Abstract
BACKGROUND:
Allogeneic hematopoietic cell transplantation (allo-HCT) improves survival in patients with acute myeloid leukemia (AML); however, posttransplant relapse remains the most common cause of treatment failure and death. Limited data exist on posttransplant health care resource utilization (HCRU) and costs, particularly for patients initiating maintenance therapy.
OBJECTIVE:
To describe HCRU and costs among commercially insured patients, Medicaid enrollees from participating US states, and Medicare-eligible beneficiaries with employer-sponsored supplemental coverage who received maintenance therapy after allo-HCT compared with those receiving allo-HCT alone.
METHODS:
We conducted a retrospective cohort study of patients with AML who underwent allo-HCT using claims data from the Merative MarketScan database from October 1, 2015, to March 31, 2024. Patients receiving maintenance therapy were identified by a claim for 1 of the following agents after allo-HCT: azacitidine, decitabine, enasidenib, gemtuzumab ozogamicin, gilteritinib, glasdegib, ivosidenib, midostaurin, quizartinib, sorafenib, or venetoclax. Groups were balanced using inverse probability treatment weighting (IPTW) based on baseline characteristics. We assessed differences in all-cause monthly HCRU and costs. HCRU included emergency department (ED) visits, inpatient (IP) admissions, outpatient (OP) visits, and hospital length of stay throughout the 12-month follow-up period. Poisson and negative binomial regression models estimated event rates. Per patient per month mean costs were reported with SEs, and between-group differences were assessed using mean differences, bootstrapped 95% CIs, and P values. Cumulative costs were summarized using the mean with bootstrapped 95% CIs and the median. IPTW-weighted mean monthly costs were also calculated for each cohort. Transfusion burden was also evaluated.
RESULTS:
Of the 373 patients who met the inclusion criteria, 43 were prescribed maintenance therapy following allo-HCT. The maintenance therapy group demonstrated significantly higher HCRU across service types. Both office (incidence rate ratio [IRR] = 3.23, P = 0.004) and OP visits (IRR = 4.26, P < 0.001) were more than tripled compared with the allo-HCT–only group, and IP admissions rose by 34% (IRR = 1.34, P = 0.034). Specialist clinic and ED visit rates were higher but not statistically significant (IRR = 3.03, P = 0.060 and IRR = 1.51, P = 0.483, respectively). Health care costs transitioned from predominantly IP at transplant to mostly pharmacy-driven by month 4, with similar trends across both groups. The maintenance therapy group had significantly higher per patient per month pharmacy costs ($6,433.81 vs $3,132.51, P = 0.002). Blood transfusion requirements were minimal in both groups. Weighted mean length of stay was significantly longer in the allo-HCT–only cohort compared with the maintenance therapy group (26.53 [SE = 1.29] vs 21.40 [SE = 1.21] days, respectively), with a mean difference of −5.13 days (95% CI = −8.60 to −1.65; P = 0.004).
CONCLUSIONS:
Those who received maintenance therapy following allo-HCT had higher IP admission and OP use rates and incurred more pharmacy costs. These findings highlight the need to balance the clinical benefits of maintenance therapy with its increased demands on health care resources.
Plain language summary
We studied 373 patients with acute myeloid leukemia (AML) receiving bone marrow transplants; 43 received maintenance therapy afterward. Those on maintenance had more doctor visits, longer hospital stays, and higher drug costs. Although total health care spending appeared higher in the maintenance therapy group, the difference was not conclusive, given the small number of patients studied. Both groups had minimal blood transfusions. These results suggest maintenance therapy increases care needs without substantially raising total health care needs during the 12-month posttransplant period.
Implications for managed care pharmacy
Maintenance therapy was associated with increased pharmacy costs as well as inpatient and outpatient use, yet overall plan expenditures remained comparable. Managed care plans should prioritize care coordination to optimize clinical outcomes and resource use after allogeneic hematopoietic cell transplantation.
