Phenotypic Considerations for Tafamidis Use in a Real-World Cohort of Older Patients with Transthyretin Amyloid Cardiomyopathy Authors: Carlo Fumagalli, Francesco Cappelli, Mathew S. Maurer, Fabian aus dem Siepen, Angela Dispenzieri, Pablo Garcia-Pavia, Mazen Hanna, Martin Carlsson, Thibaud Damy Published: 2026 ABSTRACT IntroductionDisease-modifying therapies for transthyretin amyloid cardiomyopathy (ATTR-CM) improve clinical outcomes, yet older patients are undertreated. In this analysis, we evaluated survival and cardiovascular-related hospitalizations (CVH) among patients treated with tafamidis (tafamidis free acid at 61 mg or bioequivalent tafamidis meglumine at 80 mg) aged < 80, 80–85, and > 85 years to assess effects of baseline clinical phenotype on outcomes.MethodsData were pooled from an early access cohort (ClinicalTrials.gov ID: NCT02791230) and the Transthyretin Amyloidosis Outcomes Survey registry (NCT00628745; enrollment 2019–2023; cardiac/mixed phenotype). Outcomes included all-cause mortality (ACM), CVH, and ACM or CVH. Cox proportional hazards and negative binomial regression models were adjusted for age, sex, TTR genotype, phenotype, New York Heart Association (NYHA) class, and estimated glomerular filtration rate (eGFR).ResultsAmong the 1924 patients enrolled in the study, 1188, 505, and 231 were aged < 80, 80–85, and > 85 years, respectively. Patients aged > 85 versus 80–85 and < 80 years had lower baseline mean eGFR (39.7 vs. 49.8 and 64.1 mL/min/1.73 m2, respectively) and lower modified body mass index (mBMI; 1047.1 vs. 1084.0 and 1134.3 kg g/[m2 L]). Also, more patients aged > 85 years versus 80–85 and < 80 years were NYHA class IV (3.3% vs. 0.8% and 0.9%). Kaplan–Meier estimated survival rates were significantly lower in patients aged > 85 and 80–85 years versus those aged < 80 years at months 30 (56.3% and 66.2% vs. 77.0%) and 42 (42.9% and 57.2% vs. 70.9%). ACM risk was higher in patients aged > 85 years than in those aged < 80 years (hazard ratio 1.49, 95% confidence interval [CI] 1.11–2.01). Total annual CVH rates in patients aged > 85, 80–85, and < 80 years were 0.79, 0.65, and 0.59, respectively. CVH and ACM or CVH risks were similar across groups. Patients aged 80–85 and < 80 years with preserved mBMI, eGFR, or NYHA class I/II symptoms had similar survival rates. Across groups, patients with NYHA class III symptoms or eGFR < 45 mL/min/1.73 m2 had similar survival rates.ConclusionPatients aged ≥ 80 years who were treated with tafamidis had outcomes comparable to those of younger patients when clinical vulnerability markers were considered. These findings support individualized treatment decisions based on patient’s overall clinical profile and multidimensional assessment rather than age alone.Trial registrationClinicalTrials.gov identifiers: NCT02791230, NCT00628745.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1007/s40119-026-00474-4. Key Summary Points Why carry out this study?Disease-modifying therapies (DMT) for transthyretin amyloid cardiomyopathy (ATTR-CM), including the oral medication tafamidis, benefit patients across all ages; however, older patients are less likely to receive DMT compared with younger patientsTafamidis, an approved treatment for ATTR-CM, was found to improve survival and cardiovascular-related hospitalization (CVH) in the phase 3 ATTR-ACT study and in real-world settingsThe present analysis evaluated all-cause mortality (ACM) and CVH among patients treated with tafamidis aged < 80, 80–85, and > 85 years in a contemporary real-world cohort to evaluate the impact of baseline clinical phenotype on outcomes across age groups and to assess whether advanced age alone should guide treatment candidacyWhat has been learned from the study?Patients treated with tafamidis aged < 80, 80–85, and > 85 years had broadly similar risks of CVH and the composite of ACM or CVH; the higher risk of ACM observed in patients aged > 85 years may be partly explained by a more vulnerable baseline clinical phenotype rather than by chronological age alone compared with those aged < 80 yearsIn ATTR-CM, prognosis in older patients may be better determined by clinical phenotype than by age alone, suggesting that advanced chronological age alone may not be the most informative factor to preclude consideration of DMT and supporting individualized treatment decisions based on multidimensional assessment Why carry out this study?Disease-modifying therapies (DMT) for transthyretin amyloid cardiomyopathy (ATTR-CM), including the oral medication tafamidis, benefit patients across all ages; however, older patients are less likely to receive DMT compared with younger patientsTafamidis, an approved treatment for ATTR-CM, was found to improve survival and cardiovascular-related hospitalization (CVH) in the phase 3 ATTR-ACT study and in real-world settingsThe present analysis evaluated all-cause mortality (ACM) and CVH among patients treated with tafamidis aged < 80, 80–85, and > 85 years in a contemporary real-world cohort to evaluate the impact of baseline clinical phenotype on outcomes across age groups and to assess whether advanced age alone should guide treatment candidacyWhat has been learned from the study?Patients treated with tafamidis aged < 80, 80–85, and > 85 years had broadly similar risks of CVH and the composite of ACM or CVH; the higher risk of ACM observed in patients aged > 85 years may be partly explained by a more vulnerable baseline clinical phenotype rather than by chronological age alone compared with those aged < 80 yearsIn ATTR-CM, prognosis in older patients may be better determined by clinical phenotype than by age alone, suggesting that advanced chronological age alone may not be the most informative factor to preclude consideration