itzhak Shapira, MD, Amos Pines, MD, Margalit Drory, MSW, Arie Roth, MD, Shlomo Laniado, MD, and Gad Keren, MD, Tel-Aviv Sourasky Medical Center, Tel-Aviv; Daniel David, MD, Morton Leibowitz, MD, and Hana Pausner, MD, Meir Medical Center, Kfar Sava; Zvi Schlesinger, MD, and Zvi Vered, MD, Assaf Harofeh Medical Center, Zerifin; Alexander Battler, MD, Alejandro Solodky, MD, and Samuel Sclarovsky, MD, Beilinson Medical Center, Petach Tikvah; Izhar Zehavi, MD, and Rachel MaromKlibansky, MD, Hasharon Medical Center, Petah Tikvah; and Ron Leor, MD, Laniado Medical Center, Netanya. The authors are also indebted to Zalman Kaufman, MSc, for assistance with the Geographic Information System analysis.
Appendix 1. Frailty index variables assessed 10–13 years post-MI No.
Variable
Grading
% subjects scoring 1
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Self-rated health Self-rated health deterioration Intensive exercise limitations Stairs limitations Walk limitations Housework limitations Activities of daily living limitations Work limitations Social activities limitations Physical function (NYHA) Insulin dependent diabetes mellitus Non-insulin-dependent diabetes mellitus Peripheral vascular disease Stroke Parkinson COPD Cancer Kidney disease Arthritis Eye disease Liver disease Ulcer Carotid disease Digestion disease Psychological/mental disease Thyroid gland disease Blood disease Gall bladder disease Chronic disability Obesity Leisure time physical activity Weight loss Gets sick easily Self-reported functional fitness Tension/anxiety Alzheimer's Sadness/depression Loneliness Energetic Physical pain
Poor/not so good = 1; fair = 0.5; good/excellent = 0 Worse: reduced ≥2 points = 1; worse: reduced 1 point = 0.5; same/better = 0 Major limitation = 1; minor limitation = 0.5; no limitation = 0 Major limitation = 1; minor limitation = 0.5; no limitation = 0 Major limitation = 1; minor limitation = 0.5; no limitation = 0 Major limitation = 1; minor limitation = 0.5; no limitation = 0 Major limitation = 1; minor limitation = 0.5; no limitation = 0 Major limitation = 1; minor limitation = 0.5; no limitation = 0 Major limitation = 1; minor limitation = 0.5; no limitation = 0 Major limitation = 1; minor limitation = 0.5; no limitation = 0 Present = 1; absent = 0 Present = 1; absent = 0 Present = 1; absent = 0 Present = 1; absent = 0 Present = 1; absent = 0 Present = 1; absent = 0 Present = 1; absent = 0 Present = 1; absent = 0 Present = 1; absent = 0 Present = 1; absent = 0 Present = 1; absent = 0 Present = 1; absent = 0 Present = 1; absent = 0 Present = 1; absent = 0 Present = 1; absent = 0 Present = 1; absent = 0 Present = 1; absent = 0 Present = 1; absent = 0 Present = 1; absent = 0 BMI ≥ 30 = 1; BMI b 30 = 0 Inactive = 1; irregularly active = 0.5; regularly active = 0 N20% reduction = 1; 10%–20% reduction = 0.5; b10% loss = 0 Agree = 1; somewhat = 0.5; disagree = 0 1 = poor; 0.5 = average; 0 = great 1 = all the time; 0.5 = sometimes; 0 = never Present = 1; absent = 0 1 = all the time; 0.5 = sometimes; 0 = never 1 = all the time; 0.5 = sometimes; 0 = never 1 = never; 0.5 = sometimes; 0 = always 1 = very bad; 0.5 = some; 0 = not at all
23 25 16 28 10 23 3 32 16 10 b1 43 15 14 3 14 8 20 12 19 2 18 6 5 16 9 5 13 4 22 36 b1 19 23 10 2 9 9 36 13
References [1] Ferrucci L, Guralnik JM, Studenski S, Fried LP, Cutler Jr GB, Walston JD. Designing randomized, controlled trials aimed at preventing or delaying functional decline and disability in frail, older persons: a consensus report. J Am Geriatr Soc 2004;52:625–34.
[2] Song X, Mitnitski A, Rockwood K. Prevalence and 10-year outcomes of frailty in older adults in relation to deficit accumulation. J Am Geriatr Soc 2010;58:681–7. [3] Garcia-Gonzalez JJ, Garcia-Pena C, Franco-Marina F, Gutierrez-Robledo LM. A frailty index to predict the mortality risk in a population of senior Mexican adults. BMC Geriatr 2009;9:47.