Acute myeloid leukemia (AML) comprises a heterogeneous group of rapidly progressing myeloid neoplasms characterized by the clonal expansion of immature hematopoietic stem cells in the peripheral blood and bone marrow.1–3 AML is the second most common type of leukemia and the most common type of acute leukemia in adults in the United States. In 2026, approximately 22,720 individuals in the United States will be diagnosed with AML, and an estimated 11,500 deaths will be attributed to the disease. Outcomes for patients with AML remain suboptimal, with a 5-year relative survival of 33.4%.4,5
Allogeneic hematopoietic cell transplantation (allo-HCT) has improved survival in patients at high risk of relapse. Allo-HCT has been found to improve survival in patients at high risk of relapse when compared with consolidation with chemotherapy or autologous HCT (auto-HCT), with studies showing markedly higher 5-year overall survival with allo-HCT (74% vs 38% and 49% for chemotherapy and auto-HCT, respectively) and improved progression-free survival, as well as a lower cumulative incidence of relapse (P < 0.001).6 Although allo-HCT is a curative option for some patients, posttransplant relapse remains the most common cause of treatment failure and death.7,8 Nearly 20%-30% of patients with AML will relapse within the first 1-2 years after allo-HCT and face a poor prognosis with a 1-year survival rate of less than 20%.9,10
Currently, there are 3 approaches to therapy after allo-HCT including maintenance, preemptive, and active relapse therapy.11–13 Multiple randomized controlled trials and meta-analyses have shown that maintenance therapies, such as FMS-like tyrosine kinase 3 (FLT3) inhibitors (eg, sorafenib and gilteritinib) and hypomethylating agents (eg, azacitidine), can reduce relapse rates and improve relapse-free and overall survival in patients with AML who harbor FLT3 internal tandem duplication mutations.14 Maintenance therapy following allo-HCT can be inherently costly, with patients eligible to continue treatment for up to 2 years, although many discontinue early because of clinical judgement that ongoing therapy is no longer necessary. Average wholesale unit prices (per tablet/capsule for oral agents and per vial for injectables) for these therapies range from $156.20 to $12,054, highlighting the significant financial burden associated with long-term maintenance regimens in this population (Supplementary Table 17 (1.1MB, pdf) , available in online article).15–24 Additionally, clinical benefit varies considerably by therapeutic agent, with ongoing uncertainty regarding optimal treatment duration, patient selection criteria, and long-term tolerability. To date, to our knowledge, there are no real-world data regarding the economic and clinical burden associated with maintenance therapy after patients with AML receive allo-HCT. Real-world evidence on health care resource utilization (HCRU) and direct health care costs among commercially insured, Medicaid, and Medicare beneficiaries who have undergone allo-HCT is needed to inform payer and health system decision-making processes. Therefore, this analysis examines HCRU patterns and quantifies direct health care costs from the payer perspective among patients who have and have not received maintenance therapy after allo-HCT.
Methods
STUDY DESIGN AND DATA SOURCE
We conducted a retrospective cohort study using the Merative MarketScan database. The MarketScan Commercial and Encounters and Medicare Supplemental Databases capture health care utilization data for more than 273 million individuals in employer-sponsored commercial plans and Medicare-eligible retirees with employer-sponsored supplemental plans. This includes employees, their dependents, and retirees with either primary or Medicare supplemental coverage. The Multi-State Medicaid Database covers fee-for-service and managed care enrollees from multiple geographically dispersed states. Across these databases, individuals are enrolled in various health plans such as fee-for-service, point-of-service, or capitated plans. These databases contain comprehensive information on enrollment status and health care claims, encompassing inpatient (IP) and outpatient (OP) services and prescription drug use.25 All costs were based on the net payment made by the insurance plan to the provider, excluding patient out-of-pocket costs such as copayments, coinsurance, or deductibles.
The study period spanned from October 1, 2015, the start of International Classification of Diseases, Tenth Revision (ICD-10) coding in the United States, through March 31, 2024 (Figure 1). We identified patients who had their first diagnosis after April 30, 2016, with no AML diagnosis recorded at least 6 months prior. A 7-month washout period from October 1, 2015, through April 30, 2016, was implemented to exclude patients with prior malignancies.
FIGURE 1.
Study Design
allo-HCT = allogeneic hematopoietic cell transplantation; AML = acute myeloid leukemia.