of DMT and supporting individualized treatment decisions based on multidimensional assessment Introduction Transthyretin (TTR) amyloid cardiomyopathy (ATTR-CM) is a progressive and life-threatening disease caused by the deposition of misfolded TTR amyloid fibrils in the myocardium. ATTR-CM results from the age-associated accumulation of wild-type TTR protein (ATTRwt-CM) or pathogenic variants in the TTR gene (ATTRv-CM) [1–3]. Historically, persons with untreated ATTR-CM have a poor prognosis, but early diagnosis and treatment can improve outcomes [2, 4, 5]. Tafamidis was the first disease-modifying therapy to receive regulatory approval for ATTR-CM and is the only TTR kinetic stabilizer administered orally once daily. In one study, patients receiving tafamidis had a statistically significant reduction in mortality at 30 months of treatment compared with placebo [6]. Tafamidis is approved in > 50 countries worldwide based on findings from the phase 3 Tafamidis in Transthyretin Cardiomyopathy Clinical Trial (ATTR-ACT; ClinicalTrials.gov ID: NCT01994889) [6]. In ATTR-ACT, tafamidis significantly reduced the risk of all-cause mortality (ACM) and the annual rate of cardiovascular (CV)-related hospitalizations after 30 months compared with placebo in patients with ATTRwt-CM or ATTRv-CM [6]. A long-term extension (LTE) study allowed patients completing ATTR-ACT to receive tafamidis for up to 60 months or until it became commercially available in their region (NCT02791230). Findings from this LTE study demonstrated significantly better survival among patients who initially received tafamidis in ATTR-ACT compared with those who first received placebo [7]. The treatment landscape for ATTR-CM has recently expanded to include acoramidis, a TTR stabilizer taken orally twice daily and approved by the US Food and Drug Administration (FDA) in 2024 and the European Medicines Agency (EMA) in 2025, and vutrisiran, a short-interfering ribonucleic acid gene silencer administered by subcutaneous injection every 3 months and approved by both the FDA and EMA in 2025 [8, 9]. ATTR-CM, particularly the wild-type form, predominantly affects older adults. However, underdiagnosis and undertreatment of older individuals with ATTR-CM are common. Although evidence indicates that disease-modifying therapies for ATTR-CM provide clinical benefit across all age groups, treatment uptake appears to be lower among older compared with younger patients, raising concern for potential age-related treatment bias [10–12]. Therefore, studies are needed to address whether chronological age alone should influence therapeutic decisions. Other clinical factors, including multiple age-related comorbidities, frailty, polypharmacy, and declining cognitive or functional capacity, may further complicate diagnosis and treatment decisions in patients of advanced age [3, 13–16]. Evaluating treatment outcomes in the context of clinical and geriatric vulnerabilities is therefore important to support individualized, comprehensive assessment and to identify patients most likely to benefit from disease-modifying therapy [10, 12, 17, 18]. A post hoc analysis from ATTR-ACT and the LTE study showed the benefits of tafamidis in patients aged ≥ 80 years [3]. In ATTR-ACT, octogenarians treated with tafamidis versus placebo had better outcomes, including reduced decline in quality of life (QoL) and functional capacity and smaller increases in N-terminal pro-B type natriuretic peptide (NT-proBNP) concentration at 30 months. In the LTE study, octogenarians receiving continuous versus delayed tafamidis maintained a lower decline in QoL and trended toward longer survival [3]. A protocol amendment to the LTE study allowed the addition of an independent, all-comers cohort of patients who had not participated in ATTR-ACT and were provided early access to tafamidis [19]. The Transthyretin Amyloidosis Outcomes Survey (THAOS; NCT00628745) is a global, longitudinal, phase 4 observational study that includes patients with TTR amyloidosis and asymptomatic carriers of pathogenic TTR variants [20]. THAOS ran for 16 years and represents the longest and largest TTR amyloidosis registry to date. The present analysis pooled data from the early access cohort and THAOS to evaluate mortality, CV-related hospitalizations (CVH), and safety of tafamidis among patients aged < 80, 80–85, and > 85 years with ATTR-CM in order to assess the impact of baseline clinical phenotype on outcomes and determine whether advanced age alone should inform access to disease modifying-therapy. Methods Study Design and Data Sources Full details of study design and eligibility criteria for the early access cohort (cohort B; NCT02791230) and THAOS (NCT00628745) have been published [19, 20]. In the present analysis, data were pooled from contemporary, real-world patients with ATTR-CM from the early access cohort and THAOS (Fig. 1). Patients in the early access cohort were diagnosed with ATTR-CM and received tafamidis free acid 61 mg (or bioequivalent tafamidis meglumine 80 mg when the former was unavailable; referred to here as tafamidis 80/61) for up to 60 months or until tafamidis became commercially available in their region, without phenotype designation. All treated patients from the early access cohort were included in the study and were assumed to have a predominantly cardiac phenotype for this analysis. Patients from THAOS were included if they enrolled between 2019 and 2023, had a predominantly cardiac or mixed phenotype at enrollment, and received only the approved dose of tafamidis 80/61 mg during THAOS [21].Fig. 1Study design. ATTR-CM Transthyretin amyloid cardiomyopathy, THAOS Transthyretin Amyloidosis Outcomes Survey Study design. ATTR-CM Transthyretin amyloid cardiomyopathy, THAOS