V. Myers et al. / International Journal of Cardiology 170 (2014) 338–343 [4] Rockwood K, Howlett SE, MacKnight C, et al. Prevalence, attributes, and outcomes of fitness and frailty in community-dwelling older adults: report from the Canadian study of health and aging. J Gerontol A Biol Sci Med Sci 2004;59:1310–7. [5] Freiheit EA, Hogan DB, Eliasziw M, et al. Development of a frailty index for patients with coronary artery disease. J Am Geriatr Soc 2010;58:1526–31. [6] Ekerstad N, Swahn E, Janzon M, et al. Frailty is independently associated with shortterm outcomes for elderly patients with non-ST-segment elevation myocardial infarction. Circulation 2011;124:2397–404. [7] Singh M, Alexander K, Roger VL, et al. Frailty and its potential relevance to cardiovascular care. Mayo Clin Proc 2008;83:1146–53. [8] Myers V, Drory Y, Gerber Y. Clinical relevance of frailty trajectory post-myocardial infarction. Eur J Prev Cardiol 2012 [in press]. [9] Rockwood K, Mitnitski A, Song X, Steen B, Skoog I. Long-term risks of death and institutionalization of elderly people in relation to deficit accumulation at age 70. J Am Geriatr Soc 2006;54:975–9. [10] Jones D, Song X, Mitnitski A, Rockwood K. Evaluation of a frailty index based on a comprehensive geriatric assessment in a population based study of elderly Canadians. Aging Clin Exp Res 2005;17:465–71. [11] Mackenbach JP, Cavelaars AE, Kunst AE, Groenhof F. Socioeconomic inequalities in cardiovascular disease mortality: An international study. Eur Heart J 2000;21:1141–51. [12] Salomaa V, Miettinen H, Niemela M, et al. Relation of socioeconomic position to the case fatality, prognosis and treatment of myocardial infarction events: the FINMONICA MI register study. J Epidemiol Community Health 2001;55:475–82. [13] Gerber Y, Goldbourt U, Drory Y. Interaction between income and education in predicting long-term survival after acute myocardial infarction. Eur J Cardiovasc Prev Rehabil 2008;15:526–32. [14] Alter DA, Chong A, Austin PC, et al. Socioeconomic status and mortality after acute myocardial infarction. Ann Intern Med 2006;144:82–93. [15] Gerber Y, Benyamini Y, Goldbourt U, Drory Y. Neighbourhood socioeconomic context and long-term survival after myocardial infarction. Circulation 2010;121:375–83. [16] Woo J, Goggins W, Sham A, Ho SC. Social determinants of frailty. Gerontology 2005;51:402–8. [17] Szanton SL, Seplaki CL, Thorpe Jr RJ, Allen JK, Fried LP. Socioeconomic status is associated with frailty: the Women's Health and Aging Studies. J Epidemiol Community Health 2010;64:63–7. [18] Roger VL, Jacobsen SJ, Weston SA, et al. Trends in the incidence and survival of patients with hospitalized myocardial infarction, Olmsted County, Minnesota, 1979 to 1994. Ann Intern Med 2002;136:341–8. [19] 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;5:532–40. [20] Afilalo J. Frailty in patients with cardiovascular disease: why, when, and how to measure. Curr Cardiovasc Risk Rep 2011;5:467–72. [21] Rockwood K, Song X, Mitnitski A. Changes in relative fitness and frailty across the adult lifespan: evidence from the Canadian National Population Health Survey. CMAJ 2011;183:E487–94. [22] Mitnitski AB, Mogilner AJ, Rockwood K. Accumulation of deficits as a proxy measure of aging. Sci World J 2001;1:323–36. [23] Searle SD, Mitnitski A, Gahbauer EA, Gill TM, Rockwood K. A standard procedure for creating a frailty index. BMC Geriatr 2008;8:24. [24] Rockwood K, Andrew M, Mitnitski A. A comparison of two approaches to measuring frailty in elderly people. J Gerontol A Biol Sci Med Sci 2007;62:738–43. [25] Lewin-Epstein N, Sagiv-Schifter T, Shabtai EL, Shmueli A. Validation of the 36-item short-form Health Survey (Hebrew version) in the adult population of Israel. Med Care 1998;36:1361–70. [26] Cruz LN, Camey SA, Fleck MP, Polanczyk CA. World Health Organization quality of life instrument-brief and Short Form-36 in patients with coronary artery disease: do they measure similar quality of life concepts? Psychol Health Med 2009;14:619–28. [27] Gerber Y, Myers V, Goldbourt U, Benyamini Y, Scheinowitz M, Drory Y. Long-term trajectory of leisure time physical activity and survival after first myocardial infarction: a population-based cohort study. Eur J Epidemiol 2011;26:109–16. [28] Gerber Y, Koren-Morag N, Myers V, Benyamini Y, Goldbourt U, Drory Y. Long-term predictors of smoking cessation in a cohort of myocardial infarction survivors: a longitudinal study. Eur J Cardiovasc Prev Rehabil 2011;18:533–41. [29] Pate RR, Pratt M, Blair SN, et al. Physical activity and public health. A recommendation from the Centers for Disease Control and Prevention and the American College of Sports Medicine. JAMA 1995;273:402–7.