All data comply with the Health Insurance Portability and Accountability Act of 1996. This research was exempt from institutional review board approval requirements as specified under 45 CFR 46.101 of the Department of Health and Human Services regulations, as MarketScan data are structured in a way that prevents participant identification, either directly or through linked identifiers.26
COHORT SELECTION
Adult patients with AML were identified by the presence of at least 1 IP claim or at least 2 OP claims more than 30 days apart of each other with a diagnosis of AML in any billing position.27 We defined the index date as the date of the first allo-HCT procedure claim after AML diagnosis; initiation of maintenance therapy occurred at varying times after allo-HCT, reflecting differences in clinical eligibility. We required patients to meet the following inclusion criteria: (1) more than 6 months continuous insurance enrollment before allo-HCT to calculate comorbidity scores at baseline and prior health care use and (2) at least a 12-month follow-up after index with continuous insurance enrollment. HCRU and cost outcomes were calculated over a fixed 12-month post–allo-HCT period (ie, follow-up was capped at 12 months for all patients to avoid censoring). Patients were excluded from analyses if they had a diagnosis of any malignancy before the index date except for myelodysplastic syndromes (MDS) or chronic myelomonocytic leukemia. After applying inclusion and exclusion criteria, patients were divided into maintenance therapy and allo-HCT–only groups based on receipt of maintenance therapy at any time during the follow-up period. The maintenance therapy cohort included patients who received maintenance therapy and remained in remission following allo-HCT until treatment initiation (Supplementary Table 7 (1.1MB, pdf) ). Patients who relapsed before or at the time of maintenance therapy initiation were excluded in the primary analysis, as maintenance therapy is prescribed only for patients in remission. As a sensitivity analysis, we repeated all primary analyses including patients who relapsed before maintenance therapy initiation to evaluate the robustness of our findings to this exclusion criterion. Diagnosis, procedure, and treatment codes can be found in Supplementary Tables 1-3 (1.1MB, pdf) .
Cohort selection was performed using SAS version 9.4 (SAS Institute, Cary, NC).
STUDY MEASURES AND OUTCOMES
Baseline demographics for both cohorts included patient age, age group (aged <35 years, 35-44 years, 45-54 years, 55-64 years, or ≥65 years), sex (female/male), insurance plan type (commercial, Medicaid, or Medicare Supplemental), geographic region (Northeast, North Central, South, West, or Unknown), and rurality (urban/rural). Baseline clinical characteristics included the year of AML diagnosis, year of allo-HCT, days to transplant, days to maintenance therapy, and previous MDS diagnosis. Pre–allo-HCT relapse status was not included as a covariate because of limitations in claims-based identification accuracy; instead, cohort design restrictions were used to ensure comparability between groups. Weighted Charlson comorbidity index (CCI) and Elixhauser comorbidity index (ECI) values were calculated from ICD-10 diagnosis codes during the pre-index period as complementary measures of comorbidity burden. The CCI and ECI were weighted using revised Quan and van Walraven weights, respectively, with the comorbidity package in R.28–30 Baseline HCRU, total pre-index period costs, and per patient per month (PPPM) costs for each type of visit during the pre-index period were assessed. HCRU outcomes included IP admissions, hospital length of stay, emergency department (ED) visits, OP visits, office visits, specialist clinic visits, and pharmacy visits. Hospital length of stay was calculated as the mean number of days per admission among patients with at least 1 IP admission during the post-index period. Number of prescription fills per agent was also described among patients receiving maintenance therapy. To characterize the cost over the post-index period, we assessed patient costs over the 12-month follow-up period, including costs by HCRU category, PPPM costs, and cumulative costs stratified by cohort. We adjusted costs to 2024 US dollars using the medical care component of the US Consumer Price Index.31
The primary outcomes included ED, IP, OP, and pharmacy costs and HCRU following allo-HCT. The secondary outcome was blood transfusion burden over the 12-month follow-up period. All outcomes were assessed between patients who received maintenance therapy and those who only underwent allo-HCT.
STATISTICAL ANALYSIS
Demographic and baseline characteristics were summarized using descriptive statistics. Continuous data were characterized through mean and SD, median, IQR, and range. Categorical variables were presented as frequencies and percentages. Propensity score weighted analysis was performed to mitigate confounding by generating a pseudo-randomized comparison in which the allo-HCT–only group was reweighted to have a similar distribution of observed baseline characteristics as the maintenance therapy group. The propensity score was estimated using a logistic regression model that included baseline demographic and clinical characteristics as well as costs to model the probability of receiving maintenance therapy (Supplementary Table 9 (1.1MB, pdf) ). We used inverse probability of treatment weighting (IPTW) with a weighting scheme for the average treatment effect on the treated (ATT) parameter to adjust for baseline differences between the cohorts, as the small sample size precluded the use of propensity score matching and ATT better reflects outcomes among patients who receive maintenance therapy in real-world practice.32 All subsequent analyses, including the calculation of weighted means and regression models, incorporated these weights to describe HCRU and quantify health care costs following allo-HCT among patients with AML. IPTW was implemented using the WeightIt package in R.33 Balance between cohorts was assessed using the absolute standardized mean difference (ASMD) and Kolmogorov-Smirnov statistics both before and after weighing, with an ASMD of no more than 0.05 considered indicative of adequate balance. Implementation of the weighting method resulted in an effective sample size of 105.3, compared with the original cohort of 373 patients. This reflects the down-weighting of patients from the comparison cohort who were dissimilar to patients in the maintenance therapy cohort, and the trade-off improves covariate balance but reduces the precision of effects estimates as only approximately one-third of the original information contributed to the weighted estimates.