Transthyretin Amyloidosis Outcomes Survey The main analysis compared outcomes in patients aged < 80, 80–85, and > 85 years. A supplementary analysis compared outcomes in patients aged < 80 and ≥ 80 years. Subgroup analyses were conducted based on sex (male and female), phenotype (predominantly cardiac and mixed phenotype), TTR genotype (wild-type and variant), baseline New York Heart Association (NYHA) class (I/II and III), estimated glomerular filtration rate (eGFR; < 45 and ≥ 45 mL/min/1.73 m2), and modified body mass index (mBMI) as a proxy for risk of malnutrition (< 1000 and ≥ 1000 kg g/[m2 L]) [22]. Ethical Approval All study sites received ethical or institutional review board approval before patient enrollment. All patients provided written informed consent. Both studies (main analysis and supplementary analysis) followed the Good Pharmacoepidemiology Practices guidelines and the principles of the Declaration of Helsinki. Outcomes and Statistical Analysis Baseline demographics and clinical characteristics were summarized as counts (percentage) for categorical variables and as the mean (standard deviation [SD]) or median (10th, 90th percentile) for continuous variables. Categorical and continuous variables were compared using the chi-square test and Welch’s t-test, respectively. A two-sided p < 0.05 was considered statistically significant. Concomitant heart failure (HF) medications used by patients prior to and after the initiation of tafamidis therapy were summarized descriptively. Outcomes included ACM, first CVH, recurrent CVH, and composite endpoints (ACM or first CVH, ACM or recurrent CVH) analyzed by age groups. As ATTR-CM is characterized by progressive HF leading to death and frequent CVH, a combined endpoint with ACM and CVH helps clarify whether treatment benefits result from reduced CVH, lower mortality, or both [23, 24]. Overall survival was calculated using the Kaplan–Meier method with median and 95% confidence intervals (CIs) reported, where estimable. A delayed-entry model accounted for differences among patients in time from ATTR-CM diagnosis to study entry. Time-to-event outcomes (ACM, first CVH, ACM, or first CVH) were assessed using a stratified Cox proportional hazards model adjusted for age, phenotype, TTR genotype, sex, baseline NYHA class, and eGFR as covariates. Proportional hazards assumptions were examined using Schoenfeld residuals. CVH incidence and duration of stay were summarized descriptively. Annualized frequency of CVH, ACM, or recurrent CVH was estimated using a negative binomial regression model with age, phenotype, TTR genotype, baseline NYHA class, eGFR, and sex as factors, and the logarithm of the follow-up duration as an offset term. ACM or recurrent CVH was analyzed as a repeated event using the modified Andersen-Gill model with a robust variance estimator (Lin-Wei-Yang-Ying [LWYY] model) and the same covariates as described earlier [25]. All outcome estimates by age group presented in this article were adjusted for sex, genotype, NYHA class, phenotype, and eGFR. Adverse events (AEs) were summarized by study and age group. In the early access cohort, AEs were collected from the initiation of tafamidis therapy until 28 days after the last dose [19]. In THAOS, all-causality, treatment-emergent AEs were recorded from the time of tafamidis initiation while enrolled in the study or informed consent if patients were being treated at enrollment [2, 3]. Data were collected until the end of the observation period, including ≥ 28 days after the last dose. Results Baseline Demographics and Clinical Characteristics Of the 1924 patients in the pooled dataset, 1188, 505, and 231 were aged < 80, 80–85, and > 85 years, respectively (Fig. 1); mean age at enrollment was 72.3, 82.1, and 88.0 years, respectively, and 90.2%, 87.5%, and 84.8% were male, respectively (Table 1). A higher proportion of those aged > 85 years versus < 80 and 80–85 years had a wild-type TTR genotype (97.4% vs. 80.1% and 93.3%) and a mixed phenotype (10.0% vs 6.1% and 4.0%). Table 1Baseline demographics and clinical characteristics of patients treated with tafamidisPatient baseline demographics and clinical characteristicsAge groupp value< 80 years (n = 1188)80–85 years (n = 505)> 85 years (n = 231)Age at enrollment, years Mean (SD)72.3 (6.12)82.1 (1.62)88.0 (2.22)< 0.001Sex, n (%)0.031 Male1072 (90.2)442 (87.5)196 (84.8) Female116 (9.8)63 (12.5)35 (15.2)Race, n (%)n = 1163n = 490n = 206< 0.001 White961 (80.9)441 (87.3)190 (82.3) Black or African American131 (11.0)34 (6.7)6 (2.6) Asian58 (4.9)11 (2.2)5 (2.2) Latino American1 (0.1)01 (0.4) American Hispanic1 (0.1)00 Other11 (0.9)4 (0.8)4 (1.7)Age at diagnosis, yearsn = 1114n = 485n = 208 Mean (SD)71.7 (6.13)81.3 (2.21)87.2 (2.37)< 0.001Time from diagnosis to enrollment, yearsn = 1114n = 485n = 208 Mean (SD)1.1 (1.32)1.1 (1.34)1.1 (1.26)0.990 Median (10th, 90th percentile)1.3 (0.4, 3.5)1.1 (0.4, 3.4)1.0 (0.3, 2.4)Follow-up time, years Mean (SD)1.6 (1.17)1.5 (1.12)1.2 (0.99)< 0.001TTR genotype, n (%)< 0.001 Wild-type952 (80.1)471 (93.3)225 (97.4) Variant236 (19.9)34 (6.7)6 (2.6)Phenotype, n (%)0.006 Cardiac1115 (93.9)485 (96.0)208 (90.0) Mixed73 (6.1)20 (4.0)23 (10.0)NYHA class, n (%)n = 1155n = 486n = 214< 0.001 I210 (18.2)48 (9.9)20 (9.3) II647 (56.0)254 (52.3)113 (52.8) III288 (24.9)180 (37.0)74 (34.6) IV10 (0.9)4 (0.8)7 (3.3)mBMIn = 1012n = 402n = 180 Mean (SD) <1000 and ≥1000 kg g/(m2 L)1134.3 (206.27)1084.0 (176.15)1047.1 (193.57)< 0.001 Median (10th, 90th percentile), kg g/(m2 L)1123.9 (891.5, 1383.2)1068.8 (880.8, 1315.4)1025.2 (823.7, 1311.7) <1000 kg g/m2 L, n (%)264 (26.1)127 (31.6)81 (45.0)NT-proBNP, pg/mLn = 294n = 129n = 63 Mean (SD)8933.2 (22,541.98)8069.8 (13,457.89)8664.7 (16,659.72)0.918 Median (10th, 90th