Executive Overview
For decades, disease relapse following an allogeneic hematopoietic cell transplantation (allo-HCT)—the high-stakes curative procedure for acute leukemias and myelodysplastic syndromes (MDS)—has represented one of the most disheartening outcomes in modern oncology. When a patient relapses after transplant, the window for effective intervention narrows dramatically, and long-term survival rates plummet to historical averages below 15%.
To combat post-transplant failure, clinicians have traditionally leaned on two primary cellular salvage strategies: therapeutic donor lymphocyte infusion (DLI), designed to jumpstart a renewed graft-versus-leukemia (GvL) effect with minimal upfront toxicity, and a second allogeneic hematopoietic cell transplantation (allo-HCT), a aggressive, intensive maneuver aimed at deep cytoreduction through fresh cellular grafts and conditioning.
Despite decades of widespread clinical use, direct, high-quality head-to-head comparisons between these two strategies have remained surprisingly scarce. Medical teams have long been forced to navigate this clinical crossroads without definitive evidence regarding which pathway truly confers a superior survival advantage.
A pivotal retrospective study published in Frontiers in Hematology (June 2026, Vol. 5) by researchers at the University Medical Center Göttingen (Angleitner et al.) aims to dismantle this knowledge gap. Analyzing a uniform, single-center cohort of 85 adult patients treated between 2010 and 2022, the research team deployed advanced statistical frameworks—including propensity score matching (PSM), restricted mean survival time (RMST) modeling, and Fine–Gray competing-risks regressions—to evaluate overall survival (OS), relapse-related mortality (RRM), and non-relapse mortality (NRM).
The headline finding challenges longstanding clinical assumptions: The study detected no statistically significant difference in overall survival between patients receiving a second allo-HCT and those undergoing therapeutic DLI. While second transplants demonstrated a numerical advantage in mitigating disease relapse, this benefit was counterbalanced by a trend toward higher non-relapse mortality. Consequently, the research underscores that treatment selection must remain deeply individualized, driven by patient fitness, goals of care, and toxicity constraints rather than an expectation of guaranteed survival gains.
Detailed Chronology and Methodological Rigor
To understand the weight of these findings, one must examine the meticulous methodology implemented by the Göttingen research team. Past retrospective analyses of post-transplant relapse often suffered from methodological pitfalls, most notably immortal time bias—a distortion that occurs when the time elapsed between relapse and subsequent salvage therapy is improperly accounted for, artificially inflating the survival metrics of the intervention group.
To circumvent this, the investigators established the exact moment of relapse as "time zero" for all survival analyses, rather than the date of intervention. This design choice anchors the study in real-world clinical urgency, reflecting the true trajectory a patient faces the moment recurrence is documented.
Cohort Breakdown and Treatment Pathways
The retrospective cohort comprised 85 adult patients (median age: 60.1 years) diagnosed with acute myeloid leukemia (AML, 63%), secondary AML (sAML, 24%), MDS (14%), and acute lymphoblastic leukemia (ALL, 6%). Among these, 60 patients (71%) received therapeutic DLI, while 25 patients (29%) underwent a second allo-HCT.
The selection criteria for each modality followed standard institutional protocols shaped by multidisciplinary tumor boards:
- Second Allo-HCT Candidates: Generally exhibited favorable Eastern Cooperative Oncology Group (ECOG) performance status (median ECOG score of 1), an absence of active, uncontrolled graft-versus-host disease (GvHD) from their first transplant, and were deemed robust enough to withstand intensive conditioning regimens. These patients also experienced a significantly longer latency from their first transplant to relapse (median 74 days from relapse to treatment).
- Therapeutic DLI Candidates: Often presented with more advanced relapse dynamics, limited physiological reserve, or distinct treatment preferences. DLI was administered in overt hematologic relapse via stepwise dose escalations (starting at $0.5 times 10^6$ up to $1 times 10^7$ CD3+ cells/kg) in the absence of active GvHD. These patients received treatment much sooner after relapse detection (median 42.5 days).
Furthermore, bridging salvage therapies differed markedly between the cohorts prior to cellular intervention. Patients earmarked for a second transplant frequently underwent intensive cytoreductive multi-agent chemotherapy (such as FLAG-based protocols, HAM, or ICE). In contrast, those destined for DLI were more likely to receive lower-intensity bridging strategies, such as hypomethylating agents (HMAs) with or without venetoclax, targeted therapies, or no salvage therapy at all.
Supporting Context & Core Metrics: Survival, Relapse, and Toxicity
When evaluating the primary endpoint of overall survival across the entire cohort, the numbers revealed striking parity:
- Median OS was 0.92 years (95% CI: 0.71–1.54) in the second allo-HCT group.
