Abstract
Background
Ticagrelor is recommended over clopidogrel in acute coronary syndrome based on the results of the PLATO (Study of Platelet Inhibition and Patient Outcomes) trial. We aimed to emulate PLATO in older adults with and without frailty and with acute coronary syndrome treated with percutaneous coronary intervention.
Methods and Results
We created a new‐user cohort of Medicare fee‐for‐service beneficiaries aged ≥65 years hospitalized for acute coronary syndrome from 2014 to 2018 and initiated ticagrelor or clopidogrel following percutaneous coronary intervention. Frailty was defined using a validated claims‐based frailty index ≥0.25. Coprimary outcomes were major adverse cardiovascular events and major bleeding. Follow‐up began on the date of first outpatient prescription for ticagrelor or clopidogrel and ended on the earliest date for an outcome event, death, discontinuation of the index drug, or disenrollment from Medicare. The study included 42 843 older adults; 23% were frail. After propensity score matching, the rates of major adverse cardiovascular events per 100 person‐years comparing ticagrelor versus clopidogrel groups were 7.8 and 7.3 in the frail cohort (hazard ratio [HR], 1.07 [95% CI, 0.84–1.36]) and 3.7 and 4.2 in the nonfrail cohort (HR, 0.87 [95% CI, 0.75–1.02]). The corresponding rates of major bleeding were 4.3 and 3.8 in the frail cohort (HR, 1.12 95% CI, [0.80–1.56]) and 2.2 and 1.8 in the nonfrail cohort (HR, 1.22 [95% CI, 0.98–1.51]).
Conclusions
There was a trend toward a modest reduction in risk of major adverse cardiovascular events and a trend toward a modest increase in risk of major bleeding with ticagrelor compared with clopidogrel in the nonfrail cohort. There was insufficient evidence for the benefit of ticagrelor in frail older adults.
Keywords: acute coronary syndrome, clopidogrel, frail, older adult, ticagrelor
Subject Categories: Cardiovascular Disease, Epidemiology, Aging
Nonstandard Abbreviations and Acronyms
- CFI
claims‐based frailty index
- MACE
major adverse cardiovascular event
- PLATO
Study of Platelet Inhibition and Patient Outcomes
- PS
propensity score
- RD
rate differences
Clinical Perspective.
What Is New?
There was a trend toward a modest reduction in risk of major adverse cardiovascular events and a trend toward a modest increase in risk of major bleeding with ticagrelor compared with clopidogrel in nonfrail patients.
The benefit of ticagrelor in nonfrail older adults was comparable to PLATO (Study of Platelet Inhibition and Patient Outcomes), and there was insufficient evidence for the benefit of ticagrelor in frail older adults.
What Are the Clinical Implications?
Frailty assessment may be important in tailoring antiplatelet therapy to older patients being managed for acute coronary syndrome.
Frailty affects 20% to 30% of patients treated for acute coronary syndrome and those undergoing percutaneous coronary intervention (PCI). 1 , 2 Frailty is associated with higher rates of major adverse cardiovascular events (MACE) as well as treatment‐related bleeding following acute coronary syndrome and PCI. 2 , 3 , 4 , 5 , 6 , 7 , 8 Although the practice guidelines recommend ticagrelor over clopidogrel in non‐ST‐segment–elevation myocardial infarction (NSTEMI) or ST‐segment–elevation myocardial infarction (STEMI) treated with PCI 9 , 10 based on the results of the PLATO (Study of Platelet Inhibition and Patient Outcomes) trial, 11 , 12 the trial did not examine variation in the efficacy and safety of ticagrelor by patients' frailty status. In fact, in the PLATO trial, ticagrelor increased the risk of noncoronary artery bypass graft‐related major bleeding compared with clopidogrel, 11 and the risk of bleeding may be increased further in older adults with frailty. Furthermore, compared with nonfrail patients, frail patients have greater risk of death and, therefore, may derive less benefit from ticagrelor over clopidogrel. The dilemma in choosing the optimal P2Y12 inhibitor in older adults with frailty is reflected in the drug use patterns. In our prior study of Medicare beneficiaries hospitalized with STEMI in 2018 to 2020, 58% were initiated on ticagrelor or prasugrel (mostly ticagrelor), and those with frailty were less likely to receive ticagrelor or prasugrel. 13 In the current study, we aimed to address the evidence gap by emulating the PLATO trial in Medicare beneficiaries with and without frailty who were hospitalized with NSTEMI and STEMI and subsequently treated with PCI.
Methods
Study Design and Study Cohort
Due to Centers for Medicare & Medicaid Services restrictions and the risk of breach of confidentiality, we cannot share Medicare data files generated in the project. Medicare data can be purchased from the Centers for Medicare & Medicaid Services through ResDAC. This study was approved by the institutional review board at Brigham and Women's Hospital, Boston, Massachusetts. A waiver of informed consent was obtained. We created a new‐user cohort of Medicare fee‐for‐service beneficiaries who were hospitalized with acute myocardial infarction (MI) between 2014 and 2018 and were initiated on ticagrelor or clopidogrel. P2Y12 inhibitor use was measured from Medicare Part D (prescription) file. We did not examine prasugrel because prasugrel is relatively contraindicated in patients aged ≥75 years, and the rate of initiation of prasugrel in older adults is very low. 13 Eligible patients were aged ≥65 years, hospitalized with a primary diagnosis of NSTEMI or STEMI and treated with PCI before hospital discharge, and filled a prescription for ticagrelor or clopidogrel within 60 days after the hospital discharge. We allowed 60 days because some patients might be discharged with up to a 60‐day supply of P2Y12 inhibitors based on our clinical experience (D.K. and A.P.). The index date is the date of first outpatient prescription for ticagrelor or clopidogrel. We required patients to have had continuous enrollment in Medicare Part A (inpatient), B (outpatient), and D (prescription) program for 1 year before the index date to allow a complete capture of comorbidities and drug use. We excluded patients with missing data on age, sex, or race; history of hemorrhagic stroke in the past 5 years; prior use of any P2Y12 inhibitor in the past 1 year; and treatment with an oral anticoagulant or history of major bleeding in the previous 90 days. Figures 1 and 2 show the study cohort design and cohort flow diagram, respectively.
Figure 1. Design of the study cohort.
MACE indicates major adverse cardiovascular event; NSTEMI, non‐ST‐segment–elevation myocardial infarction; PCI, percutaneous coronary intervention; and STEMI, ST‐segment–elevation myocardial infarction.