percentile)2851.5 (487.0, 19,947.2)3827.0 (827.0, 20,229.0)3387.0 (651.0, 15,683.0)eGFR, mL/min/1.73 m2n = 1119n = 472n = 219 Mean (SD)64.1 (24.11)49.8 (16.41)39.7 (12.94)< 0.001 Median (10th, 90th percentile)61.4 (37.0, 95.6)48.5 (29.7, 71.0)37.8 (24.3, 55.3)NAC stage, n (%)n = 289n = 129n = 63< 0.001 I129 (44.6)38 (29.5)13 (20.6) II123 (42.6)55 (42.6)26 (41.3) III37 (12.8)36 (27.9)24 (38.1)Baseline refers to at enrollment or the start of tafamidis treatment for patients treated with tafamidiseGFR Estimated glomerular filtration rate, mBMI modified body mass index, NAC National Amyloidosis Center, NT-proBNP N-terminal pro-B-type natriuretic peptide, NYHA New York Heart Association, SD standard deviation, TTR transthyretin Table 1Baseline demographics and clinical characteristics of patients treated with tafamidisPatient baseline demographics and clinical characteristicsAge groupp value< 80 years (n = 1188)80–85 years (n = 505)> 85 years (n = 231)Age at enrollment, years Mean (SD)72.3 (6.12)82.1 (1.62)88.0 (2.22)< 0.001Sex, n (%)0.031 Male1072 (90.2)442 (87.5)196 (84.8) Female116 (9.8)63 (12.5)35 (15.2)Race, n (%)n = 1163n = 490n = 206< 0.001 White961 (80.9)441 (87.3)190 (82.3) Black or African American131 (11.0)34 (6.7)6 (2.6) Asian58 (4.9)11 (2.2)5 (2.2) Latino American1 (0.1)01 (0.4) American Hispanic1 (0.1)00 Other11 (0.9)4 (0.8)4 (1.7)Age at diagnosis, yearsn = 1114n = 485n = 208 Mean (SD)71.7 (6.13)81.3 (2.21)87.2 (2.37)< 0.001Time from diagnosis to enrollment, yearsn = 1114n = 485n = 208 Mean (SD)1.1 (1.32)1.1 (1.34)1.1 (1.26)0.990 Median (10th, 90th percentile)1.3 (0.4, 3.5)1.1 (0.4, 3.4)1.0 (0.3, 2.4)Follow-up time, years Mean (SD)1.6 (1.17)1.5 (1.12)1.2 (0.99)< 0.001TTR genotype, n (%)< 0.001 Wild-type952 (80.1)471 (93.3)225 (97.4) Variant236 (19.9)34 (6.7)6 (2.6)Phenotype, n (%)0.006 Cardiac1115 (93.9)485 (96.0)208 (90.0) Mixed73 (6.1)20 (4.0)23 (10.0)NYHA class, n (%)n = 1155n = 486n = 214< 0.001 I210 (18.2)48 (9.9)20 (9.3) II647 (56.0)254 (52.3)113 (52.8) III288 (24.9)180 (37.0)74 (34.6) IV10 (0.9)4 (0.8)7 (3.3)mBMIn = 1012n = 402n = 180 Mean (SD) <1000 and ≥1000 kg g/(m2 L)1134.3 (206.27)1084.0 (176.15)1047.1 (193.57)< 0.001 Median (10th, 90th percentile), kg g/(m2 L)1123.9 (891.5, 1383.2)1068.8 (880.8, 1315.4)1025.2 (823.7, 1311.7) <1000 kg g/m2 L, n (%)264 (26.1)127 (31.6)81 (45.0)NT-proBNP, pg/mLn = 294n = 129n = 63 Mean (SD)8933.2 (22,541.98)8069.8 (13,457.89)8664.7 (16,659.72)0.918 Median (10th, 90th percentile)2851.5 (487.0, 19,947.2)3827.0 (827.0, 20,229.0)3387.0 (651.0, 15,683.0)eGFR, mL/min/1.73 m2n = 1119n = 472n = 219 Mean (SD)64.1 (24.11)49.8 (16.41)39.7 (12.94)< 0.001 Median (10th, 90th percentile)61.4 (37.0, 95.6)48.5 (29.7, 71.0)37.8 (24.3, 55.3)NAC stage, n (%)n = 289n = 129n = 63< 0.001 I129 (44.6)38 (29.5)13 (20.6) II123 (42.6)55 (42.6)26 (41.3) III37 (12.8)36 (27.9)24 (38.1)Baseline refers to at enrollment or the start of tafamidis treatment for patients treated with tafamidiseGFR Estimated glomerular filtration rate, mBMI modified body mass index, NAC National Amyloidosis Center, NT-proBNP N-terminal pro-B-type natriuretic peptide, NYHA New York Heart Association, SD standard deviation, TTR transthyretin Baseline demographics and clinical characteristics of patients treated with tafamidis Baseline refers to at enrollment or the start of tafamidis treatment for patients treated with tafamidis eGFR Estimated glomerular filtration rate, mBMI modified body mass index, NAC National Amyloidosis Center, NT-proBNP N-terminal pro-B-type natriuretic peptide, NYHA New York Heart Association, SD standard deviation, TTR transthyretin Overall, older patients (aged > 85 years vs. < 80 and 80–85 years) presented with a more vulnerable clinical phenotype, as evidenced by a higher proportion of the former in NYHA class IV (3.3% vs. 0.9% and 0.8%) and National Amyloidosis Center (NAC) stage III (38.1% vs. 12.8% and 27.9%) and a lower mean eGFR (39.7 vs. 64.1 and 49.8 mL/min/1.73 m2). Older patients were at higher risk of malnutrition, as indicated by lower mean mBMI (1047.1 vs. 1134.3 and 1084.0 kg g/[m2 L]). Concomitant HF medication before tafamidis initiation was reported in 76.6%, 76.6%, and 62.3% of patients aged < 80, 80–85, and > 85 years, respectively. Corresponding proportions of patients receiving concomitant HF on or after tafamidis initiation were 50.2%, 48.9%, and 47.2% (Electronic Supplementary Material [ESM] Tables S1, S2). All-Cause Mortality Mean follow-up time was 1.6, 1.5, and 1.2 years in patients aged < 80, 80–85, and > 85 years, respectively (Table 1). Overall, 378 patients died during follow-up. ACM occurred in 15.9%, 23.4%, and 30.7% of patients aged < 80, 80–85, and > 85 years, respectively (Table 2). Kaplan–Meier estimated survival rates were 77.0%, 66.2%, and 56.3% at 30 months, and 70.9%, 57.2%, and 42.9% at 42 months, respectively. Median survival was not reached in patients aged < 80 years, and was 47.4 and 38.6 months in those aged 80–85 and > 85 years, respectively (Fig. 2). These findings were consistent in the delayed-entry model adjusting for time from diagnosis to enrollment (ESM Table S3). Table 2All-cause mortality and hospitalizationsOutcomesAge group< 80 years (n = 1188)80–85 years (n = 505)> 85 years (n = 231)ACM, n (%)189 (15.9)118 (23.4)71 (30.7)First CVH, n (%)273 (23.0)116 (23.0)56 (24.2)ACM or first CVH, n (%)372 (31.3)193 (38.2)99 (42.9)All-cause hospitalizations, n (%)435 (36.6)195 (38.6)101 (43.7)CVH, n (%)273 (23.0)116 (23.0)56 (24.2) Duration of hospitalization, median (range), days5.00 (0.0, 161.0)6.00 (0.0, 65.0)7.00 (1.0, 77.0) Total annual rate among all patientsa0.590.650.79ACM All-cause mortality, CVH cardiovascular-related hospitalizationaCalculated as the total CVH across all patients divided by total years of study participation across all patientsFig. 