- Median OS was 0.61 years (95% CI: 0.45–0.94) in the therapeutic DLI group.
- This difference did not reach statistical significance ($p = 0.295$).
Long-term survival plateau rates were similarly matched, resting at 0.13 for second-transplant recipients and 0.12 for the DLI cohort.
To control for confounding baseline variables—such as performance status and latency to relapse—the researchers executed a 1:1 propensity score-matched (PSM) cohort ($n = 36$). Within this balanced subset, median overall survival remained numerically longer following a second allo-HCT (1.01 years) compared to DLI (0.49 years), but again failed to achieve statistical significance ($p = 0.519$), shadowed by wide, overlapping confidence intervals.
Competing Risks: The Tug-of-War Between RRM and NRM
The study’s most revealing insights emerge when parsing the causes of treatment failure through competing-risks analysis. Relapse and non-relapse mortality operate in a constant biological push-and-pull:
- Relapse-Related Mortality (RRM): At 1 year post-intervention, the cumulative incidence of RRM was noticeably higher in patients receiving DLI (43.3%) compared to those undergoing a second allo-HCT (24.4%). By year two, RRM climbed to 50.4% in the DLI group versus 38.3% in the transplant arm. This indicates that second allo-HCT provides superior intermediate-term disease control.
- Non-Relapse Mortality (NRM): Conversely, this disease-control benefit exacts a heavy toll. At 1 year, the incidence of NRM was significantly higher in the second allo-HCT group (34.5%) than in the DLI group (20.3%). By year two, NRM surged to 40.5% for second-transplant recipients, compared to 24.0% for DLI patients. Fine–Gray subdistribution hazard modeling confirmed a trend toward increased NRM hazards for second transplants (sHR = 1.81).
Ultimately, the lower relapse-related mortality of a second transplant was successfully neutralized by its higher non-relapse toxicity (driven largely by lethal infectious complications like sepsis and conditioning-related organ damage), explaining why overall survival curves ultimately converged.
Independent Predictors of Survival
Multivariate Cox regression modeling isolated key clinical factors driving survival outcomes across the board:
- ECOG Performance Status: Emerged as a potent prognostic indicator. Higher ECOG scores (denoting poorer functional status) trended strongly toward increased mortality risk (HR = 1.32, $p = 0.072$).
- Conditioning Intensity of First Transplant: Patients who received myeloablative conditioning during their initial transplant demonstrated improved overall survival trajectories in univariate analyses, highlighting baseline disease biology and host resilience.
- Time to Relapse: Longer intervals from the first transplant to relapse showed a positive trend toward protective prognostic value.
Official Statements and Author Insights
Lead author Alexander C. Angleitner and senior investigators emphasized that the study’s conclusions demand a shift in how clinicians counsel patients facing this devastating diagnosis.
"We did not detect an overall survival difference between second allo-HCT and therapeutic DLI, even after propensity score matching for key clinical confounders like performance status and time to relapse," the study authors noted.
Addressing the therapeutic dilemma, the researchers highlighted the distinct clinical trade-offs discovered in their competing-risks models:
"While second allo-HCT offers superior disease control with lower relapse-related mortality, this advantage is completely offset by a substantially higher risk of non-relapse mortality. Given the intensity and toxicity of a second transplant, treatment selection must remain individualized. No overarching survival benefit can be claimed for either strategy."
Furthermore, the investigative team pointed out that cumulative DLI dose exposure in their cohort was intentionally conservative—relying on stepwise dose escalations to cap the incidence of severe graft-versus-host disease. Only a small fraction of patients successfully received the maximum planned schedule of five infusions, as rapid clinical deterioration or disease progression frequently terminated the DLI course prematurely.
Future Outlook: Refining Patient Selection in Post-Transplant Care
The findings published by Angleitner and colleagues establish a critical benchmark for hematologists worldwide. They prove that reflexively opting for a second allogeneic transplant in the name of aggressive disease eradication does not automatically translate into superior survival, due to the punishing weight of transplant-related toxicities. Conversely, while DLI spares patients immediate high-intensity toxicity, it leaves them vulnerable to unchecked disease progression.
Moving forward, the medical community must pivot away from a one-size-fits-all paradigm. Future clinical investigations must leverage larger, prospective multicenter cohorts equipped with standardized molecular tracking (such as measurable residual disease [MRD] monitoring and deep mutational profiling) to identify precise biological sub-populations. Only by isolating which specific patient subsets can truly survive the toxicity of a second transplant—or conversely, which patients harbor enough immune responsiveness to clear disease via low-intensity lymphocyte infusions—can oncology truly optimize post-relapse survivorship.
Until such data matures, clinicians must carefully weigh functional performance, organ reserve, the kinetics of relapse, and personal patient goals when standing at this high-stakes clinical crossroads.