Figure 2. Cohort flow diagram.
NSTEMI indicates non‐ST‐segment–elevation myocardial infarction; PCI, percutaneous coronary intervention; and STEMI, ST‐segment–elevation myocardial infarction.
Measurement of Frailty and Baseline Covariates
The covariate assessment period was 365 days before the index date. Baseline covariates included age, sex, self‐reported race or ethnicity (Asian, Black, Hispanic, White, and Other), the chronic conditions defined by the Centers for Medicare & Medicaid Services Chronic Conditions Data Warehouse algorithms, 14 drug use, and health care use (inpatient days, skilled nursing facility days, and emergency department visits). The Gagne combined comorbidity index was used to assess comorbidity burden. 15 The Kim claims‐based frailty index (CFI), which estimates a deficit accumulation frailty index (range: 0–1) using 93 variables defined by diagnosis codes, health services codes, and durable medical equipment codes, was used to quantify frailty. 16 , 17 , 18 , 19 Patients with CFI ≥0.25 were classified as being frail. We assessed the length of the index MI hospitalization, the number of days from MI diagnosis to the initiation of a P2Y12 inhibitor, and the number of cardiologist visits, ECGs, echocardiograms, stress tests, and venipunctures in the past year.
Outcomes and Follow‐Up
The coprimary outcomes were (1) MACE, defined as a composite of recurrent MI and ischemic stroke, and (2) major bleeding. The secondary outcomes were all‐cause death and the composite of MACE, major bleeding, and all‐cause death. Major bleeding was captured using inpatient claims with critical site bleeding diagnosis codes in the primary position or inpatient claims with noncritical site bleeding diagnosis codes in the primary position and transfusion procedure or revenue codes during the same inpatient admission (positive predictive value: 89%–99%). 20 This approach has been used previously. 21 , 22 , 23
The follow‐up started on the index date and ended on the earliest date for the following events: occurrence of an outcome event, death, discontinuation of the index drug (defined as a prescription gap of ≥30 days), switching to a different P2Y12 inhibitor, or disenrollment from Medicare Part A, B, or D (Figure 1).
Statistical Analysis
The analyses were conducted for frail and nonfrail cohorts separately. Our primary analysis was on‐treatment analysis, in which patients were censored when they discontinued the drug. To adjust for imbalance in clinical characteristics between the ticagrelor and clopidogrel groups, we estimated a propensity score (PS) within each cohort using logistic regression that included the aforementioned covariates (see 1.2 Measurement of Frailty and Baseline Covariates) and calendar year to adjust for the time trends in drug use and outcomes. We conducted a 1:1 nearest‐neighbor PS matching with a caliper of 0.2 of the SD of the logit PS. Adequate matching was defined as having a standardized mean difference less than 0.10. We estimated the incident rates of the primary and secondary outcomes and calculated the hazard ratios (HR), rate differences (RD), and 95% CIs. In accordance with guidance from the American Statistical Association 24 and the International Committee of Medical Journal Editors, 25 we provide quantitative interpretation of HR and 95% CI. 26 We tested for heterogeneity in the HR between the frail and nonfrail groups by using the common effect inverse‐variance model. 27 As a sensitivity analysis, we conducted intention‐to‐treat analysis, in which patients were not censored when they discontinue or switch from the index drug, with a maximum follow‐up time of 1 year. In addition, we assessed how the HRs would differ under hypothetical scenarios involving varying levels of unmeasured confounding, and for nonnull exposure‐outcome associations, we calculated the E‐values, the minimum strength of exposure‐confounder and outcome‐confounder associations that would move the exposure‐outcome association to the null. 28 Analyses were conducted in the SAS version 9.4 and Stata version 17. A 2‐sided P value <0.05 was considered statistically significant.
Results
Characteristics of Study Cohorts
There were 2 study cohorts: the new users of ticagrelor or clopidogrel who were frail (n [%]=9762 [23]) and nonfrail (n [%]=33 081 [77]). Before PS matching, 28% and 48% of the frail and nonfrail cohorts were treated with ticagrelor, respectively (Table S1). In both the frail and nonfrail cohorts, the ticagrelor‐treated patients were younger, less likely to be dually eligible for Medicare and Medicaid, more likely to have presented with STEMI, more likely to have received a drug‐eluting stent, and had a lower mean combined comorbidity score and mean frailty index. The prevalence of cardiovascular and noncardiovascular comorbidities was generally lower in the ticagrelor‐treated patients.
After PS matching, there was an adequate balance of the baseline covariates between the ticagrelor‐ and clopidogrel‐treated patients in both the frail and nonfrail cohorts (Table 1). The frail cohort had a mean age of 78 years, 63% women, 29% dual eligibility, a mean combined comorbidity index score of 7, and a mean CFI of 0.29. The nonfrail cohort had a mean age of 76 years, 41% women, 15% dual eligibility, a mean combined comorbidity index score of 3, and a mean CFI of 0.19. The proportion of STEMI was similar (frail versus nonfrail cohort: 43% versus 44%), but the rate of drug‐eluting stent use was lower in the frail cohort (86% versus 91%).
Table 1.