2Kaplan–Meier curve of overall survival in patients with transthyretin amyloid cardiomyopathy (ATTR-CM) treated with tafamidis, stratified by age group (< 80 vs 80–85 and >85 years). CI Confidence interval, NE non-estimable Table 2All-cause mortality and hospitalizationsOutcomesAge group< 80 years (n = 1188)80–85 years (n = 505)> 85 years (n = 231)ACM, n (%)189 (15.9)118 (23.4)71 (30.7)First CVH, n (%)273 (23.0)116 (23.0)56 (24.2)ACM or first CVH, n (%)372 (31.3)193 (38.2)99 (42.9)All-cause hospitalizations, n (%)435 (36.6)195 (38.6)101 (43.7)CVH, n (%)273 (23.0)116 (23.0)56 (24.2) Duration of hospitalization, median (range), days5.00 (0.0, 161.0)6.00 (0.0, 65.0)7.00 (1.0, 77.0) Total annual rate among all patientsa0.590.650.79ACM All-cause mortality, CVH cardiovascular-related hospitalizationaCalculated as the total CVH across all patients divided by total years of study participation across all patients All-cause mortality and hospitalizations ACM All-cause mortality, CVH cardiovascular-related hospitalization aCalculated as the total CVH across all patients divided by total years of study participation across all patients Kaplan–Meier curve of overall survival in patients with transthyretin amyloid cardiomyopathy (ATTR-CM) treated with tafamidis, stratified by age group (< 80 vs 80–85 and >85 years). CI Confidence interval, NE non-estimable The results of the subgroup analyses suggested that age-related differences in survival varied according to the severity of the baseline clinical phenotype. Among patients with preserved mBMI (≥ 1000 kg g/[m2 L]) or NYHA class I/II symptoms, survival rates at 30 and 42 months in patients aged 80–85 years were similar to those of patients aged < 80 years (Figs. 3, 4). A similar pattern was observed at 30 months among patients with eGFR ≥ 45 mL/min/1.73 m2 (Fig. 5). In contrast, in clinically vulnerable patients (NYHA class III or eGFR < 45 mL/min/1.73 m2), survival rates were numerically lower across all age groups, with no significant differences by age in either subgroup (Figs. 4, 5). In patients with baseline mBMI <1000 kg g/(m2 L), survival rates were significantly lower in those aged > 85 and 80–85 versus < 80 years at 30 and 42 months. The largest absolute reduction in 42-month survival rate compared with preserved mBMI (≥ 1000 kg g/[m2 L]) was observed in the 80- to 85-year-old group. Compared with patients aged < 80 years, those aged 80–85 and > 85 years had lower survival rates at 30 and 42 months among male participants and those with ATTRwt-CM or a predominantly cardiac phenotype (ESM Figs. S1–S3). No significant survival differences across age groups were observed among female patients and those with ATTRv-CM or a mixed phenotype (ESM Figs. S1–S3).Fig. 3Kaplan–Meier curve of overall survival by age group (< 80, 80–85 and > 85 years) in patients with transthyretin amyloid cardiomyopathy (ATTR-CM) treated with tafamidis, stratified by baseline mBMI. CI Confidence interval, mBMI modified body mass index, NE non-estimableFig. 4Kaplan–Meier curve of overall survival by age group (< 80, 80–85 and > 85 years) in patients with transthyretin amyloid cardiomyopathy (ATTR-CM) treated with tafamidis, stratified by baseline NYHA class. CI Confidence interval, NE non-estimable, NYHA New York Heart AssociationFig. 5Kaplan–Meier curve of overall survival by age group (< 80, 80–85 and > 85 years) in patients with transthyretin amyloid cardiomyopathy (ATTR-CM) treated with tafamidis, stratified by baseline eGFR. CI Confidence interval, eGFR estimated glomerular filtration rate, NE non-estimable Kaplan–Meier curve of overall survival by age group (< 80, 80–85 and > 85 years) in patients with transthyretin amyloid cardiomyopathy (ATTR-CM) treated with tafamidis, stratified by baseline mBMI. CI Confidence interval, mBMI modified body mass index, NE non-estimable Kaplan–Meier curve of overall survival by age group (< 80, 80–85 and > 85 years) in patients with transthyretin amyloid cardiomyopathy (ATTR-CM) treated with tafamidis, stratified by baseline NYHA class. CI Confidence interval, NE non-estimable, NYHA New York Heart Association Kaplan–Meier curve of overall survival by age group (< 80, 80–85 and > 85 years) in patients with transthyretin amyloid cardiomyopathy (ATTR-CM) treated with tafamidis, stratified by baseline eGFR. CI Confidence interval, eGFR estimated glomerular filtration rate, NE non-estimable In a Cox proportional hazards model adjusted for age, phenotype, TTR genotype, sex, NYHA class, and eGFR, patients aged > 85 years had a 49% higher risk of ACM compared with those aged < 80 years (hazard ratio [HR] 1.49, 95% CI 1.11–2.01), whereas no significant difference was observed between patients aged 80–85 and < 80 years and between those aged > 85 and 80–85 years (Fig. 6a). Across all age groups, male participants had a 51% lower risk of ACM versus female participants (HR 0.49, 95% CI 0.33–0.72; ESM Table S4).Fig. 6a–f All-cause mortality and cardiovascular-related hospitalization. aAnalyzed using a negative binomial regression model with age, phenotype, TTR genotype, baseline NYHA class, eGFR, and sex as factors, and the logarithm of the follow-up duration as an offset term. bAnalyzed using the Lin-Wei-Yang-Ying model. ACM All-cause mortality, CI confidence interval, CVH cardiovascular-related hospitalization a–f All-cause mortality and cardiovascular-related hospitalization. aAnalyzed using a negative binomial regression model with age, phenotype, TTR genotype, baseline NYHA class, eGFR, and sex as factors, and the logarithm of the follow-up duration as an offset term. bAnalyzed using the Lin-Wei-Yang-Ying model. ACM All-cause mortality, CI confidence interval, CVH cardiovascular-related