Introduction Acute myeloid leukemia (AML) is potentially curable with allogeneic stem cell transplantation (alloHCT); however, its curative potential is often limited by treatment-related toxicity. We have previously demonstrated that frailty independently predicts both overall survival (OS) and non-relapse mortality (NRM) in all patients undergoing alloHCT. In this study, we aimed to evaluate the prognostic impact of frailty in patients under the age of 65 with AML undergoing alloHCT.

Methods We retrospectively analyzed 213 patients aged <65 years who underwent alloHCT between 2014 and 2025. The previously validated Hematopoietic Cell Transplantation Frailty Scale (HCT-FS) was used, comprising eight domains: Clinical Frailty Score (≥3 vs. not frail), instrumental activities of daily living (IADL; ≥1 limitation vs. no limitation), timed up and go test (TUGT; >10s vs. <10s), grip strength (<16 kg for females and <26 kg for males vs. higher), self-rated health (SRH; fair or poor vs. normal), history of falls in the last 6 months (yes vs. no), serum albumin (<38 g/L vs. >38 g/L), and C-reactive protein (CRP ≥11 mg/L vs. <11 mg/L). Patients were stratified into three groups: fit (≤1), pre-frail (>1 to ≤5.5), and frail (≥5.5).

Primary endpoints were NRM and OS. Secondary endpoints included cumulative incidence of relapse (CIR), relapse-free survival (RFS), and graft-versus-host disease-free, relapse-free survival (GRFS). Kaplan-Meier estimates were calculated for OS and RFS, with comparisons via the log-rank test. Cox proportional hazards models were used for prognostic analysis, reporting hazard ratios (HRs) and 95% confidence intervals (CIs). For cumulative incidence of CIR and NRM, the Fine-Gray method and Gray’s test were applied considering competing risks.

Results The median age of the cohort was 53 years (range 18–64), with 54% (n=115) female. At pre-transplant assessment, 13.2% (n=28/212) had a high/very high disease-risk index, 30.1% (n=63/209) had an HCT-CI ≥3, and 13.4% (n=28/208) had a Karnofsky Performance Status (KPS) <90. Based on the HCT-FS, 33.3% (n=71/213) were classified as fit, 52.1% (n=111) as pre-frail, and 14.6% (n=31) as frail.

Fifty-five percent (n=119) received myeloablative conditioning. Donor sources included matched unrelated (61%, n=130), matched sibling (23.9%, n=51), and haploidentical donors (15%, n=32). Most patients (88.7%) received post-transplant cyclophosphamide-based GvHD prophylaxis.

After a median follow-up of 43.4 months [range 35.7–47.9], the 1- and 2-year NRM for the entire cohort were 13.4% [9.2–18.4] and 15.2% [11.5–21.6], respectively. When stratified by HCT-FS, fit patients had a 1- and 2-year NRM of 2.9% and 4.4%, pre-frail patients 14.0% and 15.1%, and frail patients 35.5% and 46.5% (p<0.001). Compared to fit patients, pre-frail patients had a trend toward higher NRM (HR 2.82 [0.97–8.21], p=0.057), while frail patients had a significantly higher risk (HR 2.36 [1.97–5.73], p<0.001).

The 1- and 2-year OS for the full cohort were 79.7% [73.6–84.6] and 70.8% [63.9–76.7], respectively. Fit patients had a 1- and 2-year OS of 91.4% and 86.8%, pre-frail 78.2% and 69.3%, and frail 58.1% and 40.1% (p<0.001). Compared to fit patients, frail patients had a significantly higher risk of death (HR 1.88 [1.36–2.62], p<0.001), while pre-frail patients also had decreased survival though not statistically modelled here.

The 1-year CIR for the cohort was 14.5% [10.1–19.7], with no significant differences between fit, pre-frail (p=0.59), or frail (p=0.21) groups. Similarly, the 1- and 2-year RFS were 84% [77.9–88.6] and 73.3% [65.8–79.4], with no significant differences between fit, pre-frail (p=0.36), or frail (p=0.54) patients.

The 1- and 2-year GRFS were 53.2% [46.1–59.7] and 41.2% [34.2–48.0]. Fit patients had a GRFS of 59.6% and 51.0%, pre-frail 55.0% and 39.4%, and frail 32.3% and 25.1% (p=0.002). Compared to fit patients, frail patients had a significantly higher risk of inferior GRFS (HR 1.52 [1.17–1.98], p=0.001).

Conclusion Frailty was a predictor of both OS and NRM even in young patients with AML undergoing alloHCT. Our results support the growing evidence that physiological reserve may be a more meaningful determinant of transplant fitness than chronologic age alone. Routine use of the HCT-FS could better inform treatment decisions, optimize eligibility criteria, and identify patients who might benefit from prehabilitation or modified conditioning approaches.

This content is only available as a PDF.
2025
Sign in via your Institution