Characteristics of 1:1 Propensity Score‐Matched Frail and Nonfrail Patients With Acute Myocardial Infarction Treated With Ticagrelor Versus Clopidogrel (2014–2018) Following Percutaneous Coronary Intervention
| Frail | Nonfrail | |||||
|---|---|---|---|---|---|---|
| Ticagrelor | Clopidogrel | SMD | Ticagrelor | Clopidogrel | SMD | |
| Sample size | 2670 | 2670 | 10 487 | 10 487 | ||
| Mean age ±SD, y | 78.3±7.4 | 78.4±7.6 | −0.009 | 75.5±6.7 | 75.5±7.0 | 0.001 |
| Women, n (%) | 1691 (63.3) | 1701 (63.7) | 0.008 | 4282 (40.8) | 4387 (41.8) | 0.021 |
| Race, n (%) | ||||||
| Black | 176 (6.6) | 183 (6.9) | 0.044 | 603 (5.8) | 636 (5.9) | 0.000 |
| Other | 146 (5.5) | 155 (5.8) | 0.044 | 623 (5.9) | 617 (5.9) | 0.000 |
| White | 2348 (87.9) | 2332 (87.3) | 0.044 | 9261 (88.3) | 9234 (88.1) | 0.000 |
| Dual eligibility | 768 (28.8) | 806 (30.2) | −0.031 | 1538 (14.7) | 1527 (14.6) | 0.000 |
| Mean combined comorbidity index±SD | 7.0±3.2 | 7.0±3.2 | −0.020 | 3.2±2.9 | 3.3±2.9 | −0.015 |
| Mean frailty index±SD | 0.29±0.04 | 0.29±0.04 | −0.006 | 0.19±0.03 | 0.19±0.03 | −0.004 |
| ST‐segment–elevation myocardial infarction at presentation, n (%) | 1130 (42.3) | 1156 (43.3) | −0.020 | 4704 (44.9) | 4591 (43.8) | 0.022 |
| Drug‐eluting stent, n (%) | 2311 (86.6) | 2295 (86.0) | 0.071 | 9559 (91.2) | 9541 (91.0) | 0.006 |
| Cardiovascular comorbidities, n (%) | ||||||
| Atrial fibrillation | 487 (18.2) | 477 (17.9) | 0.010 | 1035 (9.9) | 1057 (10.1) | −0.007 |
| Chronic kidney disease | 1914 (71.7) | 1937 (72.6) | −0.019 | 4390 (41.9) | 4461 (42.5) | −0.014 |
| Diabetes | 1757 (65.8) | 1754 (65.7) | 0.002 | 4788 (45.7) | 4681 (44.6) | 0.021 |
| Heart failure | 1930 (72.3) | 1941 (72.7) | −0.009 | 4277 (40.8) | 4222 (40.3) | 0.011 |
| Hyperlipidemia | 2578 (96.6) | 2574 (96.4) | 0.008 | 9696 (92.5) | 9684 (92.3) | 0.004 |
| Hypertension | 2638 (98.8) | 2641 (98.9) | −0.011 | 9631 (91.8) | 9636 (91.9) | −0.002 |
| Prior myocardial infarction | 326 (12.2) | 317 (11.9) | 0.010 | 744 (7.1) | 732 (7.0) | 0.004 |
| Stroke or transient ischemic attack | 863 (32.3) | 885 (33.2) | −0.018 | 1296 (12.4) | 1299 (12.4) | −0.001 |
| Noncardiovascular comorbidities, n (%) | ||||||
| Anemia | 2265 (84.8) | 2278 (85.3) | −0.004 | 5451 (52.0) | 5480 (52.3) | −0.006 |
| Cancer | 523 (19.6) | 516 (19.3) | 0.011 | 1658 (15.8) | 1727 (16.5) | −0.018 |
| Chronic obstructive pulmonary disease | 1473 (55.2) | 1470 (55.1) | 0.002 | 2977 (28.4) | 3015 (28.8) | −0.008 |
| Dementia | 1213 (45.4) | 1219 (45.7) | 0.005 | 727 (6.9) | 735 (7.0) | −0.003 |
| Depression | 1797 (67.3) | 1781 (66.7) | 0.015 | 2758 (26.3) | 2773 (26.4) | −0.003 |
| Hip or pelvic fracture | 210 (7.9) | 179 (6.7) | 0.014 | 205 (2.0) | 205 (2.0) | 0.000 |
| Hypothyroidism | 1102 (41.3) | 1133 (42.4) | 0.003 | 2607 (24.9) | 2606 (24.9) | 0.0002 |
| Osteoporosis | 898 (33.6) | 901 (33.8) | 0.015 | 1548 (14.8) | 1521 (14.5) | 0.007 |
| Rheumatoid or osteoarthritis | 2203 (82.5) | 2194 (82.2) | −0.003 | 5983 (57.1) | 5997 (57.2) | −0.003 |
| Medications, n (%) | ||||||
| Angiotensin‐converting enzyme inhibitors | 1246 (46.7) | 1276 (47.8) | −0.023 | 5269 (50.2) | 5282 (50.4) | −0.002 |
| Angiotensin receptor blockers | 811 (30.4) | 782 (29.3) | 0.024 | 2845 (27.1) | 2805 (26.8) | 0.009 |
| Antiarrhythmics | 327 (12.3) | 333 (12.5) | −0.007 | 658 (6.3) | 658 (6.3) | 0.000 |
| Beta‐blockers | 2398 (89.8) | 2371 (88.8) | 0.033 | 9434 (90.0) | 9425 (89.9) | 0.003 |
| CCB—dihydropyridine | 965 (36.1) | 982 (36.8) | −0.013 | 2922 (27.9) | 2932 (28.0) | −0.002 |
| CCB—nondihydropyridine | 181 (6.8) | 176 (6.6) | 0.008 | 528 (5.0) | 501 (4.8) | 0.012 |
| Dementia drugs | 417 (15.6) | 411 (15.4) | 0.006 | 155 (1.5) | 169 (1.6) | −0.011 |
| High‐ceiling diuretics | 1133 (42.4) | 1147 (43.0) | −0.011 | 2120 (20.2) | 2088 (19.9) | 0.008 |
| Insulin | 612 (22.9) | 614 (23.0) | −0.002 | 1072 (10.2) | 1076 (10.3) | −0.001 |
| Opioids | 1066 (39.9) | 1081 (40.5) | −0.011 | 2213 (21.1) | 2194 (20.9) | 0.004 |
| Statins | 2315 (86.7) | 2333 (87.4) | −0.020 | 9446 (90.1) | 9469 (90.3) | −0.007 |
| Thyroid hormone replacement | 742 (27.8) | 759 (28.4) | −0.014 | 1789 (17.1) | 1769 (16.9) | 0.005 |
| Mean health care use in previous 365 d±SD | ||||||
| Cardiology visit | 11.4±7.1 | 11.6±7.4 | −0.029 | 8.3±4.9 | 8.3±5.1 | −0.008 |
| Echocardiogram | 1.6±1.1 | 1.7±1.2 | −0.039 | 1.2±0.9 | 1.2±1.0 | −0.012 |
| ECG | 4.8±3.6 | 4.8±3.4 | −0.007 | 3.3±2.2 | 3.3±2.3 | −0.025 |
| Stress test, n (%) | 438 (16.4) | 410 (15.4) | 0.029 | 1507 (14.4) | 1516 (14.5) | −0.002 |
| Emergency department visits | 2.9 (2.6) | 2.8 (2.5) | 0.006 | 1.6 (1.1) | 1.6 (1.3) | −0.025 |
| Hospital days | 14.6 (13.8) | 15.0 (13.0) | −0.029 | 6.1 (6.8) | 6.3 (6.2) | −0.016 |
| Skilled nursing facility days | 6.0 (13.7) | 6.5 (17.6) | −0.031 | 0.5 (2.8) | 0.6 (2.9) | −0.027 |
CCB indicates calcium channel blocker; and SMD, standardized mean difference.