hospitalization Hospitalizations In patients aged < 80, 80–85, and > 85 years, 36.6%, 38.6%, and 43.7% had all-cause hospitalizations, and 23.0%, 23.0%, and 24.2% had CVH, respectively (Table 2); corresponding total annual rates for CVH were 0.59, 0.65, and 0.79. Across age groups, there were no significant differences in the risk of first or recurrent CVH (Fig. 6b, c). Composite Outcome of ACM or CVH ACM or CVH was observed in 31.3%, 38.2%, and 42.9% of patients aged <80, 80–85 years, and > 85 years, respectively (Table 2). There were no significant differences in the risk of ACM or CVH (first or recurrent) across age groups (Fig. 6d–f). All-Cause Treatment-Emergent AEs Overall, tafamidis was generally well tolerated across patients of all age groups. In the early access cohort, treatment-emergent AEs occurred in 87.1%, 87.9%, and 90.2% of patients aged < 80, 80–85, and > 85 years, respectively (Table 3). Serious and severe AEs were numerically more frequent in patients aged > 85 versus < 80 and 80–85 years; however, discontinuation of the study drug due to AEs remained infrequent (0.4%, 0.3%, and 3.1%, respectively). In THAOS, no consistent age-related increase in treatment-emergent, serious, or severe AEs was observed, and temporary or permanent treatment discontinuation due to AEs was uncommon across all age groups (3.4%, 2.6%, and 4.4% [< 80, 80–85, and > 85 years, respectively]). Table 3All-cause treatment-emergent adverse events in patients treated with tafamidisAll-cause, treatment-emergent AEsEarly access cohortTHAOS< 80 years(n = 924)80–85 years(n = 389)> 85 years(n = 163)< 80 years(n = 264)80–85 years(n = 116)> 85 years(n = 68)All-causality AEs, n52572359114121711642Patients with AEs805 (87.1)342 (87.9)147 (90.2)80 (30.3)40 (34.5)19 (27.9)Patients with serious AEs436 (47.2)194 (49.9)106 (65.0)61 (23.1)32 (27.6)13 (19.1)Patients with severe AEs316 (34.2)152 (39.1)88 (54.0)50 (18.9)30 (25.9)11 (16.2)Patients discontinued treatment due to AEs4 (0.4)a1 (0.3)a5 (3.1)a9 (3.4)b3 (2.6)b3 (4.4)bPatients with dose reduced due to AEs6 (0.6)1 (0.3)1 (0.6)000Patients temporarily discontinued study due to AEs43 (4.7)30 (7.7)9 (5.5)–––Patients discontinued study due to AEs50 (5.4)28 (7.2)19 (11.7)–––Patients with medication error events00065 (24.6)31 (26.7)12 (17.6)Values are presented as n (%) unless otherwise indicatedAE Adverse event, THAOS Transthyretin Amyloidosis Outcomes SurveyaTreatment was withdrawn but patient continued in the studybFor THAOS, this includes all patients for whom the study drug was withdrawn (temporarily, permanently, or delayed) due to AEs and who continued in the study Table 3All-cause treatment-emergent adverse events in patients treated with tafamidisAll-cause, treatment-emergent AEsEarly access cohortTHAOS< 80 years(n = 924)80–85 years(n = 389)> 85 years(n = 163)< 80 years(n = 264)80–85 years(n = 116)> 85 years(n = 68)All-causality AEs, n52572359114121711642Patients with AEs805 (87.1)342 (87.9)147 (90.2)80 (30.3)40 (34.5)19 (27.9)Patients with serious AEs436 (47.2)194 (49.9)106 (65.0)61 (23.1)32 (27.6)13 (19.1)Patients with severe AEs316 (34.2)152 (39.1)88 (54.0)50 (18.9)30 (25.9)11 (16.2)Patients discontinued treatment due to AEs4 (0.4)a1 (0.3)a5 (3.1)a9 (3.4)b3 (2.6)b3 (4.4)bPatients with dose reduced due to AEs6 (0.6)1 (0.3)1 (0.6)000Patients temporarily discontinued study due to AEs43 (4.7)30 (7.7)9 (5.5)–––Patients discontinued study due to AEs50 (5.4)28 (7.2)19 (11.7)–––Patients with medication error events00065 (24.6)31 (26.7)12 (17.6)Values are presented as n (%) unless otherwise indicatedAE Adverse event, THAOS Transthyretin Amyloidosis Outcomes SurveyaTreatment was withdrawn but patient continued in the studybFor THAOS, this includes all patients for whom the study drug was withdrawn (temporarily, permanently, or delayed) due to AEs and who continued in the study All-cause treatment-emergent adverse events in patients treated with tafamidis Values are presented as n (%) unless otherwise indicated AE Adverse event, THAOS Transthyretin Amyloidosis Outcomes Survey aTreatment was withdrawn but patient continued in the study bFor THAOS, this includes all patients for whom the study drug was withdrawn (temporarily, permanently, or delayed) due to AEs and who continued in the study Supplementary Analysis Findings from the supplementary analysis comparing patients with ATTR-CM aged < 80 and ≥ 80 years are presented in the Supplementary Results section of the ESM. Discussion In this large, pooled, contemporary real-world analysis of patients with ATTR-CM treated with tafamidis, the risks of ACM, CVH, and the composite of ACM or CVH were broadly similar across age groups (< 80, 80–85, and > 85 years). Although patients aged > 85 years had an expectedly higher risk of ACM than those aged < 80 years, no statistically significant differences were observed between adjacent age groups. Patients aged > 85 years had a more advanced and clinically vulnerable baseline profile, including lower renal function, poorer nutritional status, higher NYHA class, and more advanced NAC stage, which may have contributed to their lower survival rates at 30 and 42 months. Overall, the findings suggest that prognosis in older patients may be better understood through baseline clinical phenotype rather than chronological age alone. The supplementary analysis comparing patients aged < 80 versus ≥ 80 years yielded consistent findings, with similar risks of ACM, recurrent CVH, and ACM or recurrent CVH between groups. Safety outcomes were also broadly comparable across age groups, supporting the tolerability of tafamidis in older patients. [7, 26–28]. The exploratory subgroup analyses further suggest a phenotype-based interpretation of outcomes. Survival differences by age were