HCRU was analyzed using Poisson or negative binomial regression models with bootstrapped 95% CIs to address skewed distributions. Model selection for each outcome was guided by comparative fit using the Akaike information criterion (AIC), with the model exhibiting the lowest AIC selected for reporting incidence rate ratios (IRRs). Models included treatment group and months after allo-HCT as predictors. We calculated PPPM mean costs and HCRU with SEs, mean differences, bootstrapped 95% CIs, and P values to test whether there were statistically significant differences between groups. Cumulative costs were summarized using the mean, median, and bootstrapped 95% CIs around the mean. IPTW-weighted mean monthly costs were calculated for each cohort and depicted using a stacked bar chart to visualize the contribution of different cost components to the total expenditure. Transfusion events during the post-index period were categorized as 0, 1, 2, or 3 or more events with counts and percentages presented for each category.
Statistical analyses were conducted using R version 4.4.1 (R Foundation for Statistical Computing, Vienna, Austria) via RStudio version 2025.05.0 (RStudio PBC, Boston, MA).
Results
BASELINE CHARACTERISTICS
We identified 10,495 patients with claims for AML during the index period of May 1, 2016, to March 31, 2023, allowing for 12 months of follow-up through March 31, 2024. Among this population, we identified 7,305 patients with AML who received allo-HCT. A total of 43 patients met the eligibility criteria of having maintenance therapy, and 330 only underwent allo-HCT (Figure 2).
FIGURE 2.
Description of Cohort Composition
allo-HCT = allogeneic hematopoietic cell transplantation; AML = acute myeloid leukemia.
Before IPTW, substantial imbalances were evident across multiple covariates, as the allo-HCT cohort was older, had a longer time to transplant, and exhibited higher mean CCI and ECI scores (Table 1) and higher baseline total costs for ED, OP, and pharmacy visits (Supplementary Table 11 (1.1MB, pdf) ). The maintenance therapy cohort, however, had a higher number of IP admissions, greater use of ED and OP services, more pharmacy visits, and higher total baseline IP costs during the pre-index period (Supplementary Table 10 (1.1MB, pdf) ). These distributional differences between treatment groups were reflected in the ASMDs and Kolmogorov-Smirnov statistics exceeding 0.1 for many variables.
TABLE 1.
Baseline Characteristics Before Adjusting the Allo-HCT–Only Cohort
| Characteristic | Allo-HCT only (n = 330) | Maintenance therapy (n = 43) | P value |
|---|---|---|---|
| Age | 0.011 | ||
| Mean ± SD | 50.76 ± 13.03 | 46.26 ± 12.71 | |
| Median (IQR) | 53.00 (43.00-61.00) | 49.00 (39.50-56.00) | 0.011 |
| Range | 19.00-77.00 | 20.00-72.00 | |
| Age group, n (%) | 0.279 | ||
| <35 y | 54 (16.4) | 9 (20.9) | |
| 35-44 y | 38 (11.5) | 9 (20.9) | |
| 45-54 y | 81 (24.5) | 10 (23.3) | |
| 55-64 y | 131 (39.7) | 14 (32.6) | |
| ≥65 y | —a | —a | |
| Sex, n (%) | 0.871 | ||
| Female | 147 (44.5) | 18 (41.9) | |
| Male | 183 (55.5) | 25 (58.1) | |
| Insurance plan type, n (%) | NA | ||
| CDHP | 22 (6.7) | 5 (11.6) | |
| Comprehensive | —a | —a | |
| Fee-for-service | —a | —a | |
| HDHP | —a | —a | |
| HMO | —a | —a | |
| Medicaid | 97 (29.4) | 11 (25.6) | |
| POS | —a | —a | |
| POS with capitation | —a | —a | |
| PPO | 122 (37.0) | 13 (30.2) | |
| EPO | —a | —a | |
| Region, n (%) | 0.019 | ||
| North Central | —a | —a | |
| Northeast | 50 (21.5) | 10 (31.2) | |
| South | 89 (38.2) | 18 (56.2) | |
| West | —a | —a | |