MACE by Frailty Levels
In the overall population, the mean±SD durations of follow‐up were 0.62 (0.64) patient‐years for the ticagrelor group and 1.06 (0.95) patient‐years for the clopidogrel group. The rate of MACE per 100 person‐years was 4.4 for the ticagrelor group and 4.8 for the clopidogrel group (HR, 0.80 [95% CI, 0.70–0.91]; RD, −0.34 [95% CI, −0.92 to 0.24]). These results indicate that ticagrelor was associated with a modest 20% reduction (the effect may range from 20% reduction to 9% reduction) in MACE compared with clopidogrel.
In the frail cohort, the mean±SD durations of follow‐up were 0.54 (0.61) patient‐years for the ticagrelor group and 0.93 (0.94) patient‐years for the clopidogrel group. In the nonfrail cohort, they were 0.64 (0.65) patient‐years for the ticagrelor group and 1.10 (0.95) for the clopidogrel group. The reasons for censoring and their frequencies are listed in Table S2. In the frail cohort, the rates of MACE per 100 person‐years were 7.8 for the ticagrelor group and 7.3 for the clopidogrel group in the frail cohort (HR, 1.07 [95% CI, 0.84–1.36]; RD, 0.51 [95% CI, −1.27 to 2.30]) (Table 2 and Figure 3). The results are consistent with both a modest benefit (16% reduction) to a modest harm (36% increase) in MACE associated with ticagrelor versus clopidogrel, precluding any definitive conclusion. In the nonfrail cohort, the corresponding rates were 3.7 and 4.2 per 100 person‐years, respectively (HR, 0.87 [95% CI, 0.75–1.02]; RD, −0.53 [95% CI, −1.12 to 0.07]). These results are consistent with modest (25% reduction) benefit to little (2% increase) harm of ticagrelor versus clopidogrel. There was no evidence for the heterogeneity of treatment effect by frailty levels (P value for heterogeneity=0.16).
Table 2.
Frailty and Association of Ticagrelor Versus Clopidogrel With Clinical Outcomes in Acute Coronary Syndrome Treated With Percutaneous Coronary Intervention
| Outcomes | Ticagrelor | Clopidogrel | HR (95%) | RD (95%) | ||||
|---|---|---|---|---|---|---|---|---|
| No. of events, n | Person‐years | Rate per 100 person‐years | No. of events, n | Person‐years | Rate per 100 person‐years | |||
| Major adverse cardiovascular events | ||||||||
| Total | 360 | 8131 | 4.4 | 668 | 140 009 | 4.8 | 0.80 (0.70 to 0.91) | −0.34 (−0.92 to 0.24) |
| Frail | 113 | 1451 | 7.8 | 181 | 2488 | 7.3 | 1.07 (0.84 to 1.36) | 0.51 (−1.27 to 2.30) |
| Nonfrail | 247 | 6680 | 3.7 | 487 | 11 521 | 4.2 | 0.87 (0.75 to 1.02) | −0.53 (−1.12 to 0.07) |
| P value for heterogeneity | … | … | … | … | … | … | 0.16 | 0.28 |
| Major bleeding | ||||||||
| Total | 211 | 8234 | 2.6 | 308 | 14 311 | 2.2 | 1.02 (0.86 to 1.22) | 0.41 (−0.01 to 0.83) |
| Frail | 63 | 1470 | 4.3 | 97 | 2543 | 3.8 | 1.12 (0.80 to 1.56) | 0.47 (−0.83 to 1.77) |
| Nonfrail | 148 | 6764 | 2.2 | 211 | 11 768 | 1.8 | 1.22 (0.98 to 1.51) | 0.40 (−0.03 to 0.82) |
| P value for heterogeneity | … | … | … | … | … | … | 0.67 | 0.92 |
| All‐cause death | ||||||||
| Total | 379 | 8295 | 4.6 | 699 | 14 467 | 4.8 | 0.84 (0.74 to 0.96) | −0.26 (−0.85 to 0.32) |
| Frail | 181 | 1488 | 12.2 | 288 | 2585 | 11.1 | 1.09 (0.90 to 1.32) | 1.03 (−1.17 to 3.22) |
| Nonfrail | 198 | 6807 | 2.9 | 411 | 11 882 | 3.5 | 0.84 (0.71 to 1.00) | −0.55 (−1.08 to −0.02) |
| P value for heterogeneity | … | … | … | … | … | … | 0.049 | <0.01 |
| Composite event | ||||||||
| Total | 876 | 8075 | 10.8 | 1519 | 13 867 | 11.0 | 0.86 (0.79 to 0.93) | −0.11 (−1.01 to 0.80) |
| Frail | 324 | 1434 | 22.6 | 504 | 2450 | 20.1 | 1.10 (0.95 to 1.26) | 2.02 (−1.03 to 5.06) |
| Nonfrail | 552 | 6641 | 8.3 | 1015 | 11 418 | 8.9 | 0.94 (0.84 to 1.04) | −0.58 (−1.46 to 0.31) |
| P value for heterogeneity | … | … | … | … | … | … | 0.052 | 0.03 |
HR indicates hazard ratio; MACE, major adverse cardiovascular event; and RD, rate difference.
Figure 3. Kaplan–Meier curves for on‐treatment analysis of the primary outcomes.
(A) MACE—frail; (B) Major bleeding—frail; (C) MACE—Nonfrail; (D) Major bleeding—nonfrail. MACE indicates major adverse cardiovascular event.
Major Bleeding by Frailty Levels
In the overall population, the mean±SD durations of follow‐up were 0.63 (0.65) patient‐years for the ticagrelor group and 1.09 (0.92) patient‐years for the clopidogrel group. The rates of major bleeding per 100 person‐years were 2.6 for the ticagrelor group and 2.2 for the clopidogrel group (HR, 1.02 [95% CI, 0.86–1.22]; RD 0.41 [95% CI, −0.01 to 0.83]). These results indicate that ticagrelor was not associated with major bleeding compared with clopidogrel (the effect may range from 14% reduction to 22% increase).