most evident among male patients, those with ATTRwt-CM, and those with a predominantly cardiac phenotype. However, when key vulnerability domains were considered, patients aged 80–85 years often showed survival patterns closer to those aged < 80 years, particularly when nutritional status, renal function, or symptom burden were preserved. Among patients with mBMI ≥ 1000 kg g/(m2 L) or NYHA class I/II symptoms, survival in the 80- to 85-year-old age group was broadly comparable to that of patients aged < 80 years. Conversely, among patients with impaired renal function or NYHA class III symptoms, survival was lower overall, and age-related differences were not significant, suggesting that advanced clinical vulnerability may partly outweigh chronological age as a prognostic determinant. Clinical trials and real-world studies have shown that disease-modifying therapies improve survival and reduce the decline in functional status and QoL in patients with ATTR-CM [6, 28, 29]. The results of prior studies support the use of tafamidis in older patients with ATTR-CM [2, 3, 12, 15–17, 29–31]. In a post hoc analysis of ATTR-ACT and its long-term extension study, patients aged ≥ 80 years treated with tafamidis versus placebo had smaller declines in QoL and functional capacity, as well as smaller increases in NT-proBNP at 30 months [3]. In the long-term extension study, continuous versus delayed tafamidis treatment in octogenarians was associated with less deterioration in QoL and a trend toward longer survival; median survival among patients aged ≥ 80 years was 45 months with continuous tafamidis versus 27 months with delayed treatment, compared with 80 and 41 months, respectively, among patients aged < 80 years [3]. Consistently, in the present analysis, median survival was 47 months in patients aged 80–85 years and 39 months in those aged > 85 years, whereas it was not reached in patients aged < 80 years. Additional real-world evidence supports the relevance of tafamidis treatment in older adults: an international ATTRwt-CM cohort study reported lower mortality among patients treated with tafamidis aged ≥ 80 years, although treatment initiation at ≥ 90 years was associated with higher mortality [16]. Analysis from the Healthcare European Amyloidosis Registry (HEAR) demonstrated that patients treated with tafamidis versus untreated patients aged 80–85 and > 85 years had less severe disease and higher 3-year survival rates [29]. In contrast, findings from an interinstitutional cardiac amyloid registry showed no significant differences in 30-month survival between patients aged < 80 and ≥ 80 years who were treated with tafamidis, likely reflecting differences in study design and baseline characteristics [31]. In clinical practice, uncertainty regarding disease-modifying therapy is greatest in older patients (octogenarians and nonagenarians) because chronological age alone may have been used, implicitly or explicitly, to guide treatment decisions [10]. Our findings suggest that patients aged 80–85 years with preserved nutritional status, renal function, and lower symptom burden have outcomes approaching those observed in younger patients, whereas a vulnerable clinical phenotype is associated with poorer outcomes irrespective of age. The added value of the present study lies in its focus on age-related phenotypic heterogeneity within a large real-world cohort, and in identifying clinical factors that may be more informative than chronological age when evaluating treatment candidacy and prognosis. Furthermore, the study addresses a fundamental question that remains important across disease-modifying therapies: whether advanced age alone should influence treatment decisions. Our findings suggest that individualized treatment decisions should be guided by overall clinical status and multidimensional assessment. The impact of a more severe baseline phenotype in the present analysis is consistent with the broader geriatric principle that, after 80 years of age, prognosis is often driven more by functional status and biological vulnerability than chronological age alone. In a population-based cohort of individuals aged ≥ 80 years, disability predicted mortality more strongly than multimorbidity, independent of age and other clinical variables, supporting a shift from a disease-centered to a function-centered model of care [32]. Similarly, in ATTR-CM, comprehensive geriatric domains including disability, malnutrition, impaired physical performance, frailty, and cumulative deficits were associated with mortality, whereas age was not independently associated with survival after adjustment. Older patients were also less likely to receive disease-modifying therapy, suggesting potential age-related treatment bias [11]. In a larger ATTR-CM cohort, frailty was independently associated with mortality across age groups, genotypes, disease stages, renal function, NT-proBNP levels, and diuretic use, further supporting frailty assessment as a tool to refine risk stratification beyond conventional disease markers [17]. Finally, hierarchical vulnerability phenotypes based on disability, malnutrition, and physical performance showed stepwise survival differences and improved prognostic discrimination beyond age, sex, NYHA class, and clinical context in older adults with complex cardiovascular disease, including ATTR-CM [18]. Our data suggest considering disease-modifying therapy whenever a patient has a reasonable life expectancy and a clinical phenotype consistent with the potential to derive benefit, while recognizing that treatment decisions in patients with advanced disease (e.g., severe functional limitation, marked frailty, or substantial competing risks) should be individualized. Treatment candidacy should