| Missing | 97 (29.4) | 11 (25.6) | |
| Rurality, n (%) | 0.453 | ||
| Metro area | 105 (45.1) | 12 (37.5) | |
| Missing | 97 (29.4) | 11 (25.6) | |
| Year of AML diagnosis, n (%) | 0.649 | ||
| 2016 | 37 (11.2) | 9 (20.9) | |
| 2017 | 51 (15.5) | 7 (16.3) | |
| 2018 | 68 (20.6) | 7 (16.3) | |
| 2019 | 56 (17.0) | 5 (11.6) | |
| 2020 | 46 (13.9) | 7 (16.3) | |
| 2021-2023 | 72 (21.8) | 8 (18.6) | |
| Year of allo-HCT, n (%) | 0.544 | ||
| 2016 | 15 (4.5) | —a | |
| 2017 | 54 (16.4) | 8 (18.6) | |
| 2018 | 60 (18.2) | 9 (20.9) | |
| 2019 | 60 (18.2) | 6 (14.0) | |
| 2020 | 44 (13.3) | 7 (16.3) | |
| 2021-2023 | 97 (29.4) | 9 (20.9) | |
| Days to transplant from diagnosis | |||
| Mean ± SD | 166.63 ± 212.06 | 146.33 ± 148.94 | |
| Median (IQR) | 120.50 (78.25-171.25) | 105.00 (76.00-157.00) | 0.722 |
| Range | 0.00-1,601.00 | 0.00-931.00 | |
| Days to maintenance therapy | |||
| Mean ± SD | — | 333.30 ± 369.70 | |
| Median (IQR) | — | 244.00 (108.00-343.00) | NA |
| Myelodysplastic syndrome, n (%) | 24 (7.3) | 7 (16.3) | 0.070 |
| Cardiac arrhythmia, n (%) | 27 (8.2) | 7 (16.3) | 0.092 |
| Weighted CCI | 0.005 | ||
| Mean ± SD | 3.18 ± 1.09 | 2.67 ± 1.00 | |
| 0, n (%) | 65 (19.7) | —a | |
| 1, n (%) | —a | —a | |
| 2, n (%) | 88 (33.2) | 23 (57.5) | |
| 3, n (%) | 79 (29.8) | 11 (27.5) | |
| ≥4, n (%) | 98 (37.0) | 6 (15.0) | |
| Missing | 65 (19.7) | —a | |
| Weighted ECI | 0.776 | ||
| Mean ± SD | 3.75 ± 4.93 | 3.30 ± 5.13 | |
| <0, n (%) | 20 (7.5) | —a | |
| 0, n (%) | 89 (33.6) | 14 (35.0) | |
| 1, n (%) | —a | —a | |
| 2, n (%) | —a | —a | |
| 3, n (%) | 45 (17.0) | 5 (12.5) | |
| ≥4, n (%) | 107 (40.4) | 16 (40.0) | |
| Missing | 65 (19.7) | —a | |
“Missing” ECI indicates no claims during this baseline period.
Denotes patient counts of less than 5.
allo-HCT = allogeneic hematopoietic cell transplantation; AML = acute myeloid leukemia; CCI = Charlson comorbidity index; CDHP = consumer-driven health plan; ECI = Elixhauser comorbidity index; EPO = exclusive provider organization; HDHP = high-deductible health plan; HMO = health maintenance organization; NA = not available; POS = point of service; PPO = preferred provider organization.
After IPTW, baseline characteristics were well balanced between the adjusted allo-HCT–only cohort (weighted n = 43.7) and the maintenance therapy cohort (n = 43), with many covariates achieving ASMDs less than or equal to 0.05 and substantially improved Kolmogorov-Smirnov statistics (Figure 3; Supplementary Table 9 (1.1MB, pdf) ). The mean age was similar (46.05 ± 13.03 vs 46.26 ± 12.71 years). Regional and insurance distributions were comparable, with the South being most represented region and commercial insurance as the predominant coverage type. Sex distribution was balanced, and the prevalence of MDS was similar between groups. The mean time to transplant did not differ (144.97 vs 146.33 days). Comorbidity scores were similar after weighting, with a slightly higher mean weighted CCI (2.73 ± 0.90 vs 2.67 ± 1.00) and weighted ECI (3.19 ± 4.48 vs 3.30 ± 5.13) in the adjusted allo-HCT cohort compared with the maintenance therapy cohort. This improved balance, achieved by weighting the allo-HCT group to resemble the maintenance therapy cohort, was expected to reduce confounding bias in outcome estimates.
FIGURE 3.
Covariate Balance Before and After Weighting
allo-HCT = allogeneic hematopoietic cell transplantation; AML = acute myeloid leukemia; CCI = Charlson comorbidity index; ECI = Elixhauser comorbidity index; ED = emergency department; MDS = myelodysplastic syndromes.