In the frail cohort, the mean±SD durations of follow‐up were 0.55 (0.61) patient‐years for the ticagrelor group and 0.95 (0.95) patient‐years for the clopidogrel group. In the nonfrail cohort, they were 0.65 (0.66) patient‐years for the ticagrelor group and 1.12 (0.96) for the clopidogrel group. In the frail cohort, the rates of major bleeding per 100 person‐years were 4.3 for the ticagrelor group and 3.8 for the clopidogrel group (HR, 1.12 [95% CI, 0.80–1.56]; RD, 0.47 [95% CI, −0.83 to 1.77]) (Table 2 and Figure 3). The results are consistent with both a modest benefit (20% reduction) and a large harm (56%). In the nonfrail cohort, the corresponding rates in the nonfrail cohort were 2.2 and 1.8, respectively (HR, 1.22 [95% CI, 0.98–1.51]; RD, 0.40 [95% CI, −0.03 to 0.82]). These results are consistent with little (2% reduction) to large harm (51% increase) of ticagrelor compared with clopidogrel. There was no evidence for the treatment effect heterogeneity by frailty levels (P value for heterogeneity=0.67).
All‐Cause Death by Frailty Levels
In the overall population, the mean±SD durations of follow‐up were 0.63 (0.65) patient‐years for the ticagrelor group and 1.10 (0.96) patient‐years for the clopidogrel group. The rate of death per 100 person‐years was 4.6 for the ticagrelor group and 4.8 for the clopidogrel group (HR, 0.84 [95% CI, 0.74–0.96]; RD −0.26 [95% CI, −0.85 to 0.32]). Ticagrelor (versus clopidogrel) was associated with a modest 16% reduction in the risk of death with 95% CI ranging from 0.74 (26% reduction) to 0.96 (4% reduction).
In the frail cohort, the mean±SD durations of follow‐up were 0.56 (0.62) patient‐years for the ticagrelor group and 0.97 (0.96) patient‐years for the clopidogrel group. In the nonfrail cohort, they were 0.65 (0.66) patient‐years for the ticagrelor group and 1.13 (0.96) for the clopidogrel group. In the frail cohort, the rates of all‐cause death per 100 person‐years were 12.2 for the ticagrelor group and 11.1 for the clopidogrel group (HR, 1.09 [95% CI, 0.90–1.32]; RD, 1.03 [95% CI, −1.17 to 3.22]) (Table 2 and Figure S1). These results are consistent with both a modest benefit (10% reduction) to a modest harm (32% increase) in death associated with ticagrelor versus clopidogrel. In the nonfrail cohort, the corresponding rates were 2.9 and 3.5 per 100 person‐years, respectively (HR, 0.84 [95% CI, 0.71–1.00]; RD, −0.55 [−1.08 to −0.02]). These results are consistent with a modest benefit (29% reduction) to no benefit. There was statistically significant heterogeneity of treatment effect (P value for heterogeneity=0.049).
Composite Event of MACE, Major Bleeding, and All‐Cause Death by Frailty Levels
In the overall population, the mean±SD durations of follow‐up were 0.61 (0.64) patient‐years for the ticagrelor group and 1.05 (0.94) patient‐years for the clopidogrel group. The rate of composite event per 100 person‐years was 10.8 for the ticagrelor group and 11.0 for the clopidogrel group (HR, 0.86 [95% CI, 0.79–0.93]; RD −0.11 [95% CI, −1.01 to 0.80]). These results indicate that ticagrelor was associated with a modest 14% reduction in the risk of composite event (the effect may range from 21% reduction to 7% reduction).
In the frail cohort, the mean±SD durations of follow‐up were 0.54 (0.61) patient‐years for the ticagrelor group and 0.92 (0.94) patient‐years for the clopidogrel group. In the nonfrail cohort, they were 0.63 (0.65) patient‐years for the ticagrelor group and 1.09 (0.94) for the clopidogrel group. In the frail cohort, the rates of the composite event per 100 person‐years were 22.6 for the ticagrelor group and 20.1 for the clopidogrel group (HR, 1.10 [95% CI, 0.95–1.26]; RD, 2.02 [95% CI, −1.03 to 5.06]) (Table 2 and Figure S1). These results are consistent with little benefit (5% reduction) to a modest harm (26% increase) in the composite event associated with ticagrelor versus clopidogrel. In the nonfrail cohort, the corresponding rates in the nonfrail cohort were 8.3 and 8.9 per 100 person‐years, respectively (HR, 0.94 [95% CI, 0.84–1.04]; RD, −0.58 [95% CI, −1.46 to 0.31]). These results are consistent with modest benefit (16% reduction) to little harm (4% increase). There was insufficient evidence to determine whether the effect of ticagrelor versus clopidogrel was different by frailty levels (P value for heterogeneity=0.052).
Sensitivity Analysis
The HR estimates from the intention‐to‐treat analyses were qualitatively consistent to the results from the primary analyses (Table S3). Our results were robust in most scenarios of unmeasured confounding, except the scenario in which there is a large imbalance in a strong risk factor (Table S4). The E‐value was 1.67 for the nonnull association of all‐cause death for ticagrelor versus clopidogrel in the nonfrail cohort (HR, 0.84). This means that the observed HR of 0.84 between all‐cause death (outcome) and ticagrelor versus clopidogrel (exposure) can be explained away by an unmeasured confounder that is associated with the outcome with HR 1.67 and is associated with the exposure with HR 1.67.
Discussion
In the current study, we emulated the PLATO trial comparing ticagrelor versus clopidogrel in a Medicare population with PCI‐treated NSTEMI and STEMI stratified by frailty status. There was a trend toward a modest reduction in risk of MACE (HR, 0.87 [95% CI, 0.75–1.02]) and a trend toward a modest increase in risk of major bleeding (HR, 1.22 [95% CI, 0.98–1.51]) with ticagrelor in the nonfrail cohort. In the frail cohort, there was insufficient evidence to conclude the difference in the rates of MACE and major bleeding between ticagrelor and clopidogrel. Ticagrelor was associated with a reduced risk of all‐cause death in the nonfrail cohort (HR, 0.84, [95% CI, 0.71–1.00]) but not in the frail cohort. PLATO reported the following HR estimates: 0.84 (95% CI, 0.77–0.92) for MACE (a composite outcome of death from vascular cause, MI, or stroke), 0.78 (95% CI, 0.69–0.89) for death from any cause, 1.04 (95% CI, 0.95–1.13) for major bleeding, and 1.19 (95% CI, 1.02–1.38) for noncoronary artery bypass graft related major bleeding. 11 Our results for MACE in the nonfrail cohort were consistent with the estimates for MACE in PLATO and fell within the trial's 95% CIs. Our results for major bleeding, which did not include bleeding events during the index hospitalization due to our study design, were consistent with the trial results for noncoronary artery bypass graft related major bleeding. In addition, consistent with PLATO, we found mortality benefit with ticagrelor in the nonfrail cohort.