be informed by multidimensional clinical assessment, including disease stage, renal function, nutritional status, symptom burden, frailty, functional capacity, comorbidity burden, competing risks, and patient goals. A structured comprehensive geriatric assessment may be particularly useful in this setting, both to improve prognostic stratification and to reduce the risk of age-based therapeutic inertia [8, 11]. In this context, survival differences observed in older patients may therefore reflect greater baseline disease severity, reduced physiological reserve, and non-cardiovascular competing risks (e.g., infection and malignancy) [3, 7, 26–28, 33, 34]. Other non-cardiovascular competing risks related to frailty and multimorbidity may have also contributed to mortality in older patients, particularly in those aged > 85 years, and may partially explain the observed age-related differences in ACM despite broadly similar risks of CVH across age groups [13, 35]. The treatment landscape for ATTR-CM is rapidly evolving with the availability of recently approved disease-modifying therapies [8, 9]. Patients in THAOS and the tafamidis Early Access Program were enrolled when disease awareness was lower, diagnoses were often made at more advanced stages, and alternative disease-modifying therapies were unavailable or not yet part of routine clinical practice [2, 6]. Consequently, the cohort in this analysis may not completely reflect the contemporary ATTR-CM population, in which patients are increasingly diagnosed earlier and have a broader range of therapeutic options [36, 37]. Comparisons of survival outcomes with more recent cohorts should be interpreted cautiously as differences may reflect evolving diagnostic practices, earlier referral patterns, and changes in baseline disease severity over time rather than treatment effects alone. Although these differences limit direct comparisons with more recent clinical trials and registries, the study addresses a fundamental question that remains important across disease-modifying therapies: whether advanced age alone should influence treatment decisions. The analysis has several limitations. The observational design restricted to patients treated with tafamidis, without an untreated comparator, precludes assessment of treatment effectiveness. Inclusion was restricted to patients selected for treatment, at the discretion of the treating physician, thereby potentially introducing selection bias particularly among the oldest patients. The findings should be interpreted as descriptive associations and not as evidence of treatment benefit or equivalence across age groups. The analysis pooled data from patients with hereditary and wild-type ATTR-CM to maximize statistical power and to increase representation of older patients. Although genotype was included as a covariate in adjusted models and genotype-specific subgroup analyses, genotype-related differences in natural history and prognosis should be considered when interpreting the results. NT-proBNP and NAC stage data were available in only a subset of patients, limiting our ability to comprehensively characterize baseline disease severity across the entire cohort. The study had a relatively short median follow-up duration. Although Kaplan–Meier survival estimates are presented through 42 months, survival estimates at later time points should be interpreted with caution as the number of patients remaining at risk was limited, particularly in older age groups. Moreover, as cause-specific mortality data were not systematically available, we could not directly quantify non-cardiovascular competing risks. Data on frailty, disability, physical performance, cognitive status, and other geriatric domains were not systematically collected and therefore could not be fully assessed. The findings relate only to the clinical variables available within the pooled dataset and should not be interpreted as a comprehensive assessment of frailty. Consequently, multidimensional treatment decision-making cannot be directly evaluated. The limitations might limit the applicability of the findings in contemporary cohorts and in clinical practice. Key strengths of this analysis include its large sample size of patients with ATTR-CM aged ≥ 80 years and the use of a multinational, real-world cohort comparing patients treated with tafamidis across age groups. The pooled cohort encompassed both ATTRwt-CM and ATTRv-CM, as well as cardiac and mixed phenotypes. The study further evaluated overall survival across clinically relevant subgroups defined by sex, TTR genotype, phenotype, baseline NYHA class, and eGFR. In addition, the analysis included 231 patients aged > 85 years, providing insights for a more advanced age group. Conclusions Findings from this real-world analysis suggest that outcomes in older patients treated with tafamidis were driven more by clinical phenotype than chronological age. Patients aged ≥ 80 years with relatively preserved nutritional status, renal function, and symptom burden had outcomes closer to younger patients, which may suggest that clinical markers of vulnerability could be more informative than chronological age alone when interpreting outcomes in older patients. The association between mBMI and survival suggests that nutritional status may be a modifiable prognostic determinant in older patients with ATTR-CM. Similar risks of CVH and ACM or CVH were observed across age groups. Overall, the findings support phenotype-guided, individualized treatment decisions based on multidimensional clinical assessment rather than chronological age alone. Supplementary Information Below is the link to the electronic supplementary material.Supplementary file1 (PDF 2327 KB) Supplementary file1 (PDF 2327 KB)