MAINTENANCE THERAPY UTILIZATION
Among the 43 patients who received maintenance therapy following allo-HCT, venetoclax was the most commonly prescribed agent (n = 24, 55.8%), followed by sorafenib (n = 9, 20.9%) and gilteritinib (n = 6, 14.0%) No patients received glasdegib, quizartinib, or gemtuzumab ozogamicin. Venetoclax-based combination regimens with hypomethylating agents such as azacitidine or decitabine were not commonly prescribed.
The median time from allo-HCT to first maintenance therapy fill was 244 days (IQR = 108-343 days). Among patients who received maintenance therapy, the median treatment duration from first to last prescription fill was 59 days (IQR = 0-259 days; mean ± SD=166 ± 248 days; maximum = 1,364 days), pointing toward variability in observed treatment duration. The number of prescription fills also varied by medication: among patients receiving venetoclax, the median number of fills was 2 (IQR = 1-4); for sorafenib, the median fills were 6 (IQR = 3-7); and for gilteritinib, the median fills were 3 (IQR = 1-13.2).
PRIMARY OUTCOME
Overall Health Care Utilization
Patients in the maintenance therapy group had a significantly higher rate of office (IRR = 3.23, P = 0.004) and OP visits (IRR = 4.26, P < 0.001) compared with the control group, indicating more than a 3-fold increase in the use of these resources relative to those who only had allo-HCT (Supplementary Table 16 (1.1MB, pdf) ). The rate of IP admissions was also significantly higher among patients who received maintenance therapy (IRR = 1.34, P = 0.034). We found that the rate of specialist clinic visits, which included hematology and oncology clinics, was higher in the maintenance therapy cohort (IRR = 3.03, P = 0.060), although this difference was not statistically significant. The same was true for rates of ED visits (IRR = 1.51, P = 0.483). Among patients with at least 1 admission during the post-index period, the weighted mean length of stay was significantly longer in the allo-HCT–only cohort compared with the maintenance therapy group (26.53 [SE = 1.29] vs 21.40 [SE = 1.21] days, respectively), with a mean difference of −5.13 days (95% CI = −8.60 to −1.65; P = 0.004). (Supplementary Table 15 (1.1MB, pdf) ).
Health Care Expenditure by Cost Type
Pharmacy costs make up the largest proportion of monthly health care expenditures in both cohorts throughout most of the 12-month follow-up period. In the month when patients underwent transplantation, IP costs accounted for approximately 97% and 96% of total costs in the allo-HCT–only and maintenance therapy cohorts, respectively. However, from the first month after allo-HCT onward, pharmacy costs became the primary cost driver, comprising roughly 57%-99% and 34%-98% of total monthly costs in the allo-HCT and maintenance therapy groups, respectively. The maintenance therapy group had higher pharmacy costs than the allo-HCT–only group, reflecting increased medication use.
In contrast, IP costs decreased substantially following transplantation, contributing approximately 18%-43% of monthly costs in the allo-HCT–only group and 10%-65% in the maintenance therapy group within the first 3 months and generally below 40% thereafter (Supplementary Table 14 (1.1MB, pdf) ). The maintenance therapy group incurred noticeably higher IP costs after month 5 compared with the allo-HCT group. Maintenance therapy was not started immediately after allo-HCT; the median time to first fill was 244 (IQR = 108-343) days (Table 1). This pattern likely reflects continued clinical monitoring and treatment support during the active treatment phase for those receiving maintenance therapy (Supplementary Figure 1 (1.1MB, pdf) ).
Patients receiving maintenance therapy following allo-HCT experienced significantly higher weighted mean PPPM pharmacy costs than those treated with allo-HCT alone ($6,433.81 vs $3,132.51, P = 0.002; Table 2). The maintenance therapy group also had higher IP PPPM costs ($10,687.84 vs $8,471.14, P = 0.394), OP PPPM costs ($13.81 vs $8.07, P = 0.542), and ED PPPM costs ($9.11 vs $3.10, P = 0.222), although these differences were not statistically significant. Total health care PPPM costs were higher in the maintenance therapy group ($17,137.09 vs $11,613.85, P = 0.054), although these differences did not reach statistical significance. This cost differential widened over time with increased uptake in maintenance therapy and suggests a sustained cost burden associated with maintenance therapy beyond the immediate posttransplant period. Nevertheless, the observed differences were not statistically significant, and the overlapping CIs indicate that although maintenance therapy might be associated with modestly higher health care spending, variability across patients limits the ability to draw definitive conclusions regarding incremental differences within 1 year following allo-HCT (Supplementary Figures 2 and 3 (1.1MB, pdf) ).