Our results from the frail cohort were different from the results for the nonfrail cohort and from the results of PLATO. One possible explanation of this discrepancy is competing risk due to death. Our frail cohort had a much higher rate of death compared with that in the trial (eg, 12.2 per 100 person years in the ticagrelor‐treated group in our study versus 4.5% in 12 months of follow‐up the ticagrelor arm in the trial). Despite their higher rate of MACE, frail patients may derive little benefit from ticagrelor over clopidogrel because of their high risk of death.
PLATO did not show significant interaction by age for MACE and major bleeding. A randomized controlled trial of ticagrelor or prasugrel versus clopidogrel in adults aged ≥70 years with NSTEMI showed lower risk of major bleeding with clopidogrel. 29 There has not been any study to date investigating the efficacy and safety of ticagrelor versus clopidogrel in frail older adults. PLATO did not measure frailty status at baseline and prior comparative effectiveness studies did not include frailty. 30 , 31 Our study demonstrates the importance of geriatric pharmacoepidemiology with focus on frailty in complementing trial evidence. We were able to confirm the benefit of ticagrelor in nonfrail older adults, who constituted 77% of the PCI‐treated NSTEMI or STEMI population. Despite the higher mean age (75 years) compared with the trial participants (median 62 years), the nonfrail patients in our study had lower 1‐year mortality than the trial participants. The remaining 23% of the PCI‐treated NSTEMI or STEMI population were frail, and the benefit of ticagrelor over clopidogrel was unclear in the frail population. Our study suggests that frailty, a marker of physiologic vulnerability, may be much more important than chronological age in determination of potential benefit from ticagrelor. Furthermore, our study suggests that treatment for older adults could be individualized by frailty status.
In‐person frailty assessment is not routinely performed in cardiovascular trials because it can be time consuming. A few cardiovascular trials that have published subgroup analyses by frailty status created various frailty indices ad hoc using the patient characteristics available within the trials. 32 , 33 , 34 These ad hoc frailty measures have not been validated nor used in clinical practice. In contrast, our CFI has been validated against clinical frailty measures, physical performance, disability, and health outcomes. 16 , 18 , 35 , 36 One way to use a validated frailty index within a randomized controlled trial is to link the trial to an external data set from which the validated frailty index can be measured. For example, in a study by Faridi et al., the investigators successfully linked a subset of the DAPT (Dual Antiplatelet Therapy) study participants who were aged 65 years or older and enrolled in fee‐for‐service Medicare to their claims data. 8 They used the validated Hopkins Claims‐Based Frailty Indicator to measure frailty within the linked Medicare data. 37 Although this approach benefits from randomization of study participants, as shown in the Faridi study, it can lack power to study outcomes in frail patients, who are often excluded from clinical trials.
Although our study suggests that frailty may provide a meaningful heterogeneity in treatment effect for ticagrelor versus clopidogrel in older adults with PCI‐treated acute coronary syndrome, there are several important limitations of our study. First, although stent thrombosis occurs early after PCI and prevention of stent thrombosis is one of the major benefits of ticagrelor over clopidogrel, only the patients who survived up to the index date (ie, the date of the first prescription fill up to 60 days after hospital discharge) were included in our new‐user analysis. Similarly, our study did not include MI, stroke, or major bleeding that occurred before the index date. Second, we did not have information on the angiography, blood biomarkers, or echocardiography during the index hospitalization. Third, we did not include unstable angina because validity of claims‐based definition of unstable angina is unreliable. Fourth, our study does not specify cardiovascular death because one cannot ascertain cause‐specific death with reasonable specificity using claims data. Fifth, our study using claims data could not ascertain aspirin use. However, dual antiplatelet therapy with aspirin 81 mg after PCI is the standard of care in the United States, and therefore, we made the reasonable assumption that all patients received aspirin 81 mg daily during follow‐up. Sixth, we adjusted for baseline comorbidities and prescription drug use, but unmeasured confounding remains possible, and we did not adjust for time‐varying confounders (eg, initiation of oral anticoagulation during follow‐up).
Conclusions
In conclusion, despite being older than the trial participants in PLATO, nonfrail older adults with STEMI or NSTEMI treated with PCI appeared to derive a similar modest benefit from taking ticagrelor compared with clopidogrel in regard to MACE and all‐cause death, with a trend toward a modestly increased rate of major bleeding. Despite having higher rates of MACE compared with nonfrail older adults, there was insufficient evidence for the benefit of ticagrelor over clopidogrel in frail older adults. Our results should be confirmed in a future randomized controlled trial targeting older adults with frailty.
Sources of Funding
This study was funded by grants R01AG062713 and K24AG073527 from the National Institute on Aging to D.H.K. and grant K23HL151903‐01A1 from the National Heart, Lung, and Blood Institute to D.K.
Disclosures
Dr Ko reports consulting fee from Windrose Consulting Group and an investigator‐sponsored research grant from Boston Scientific Corporation outside the submitted work. Dr Kim reports personal fees from Alosa Health and VillagesMD outside the submitted work. The remaining authors have no disclosures to report.
Supporting information
Tables S1–S4
Figure S1
This article was sent to Saket Girotra, MD, SM, Associate Editor, for review by expert referees, editorial decision, and final disposition.
Supplemental Material is available at https://www.ahajournals.org/doi/suppl/10.1161/JAHA.124.034529
For Sources of Funding and Disclosures, see page 9.
