Beyond Survival: A Single-Center Analysis of Complication Burden, Healthcare Utilization, and Treatment Feasibility in Higher-Risk Myelodysplastic Neoplasms

Background Higher-risk myelodysplastic neoplasms (MDS) are commonly evaluated using overall survival (OS) and leukemic transformation, but these endpoints may underrepresent the day-to-day burden of cytopenia and treatment-related complications. We aimed to characterize complications as central outcomes in a real-world, single-center cohort of the Revised International Prognostic Scoring System (IPSS-R) high/very high-risk MDS and to quantify complications using a simple burden framework. Methods Retrospective study of adults with IPSS-R high/very high-risk MDS managed in a tertiary center. Infectious complications (any infection, severe infection, or sepsis); hemorrhagic complications (any bleeding or major bleeding); ICU admission; hospitalizations; hospital days; transfusion requirements; and cause of death were collected. A severe complication composite was defined as severe infection and/or sepsis, major bleeding, and/or ICU admission. A severe complication count captured the number of severe domains (zero, one, two-three). Healthcare utilization and transfusion burden were expressed as rates per patient-month. OS was assessed using Kaplan-Meier methods with log-rank testing; survivor-conditioned landmark-proxy analyses were performed at three and six months. Results Seventy-two patients were included (median age 71 years; 55.6% male); 45.8% were IPSS-R high and 54.2% very high. Hypomethylating agents (HMAs) were administered to 61.1%, and the median follow-up was 11 months. Complications were frequent: any infection occurred in 58.3%, severe infection in 36.1%, sepsis in 25.0%, any bleeding in 55.6%, major bleeding in 26.4%, and ICU admission in 16.7%. The severe complication composite occurred in 52.8%; the severe count was zero in 47.2%, one in 33.3%, and two-three in 19.4%. The severe composite was associated with inferior OS (median 11 vs. 16 months; p=0.0238), and OS worsened across severe-count categories (16 vs. 12 vs. nine months; p=0.0409). Landmark-proxy analyses supported these findings (three-month: eight vs. 13 months; p=0.0275; six-month: six vs. 18 months; p=0.0056). Median rates were 0.75 hospitalizations/month, 5.79 hospital days/month, 1.95 RBC units/month, and 0.78 platelet units/month. Patients with the severe composite had higher time-normalized burden, including hospital days/month (7.28 vs. 4.69), RBC units/month (2.48 vs. 1.47), and platelet units/month (2.14 vs. 0.00). Among 57 deaths, complication-related deaths (infection+bleeding) accounted for 52.6%, exceeding progression/acute myeloid leukemia (AML)-attributed deaths (28.1%). In multivariable models, Eastern Cooperative Oncology Group (ECOG) ≥2 and baseline platelets <50×10⁹/L independently predicted the severe composite (OR 9.28 and 4.13, respectively; area under the curve (AUC) 0.759). Dose delay/reduction occurred in 56.8% of HMA-treated patients and was associated with fewer cycles and lower response. Conclusions In higher-risk MDS, severe complications are common and define clinically meaningful subgroups with higher care intensity, worse OS, and predominantly complication-related mortality. A composite and count-based framework, paired with patient-month normalization, offers a practical approach to quantify real-world burden and identify patients at the highest risk using readily available baseline variables.