TABLE 2.
Weighted per Patient per Month Costs in the Post-Index Period
| Cost type | Allo-HCT only (n = 330), mean (SE) | Maintenance therapy (n = 43), mean (SE) | Δ Mean difference | 95% CI | P value |
|---|---|---|---|---|---|
| Pharmacy | 3,132.51 (406.14) | 6,433.81 (1,049.17) | 3,301.30 | 1,451.18 to 5,875.89 | 0.002 |
| Inpatient | 8,471.14 (1,288.75) | 10,687.84 (1,982.93) | 2,216.70 | −2,188.67 to 7,560.15 | 0.394 |
| Outpatient | 8.07 (4.83) | 13.81 (7.62) | 5.75 | −14.47 to 22.04 | 0.542 |
| ED | 3.10 (1.10) | 9.11 (4.93) | 6.01 | −1.76 to 17.20 | 0.222 |
| Total | 11,613.85 (1,421.29) | 17,137.09 (2,577.27) | 5,523.24 | 188.00 to 11,603.69 | 0.054 |
allo-HCT = allogeneic hematopoietic cell transplantation; ED = emergency department.
SECONDARY OUTCOME
During the 12-month post-index period, the weighted mean monthly number of blood transfusions was comparable between cohorts (Supplementary Figure 4 (1.1MB, pdf) ). The highest transfusion rate was observed in the month allo-HCT was performed for the allo-HCT–only cohort and had a marked decline thereafter for both groups. Across the follow-up period the mean monthly transfusion rates remained below 0.1 for both the allo-HCT–only and maintenance therapy groups.
Analysis of blood transfusions revealed that most patients required no transfusions during the post-index period with 97.6% (n = 322) in the allo-HCT–only cohort and 93% (n = 40) in the maintenance therapy cohort. A small portion of patients received 1 blood transfusion (0.9% in the allo-HCT–only group compared with 4.7% in the maintenance therapy group).
SENSITIVITY ANALYSIS
To assess the robustness of findings to the exclusion of patients who relapsed before maintenance therapy initiation, all primary analyses were repeated in a sensitivity cohort that retained the 3 patients who had 12 months of continuous enrollment after allo-HCT only, yielding a maintenance therapy group (n = 46), whereas the allo-HCT cohort remained unchanged (n = 330) (Supplementary Figure 5 (1.1MB, pdf) ).
Baseline Characteristics
Demographic and clinical characteristics in the sensitivity cohort were largely consistent with those observed in the primary analysis (Supplementary Table 18 (1.1MB, pdf) ). The maintenance therapy group remained significantly younger than the allo-HCT–only group (median 49.0 vs 53.0 years; P = 0.032), and geographic distribution by region remained statistically significant (P = 0.025), with the South and Northeast predominantly representing the patient population. Sex, insurance type, rurality, year of AML diagnosis, year of allo-HCT, and days to transplant were all nonsignificant in both the primary and sensitivity cohorts. CCI remained significantly higher in the allo-HCT group (P = 0.005), and ECI was nonsignificant in both analyses (P = 0.776). After IPTW, both cohorts were well balanced with most SMDs less than 0.1 (Supplementary Table 19, Supplementary Figure 6 (1.1MB, pdf) ).
Maintenance Therapy Utilization
The distribution of maintenance therapy agents in the sensitivity cohort was largely consistent with the primary analysis. Venetoclax remained the most commonly prescribed agent (n = 25, 54.3%, vs n = 24, 55.8%, in the primary), followed by sorafenib (n = 9, 19.6% vs 20.9%) and gilteritinib (n = 6, 13.0% vs 14.0%). Prescription fill counts were also similar between analyses. Among patients receiving venetoclax, the median number of fills was 2 (IQR = 1-4); for sorafenib, the median fills were 6 (IQR = 3-7); and for gilteritinib, the median fills were 3 (IQR = 1-13.2). The median treatment duration was 61 days (IQR = 0-245.8 days; mean = 162.1 ± 240.8 days; maximum = 1,364 days) compared with 59 days (IQR = 0-259 days; mean = 166 ± 248 days) in the primary analysis.