References
- 1. Damluji AA, Huang J, Bandeen‐Roche K, Forman DE, Gerstenblith G, Moscucci M, Resar JR, Varadhan R, Walston JD, Segal JB. Frailty among older adults with acute myocardial infarction and outcomes from percutaneous coronary interventions. J Am Heart Assoc. 2019;8:e013686. doi: 10.1161/JAHA.119.013686 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Singh M, Rihal CS, Lennon RJ, Spertus JA, Nair KS, Roger VL. Influence of frailty and health status on outcomes in patients with coronary disease undergoing percutaneous revascularization. Circ Cardiovasc Qual Outcomes. 2011;4:496–502. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Dodson JA, Hochman JS, Roe MT, Chen AY, Chaudhry SI, Katz S, Zhong H, Radford MJ, Udell JA, Bagai A, et al. The association of frailty with in‐hospital bleeding among older adults with acute myocardial infarction: insights from the ACTION Registry. JACC Cardiovasc Interv. 2018;11:2287–2296. doi: 10.1016/j.jcin.2018.08.028 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Alonso Salinas GL, Sanmartín Fernández M, Pascual Izco M, Marco Del Castillo Á, Rincón Díaz LM, Lozano Granero C, Valverde Gómez M, Pastor Pueyo P, Del Val Martín D, Pardo Sanz A, et al. Frailty predicts major bleeding within 30 days in elderly patients with acute coronary syndrome. Int J Cardiol. 2016;222:590–593. doi: 10.1016/j.ijcard.2016.07.268 [DOI] [PubMed] [Google Scholar]
- 5. Dou Q, Wang W, Wang H, Ma Y, Hai S, Lin X, Liu Y, Zhang X, Wu J, Dong B. Prognostic value of frailty in elderly patients with acute coronary syndrome: a systematic review and meta‐analysis. BMC Geriatr. 2019;19:222. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. White HD, Westerhout CM, Alexander KP, Roe MT, Winters KJ, Cyr DD, Fox KA, Prabhakaran D, Hochman JS, Armstrong PW, et al. Frailty is associated with worse outcomes in non‐ST‐segment elevation acute coronary syndromes: insights from the TaRgeted platelet inhibition to cLarify the Optimal strateGy to medicallY manage Acute Coronary Syndromes (TRILOGY ACS) trial. Eur Heart J Acute Cardiovasc Care. 2016;5:231–242. doi: 10.1177/2048872615581502 [DOI] [PubMed] [Google Scholar]
- 7. Ekerstad N, Swahn E, Janzon M, Alfredsson J, Löfmark R, Lindenberger M, Carlsson P. Frailty is independently associated with short‐term outcomes for elderly patients with non‐ST‐segment elevation myocardial infarction. Circulation. 2011;124:2397–2404. doi: 10.1161/CIRCULATIONAHA.111.025452 [DOI] [PubMed] [Google Scholar]
- 8. Faridi KF, Strom JB, Kundi H, Butala NM, Curtis JP, Gao Q, Song Y, Zheng L, Tamez H, Shen C, et al. Association between claims‐defined frailty and outcomes following 30 versus 12 months of dual antiplatelet therapy after percutaneous coronary intervention: findings from the EXTEND‐DAPT study. J Am Heart Assoc. 2023;12:e029588. doi: 10.1161/JAHA.123.029588 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Levine GN, Bates ER, Bittl JA, Brindis RG, Fihn SD, Fleisher LA, Granger CB, Lange RA, Mack MJ, Mauri L, et al. 2016 ACC/AHA guideline focused update on duration of dual antiplatelet therapy in patients with coronary artery disease: a report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines: an update of the 2011 ACCF/AHA/SCAI guideline for percutaneous coronary intervention, 2011 ACCF/AHA guideline for coronary artery bypass graft surgery, 2012 ACC/AHA/ACP/AATS/PCNA/SCAI/STS guideline for the diagnosis and management of patients with stable ischemic heart disease, 2013 ACCF/AHA guideline for the management of ST‐elevation myocardial infarction, 2014 AHA/ACC guideline for the management of patients with non‐ST‐elevation acute coronary syndromes, and 2014 ACC/AHA guideline on perioperative cardiovascular evaluation and management of patients undergoing noncardiac surgery. Circulation. 2016;134:e123–e155. doi: 10.1161/CIR.0000000000000404 [DOI] [PubMed] [Google Scholar]
- 10. Lawton JS, Tamis‐Holland JE, Bangalore S, Bates ER, Beckie TM, Bischoff JM, Bittl JA, Cohen MG, Dimaio JM, Don CW, et al. 2021 ACC/AHA/SCAI guideline for coronary artery revascularization: a report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. Circulation. 2022;145:e18–e114. doi: 10.1161/CIR.0000000000001038 [DOI] [PubMed] [Google Scholar]
- 11. Wallentin L, Becker RC, Budaj A, Cannon CP, Emanuelsson H, Held C, Horrow J, Husted S, James S, Katus H, et al. Ticagrelor versus clopidogrel in patients with acute coronary syndromes. N Engl J Med. 2009;361:1045–1057. doi: 10.1056/NEJMoa0904327 [DOI] [PubMed] [Google Scholar]
- 12. Steg PG, James S, Harrington RA, Ardissino D, Becker RC, Cannon CP, Emanuelsson H, Finkelstein A, Husted S, Katus H, et al. Ticagrelor versus clopidogrel in patients with ST‐elevation acute coronary syndromes intended for reperfusion with primary percutaneous coronary intervention: a Platelet Inhibition and Patient Outcomes (PLATO) trial subgroup analysis. Circulation. 2010;122:2131–2141. doi: 10.1161/CIRCULATIONAHA.109.927582 [DOI] [PubMed] [Google Scholar]
- 13. Ko D, Pande A, Lin KJ, Cervone A, Bessette LG, Lee SB, Cheng S, Glynn RJ, Kim DH. Utilization of P2Y(12) inhibitors in older adults with ST‐elevation myocardial infarction and frailty. Am J Cardiol. 2023;207:245–252. doi: 10.1016/j.amjcard.2023.08.059 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Centers for Medicare & Medicaid services (CMS) . Chronic Conditions Data Warehouse . Accessed May 17, 2023. https://www2.ccwdata.org/web/guest/condition‐categories.