343
[30] Florian V, Drory Y. The Mental Health Inventory (MHI): psychometric properties and normative data in the Israeli population. Psychol Israel J Psychol 1990;2:26–35. [31] Molshatzki N, Drory Y, Myers V, et al. Role of socioeconomic status measures in long-term mortality risk prediction after myocardial infarction. Med Care 2011;49:673–8. [32] Burck L, Feinstein Y. Characterization and classification of geographical units by the socioeconomic level of the population. Series of 1995 census of population and housing publications. Jerusalem: Israel Central Bureau of Statistics; 2000. [33] Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987;40:373–83. [34] Diehr P, Patrick DL, Spertus JA, Kiefe CI, McDonell M, Fihn SD. Transforming selfrated health and the SF-36 scales to include death and improve interpretability. Med Care 2001;39:670–80. [35] Diehr P, Patrick DL. Trajectories of health for older adults over time: accounting fully for death. Ann Intern Med 2003;139:416–20. [36] Pencina MJ, D'Agostino Sr RB, D'Agostino Jr RB, Vasan RS. Evaluating the added predictive ability of a new marker: from area under the ROC curve to reclassification and beyond. Stat Med 2008;27:157–72 [discussion 207–12]. [37] Hanley JA, McNeil BJ. A method of comparing the areas under receiver operating characteristic curves derived from the same cases. Radiology 1983:148. [38] Rubin D. Multiple imputations for nonresponse in surveys. New York: John Wiley & Sons, Inc.; 1987. [39] Tonne C, Schwartz J, Mittelman M, Melly S, Suh H, Goldberg R. Long-term survival after acute myocardial infarction is lower in more deprived neighbourhoods. Circulation 2005;111:3063–70. [40] Lang IA, Hubbard RE, Andrew MK, Llewellyn DJ, Melzer D, Rockwood K. Neighborhood deprivation, individual socioeconomic status, and frailty in older adults. J Am Geriatr Soc 2009;57:1776–80. [41] Santos-Eggimann B, Cuenoud P, Spagnoli J, Junod J. Prevalence of frailty in middleaged and older community-dwelling Europeans living in 10 countries. J Gerontol A Biol Sci Med Sci 2009;64:675–81. [42] Alter DA, Naylor CD, Austin P, Tu JV. Effects of socioeconomic status on access to invasive cardiac procedures and on mortality after acute myocardial infarction. N Engl J Med 1999;341:1359–67. [43] Bostock S, Steptoe A. Association between low functional health literacy and mortality in older adults: longitudinal cohort study. BMJ 2012;344:e1602. [44] Hotchkiss JW, Davies C, Gray L, Bromley C, Capewell S, Leyland AH. Trends in adult cardiovascular disease risk factors and their socio-economic patterning in the Scottish population 1995–2008: cross-sectional surveys. BMJ Open 2011;1:e000176. [45] Wilson TW, Kaplan GA, Kauhanen J, et al. Association between plasma fibrinogen concentration and five socioeconomic indices in the Kuopio Ischemic Heart Disease Risk Factor Study. Am J Epidemiol 1993;137:292–300. [46] Leng SX, Xue QL, Tian J, Walston JD, Fried LP. Inflammation and frailty in older women. J Am Geriatr Soc 2007;55:864–71. [47] Puts MT, Lips P, Deeg DJ. Sex differences in the risk of frailty for mortality independent of disability and chronic diseases. J Am Geriatr Soc 2005;53:40–7. [48] Alvarado BE, Zunzunegui MV, Beland F, Bamvita JM. Life course social and health conditions linked to frailty in Latin American older men and women. J Gerontol A Biol Sci Med Sci 2008;63:1399–406. [49] Mitnitski A, Song X, Skoog I, et al. Relative fitness and frailty of elderly men and women in developed countries and their relationship with mortality. J Am Geriatr Soc 2005;53:2184–9. [50] Goggins WB, Woo J, Sham A, Ho SC. Frailty index as a measure of biological age in a Chinese population. J Gerontol A Biol Sci Med Sci 2005;60:1046–51. [51] Kulminski AM, Ukraintseva SV, Kulminskaya IV, Arbeev KG, Land K, Yashin AI. Cumulative deficits better characterize susceptibility to death in elderly people than phenotypic frailty: lessons from the Cardiovascular Health Study. J Am Geriatr Soc 2008;56:898–903. [52] Mcnallan SM, Chamberlain AM, Gerber Y, et al. Measuring frailty in heart failure: a community perspective. Am Heart J 2013;166(4):768–74. [53] Lawler PR, Filion KB, Eisenberg MJ. Efficacy of exercise-based cardiac rehabilitation post-myocardial infarction: a systematic review and meta-analysis of randomized controlled trials. Am Heart J 2011;162:571–84. [54] Tinetti ME, Bogardus Jr ST, Agostini JV. Potential pitfalls of disease-specific guidelines for patients with multiple conditions. N Engl J Med 2004;351:2870–4. [55] Michel JP, Newton JL, Kirkwood TB. Medical challenges of improving the quality of a longer life. JAMA 2008;299:688–90.