- 15. Gagne JJ, Glynn RJ, Avorn J, Levin R, Schneeweiss S. A combined comorbidity score predicted mortality in elderly patients better than existing scores. J Clin Epidemiol. 2011;64:749–759. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Kim DH, Glynn RJ, Avorn J, Lipsitz LA, Rockwood K, Pawar A, Schneeweiss S. Validation of a claims‐based frailty index against physical performance and adverse health outcomes in the health and retirement study. J Gerontol A Biol Sci Med Sci. 2019;74:1271–1276. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Kim DH, Schneeweiss S, Glynn RJ, Lipsitz LA, Rockwood K, Avorn J. Measuring frailty in medicare data: development and validation of a claims‐based frailty index. J Gerontol A Biol Sci Med Sci. 2018;73:980–987. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Kim DH, Patorno E, Pawar A, Lee H, Schneeweiss S, Glynn RJ. Measuring frailty in administrative claims data: comparative performance of four claims‐based frailty measures in the U.S. Medicare data. J Gerontol A Biol Sci Med Sci. 2020;75:1120–1125. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Gautam N, Bessette L, Pawar A, Levin R, Kim DH. Updating International Classification of Diseases Ninth Revision to Tenth Revision of a claims‐based frailty index. J Gerontol A Biol Sci Med Sci. 2021;76:1316–1317. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Cunningham A, Stein CM, Chung CP, Daugherty JR, Smalley WE, Ray WA. An automated database case definition for serious bleeding related to oral anticoagulant use. Pharmacoepidemiol Drug Saf. 2011;20:560–566. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Graham DJ, Reichman ME, Wernecke M, Hsueh YH, Izem R, Southworth MR, Wei Y, Liao J, Goulding MR, Mott K, et al. Stroke, bleeding, and mortality risks in elderly Medicare beneficiaries treated with dabigatran or rivaroxaban for nonvalvular atrial fibrillation. JAMA Intern Med. 2016;176:1662–1671. doi: 10.1001/jamainternmed.2016.5954 [DOI] [PubMed] [Google Scholar]
- 22. Graham DJ, Reichman ME, Wernecke M, Zhang R, Southworth MR, Levenson M, Sheu TC, Mott K, Goulding MR, Houstoun M, et al. Cardiovascular, bleeding, and mortality risks in elderly Medicare patients treated with dabigatran or warfarin for nonvalvular atrial fibrillation. Circulation. 2015;131:157–164. doi: 10.1161/CIRCULATIONAHA.114.012061 [DOI] [PubMed] [Google Scholar]
- 23. Kim DH, Pawar A, Gagne JJ, Bessette LG, Lee H, Glynn RJ, Schneeweiss S. Frailty and clinical outcomes of direct oral anticoagulants versus warfarin in older adults with atrial fibrillation: a cohort study. Ann Intern Med. 2021;174:1214–1223. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Wasserstein RL, Lazar NA. The ASA statement on p‐values: context, process, and purpose. Am Stat. 2016;70:129–133. doi: 10.1080/00031305.2016.1154108 [DOI] [Google Scholar]
- 25. Internationl Committee of Medical Journal Editors . Recommendations for the Conduct, Reporting, Editing, and Publications of Scholarly Work in Medical Journals . 2024. Accessed April 29, 2024. www.icmje.org/icme‐recommendations.pdf.
- 26. Savitz DA, Wise LA, Bond JC, Hatch EE, Ncube CN, Wesselink AK, Willis MD, Yland JJ, Rothman KJ. Responding to reviewers and editors about statistical significance testing. Ann Intern Med. 2024;177:385–386. [DOI] [PubMed] [Google Scholar]
- 27. Fleiss JL. The statistical basis of meta‐analysis. Stat Methods Med Res. 1993;2:121–145. doi: 10.1177/096228029300200202 [DOI] [PubMed] [Google Scholar]
- 28. VanderWeele TJ, Ding P. Sensitivity analysis in observational research: introducing the E‐value. Ann Intern Med. 2017;167:268–274. doi: 10.7326/M16-2607 [DOI] [PubMed] [Google Scholar]
- 29. Gimbel M, Qaderdan K, Willemsen L, Hermanides R, Bergmeijer T, de Vrey E, Heestermans T, Tjon Joe Gin M, Waalewijn R, Hofma S, et al. Clopidogrel versus ticagrelor or prasugrel in patients aged 70 years or older with non‐ST‐elevation acute coronary syndrome (POPular AGE): the randomised, open‐label, non‐inferiority trial. Lancet. 2020;395:1374–1381. doi: 10.1016/S0140-6736(20)30325-1 [DOI] [PubMed] [Google Scholar]
- 30. Wang SV, Schneeweiss S, Franklin JM, Desai RJ, Feldman W, Garry EM, Glynn RJ, Lin KJ, Paik J, Patorno E, et al. Emulation of randomized clinical trials with nonrandomized database analyses: results of 32 clinical trials. JAMA. 2023;329:1376–1385. doi: 10.1001/jama.2023.4221 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Turgeon RD, Koshman SL, Youngson E, Har B, Wilton SB, James MT, Graham MM. Association of ticagrelor vs clopidogrel with major adverse coronary events in patients with acute coronary syndrome undergoing percutaneous coronary intervention. JAMA Intern Med. 2020;180:420–428. doi: 10.1001/jamainternmed.2019.6447 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Butt JH, Jhund PS, Belohlávek J, de Boer RA, Chiang CE, Desai AS, Drożdzż J, Hernandez AF, Inzucchi SE, Katova T, et al. Efficacy and safety of dapagliflozin according to frailty in patients with heart failure: a prespecified analysis of the DELIVER trial. Circulation. 2022;146:1210–1224. doi: 10.1161/CIRCULATIONAHA.122.061754 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Butt JH, Dewan P, Merkely B, Belohlávek J, Drożdż J, Kitakaze M, Inzucchi SE, Kosiborod MN, Martinez FA, Tereshchenko S, et al. Efficacy and safety of dapagliflozin according to frailty in heart failure with reduced ejection fraction: a post hoc analysis of the DAPA‐HF trial. Ann Intern Med. 2022;175:820–830. doi: 10.7326/M21-4776 [DOI] [PubMed] [Google Scholar]
- 34. Pandey A, Segar MW, Singh S, Reeves G, O'Connor C, Pina I, Whellan D, Kraus W, Mentz R, Kitzman D. Frailty status modifies the efficacy of exercise training among patients with chronic heart failure and reduced ejection fraction: an analysis from the HF‐ACTION trial. Circulation. 2022;146:80–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Faurot KR, Jonsson Funk M, Pate V, Brookhart MA, Patrick A, Hanson LC, Castillo WC, Sturmer T. Using claims data to predict dependency in activities of daily living as a proxy for frailty. Pharmacoepidemiol Drug Saf. 2015;24:59–66. doi: 10.1002/pds.3719 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Festa N, Shi SM, Kim DH. Accuracy of diagnosis and health service codes in identifying frailty in Medicare data. BMC Geriatr. 2020;20:329. doi: 10.1186/s12877-020-01739-w [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Gilbert T, Neuburger J, Kraindler J, Keeble E, Smith P, Ariti C, Arora S, Street A, Parker S, Roberts HC, et al. Development and validation of a Hospital Frailty Risk Score focusing on older people in acute care settings using electronic hospital records: an observational study. Lancet. 2018;391:1775–1782. doi: 10.1016/S0140-6736(18)30668-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Tables S1–S4
Figure S1