Overall HCRU
Post-index HCRU results were broadly consistent with the primary analysis (Supplementary Table 25 (1.1MB, pdf) ; Supplementary Table 26 (1.1MB, pdf) ). IP admission frequency remained significantly higher in the maintenance therapy group (0.21 vs 0.15, P = 0.017 vs P = 0.023 in the primary), and mean length of stay was significantly shorter in the maintenance therapy group in both analyses (sensitivity: 21.79 vs 26.89 days; primary: 21.40 vs 26.53 days, P = 0.004). OP visits remained significantly higher in the maintenance therapy group (0.29 vs 0.12, P = 0.011), as were office visits (0.11 vs 0.05, P = 0.024). ED visit rates were nonsignificant in both analyses (P = 0.470). Specialist clinic visit rates, which were borderline significant in the primary analysis (IRR = 3.03, P = 0.060), lost statistical significance in the sensitivity cohort (IRR = 2.07, P = 0.191), suggesting this endpoint is sensitive to the inclusion of relapsed patients and should be interpreted with caution. IP admission IRRs were nearly identical between analyses (primary: IRR = 1.34, P = 0.034; sensitivity: IRR = 1.35, P = 0.030), further supporting the robustness of this finding. OP and office visit IRRs remained significant but attenuated (OP: IRR = 2.91, P = 0.006; office: IRR = 2.23, P = 0.042), consistent with relapsed patients in the allo-HCT group contributing additional utilization that narrows the between-group difference.
Health Care Expenditure by Cost Type
Post-index PPPM cost results were consistent between analyses (Supplementary Table 23 (1.1MB, pdf) , Supplementary Table 24 (1.1MB, pdf) , Supplementary Figures 7-9). The finding of significantly higher pharmacy costs in the maintenance therapy group was replicated in the sensitivity cohort ($6,226.64 vs $3,087.22; P < 0.001 vs $6,433.81 vs $3,132.51; P = 0.002 in the primary), confirming that this differential is driven by the maintenance therapy. IP, OP, and ED PPPM costs were nonsignificant in both analyses with fairly minimal numerical differences. Total PPPM costs, which approached significance in the primary analysis ($17,137.09 vs $11,613.85; P = 0.054), were not significant in the sensitivity cohort ($16,548.87 vs $11,448.22; P = 0.076). Monthly cost composition patterns were qualitatively similar between analyses, with costs transitioning from predominantly IP at transplant to pharmacy-driven from the third month onward in the maintenance therapy group (Supplementary Table 24, Supplementary Figure 7 (1.1MB, pdf) ).
Discussion
We conducted a retrospective cohort study using Merative MarketScan claims data to quantify health care costs and resource utilization among patients with AML who received maintenance therapy following allo-HCT compared with those who only underwent allo-HCT. As a secondary outcome, we also assessed the differences in transfusion burden between the 2 cohorts.
The main findings of our study highlight several important differences between patients receiving maintenance therapy after allo-HCT and those who did not. One major finding was that the maintenance therapy group demonstrated significantly higher HCRU across multiple service types, including office, OP, and IP admission. Additionally, patients in the maintenance therapy cohort experienced significantly higher weighted mean PPPM pharmacy costs compared with those of the allo-HCT–only group and had higher IP, OP, and ED PPPM costs, although these latter differences were not statistically significant. Total health care PPPM costs were also higher in the maintenance therapy group, approaching but not reaching statistical significance. The observed differences in costs between groups were primarily attributable to increased pharmacy expenditures in the maintenance therapy cohort, underscoring the impact of maintenance therapy on prescription medication spending throughout the follow-up period. Another notable observation was that blood transfusion use remained minimal across groups throughout the follow-up period with a slight increase in the maintenance cohort. Overall, although maintenance therapy was associated with higher pharmacy costs and greater IP and OP use, the mean length of stay was significantly shorter in the maintenance therapy group, suggesting hospitalizations may have been less resource intensive. Furthermore, although the total PPPM costs were numerically higher, they did not reach statistical significance, collectively suggesting that the increased care needs associated with maintenance therapy may not translate into a proportionally higher overall economic burden.
These findings can be attributed to the clinical intent and consequences of maintenance therapy, which primarily aims to reduce relapse risk yet necessitates increased monitoring and management of therapy-related toxicities.34 This clinical burden likely contributes to greater pharmacy costs, potentially alongside increased OP visits and interventions to manage complications. This pattern aligns with existing literature emphasizing the need for close surveillance in patients receiving maintenance therapy.12 The predominance of pharmacy costs reflects the ongoing use of agents integral to maintenance regimens, including hypomethylating agents and targeted therapies. Although slightly more patients in the maintenance therapy group received at least 1 transfusion compared with the allo-HCT–only group, it is important to note that the maintenance cohort was smaller, which may influence observed differences in transfusion patterns. Overall, although maintenance therapy was associated with higher pharmacy costs and greater use of IP and OP services, the difference in transfusion burden between groups appeared minimal, highlighting that the impact of maintenance therapy use was concentrated in costs and health care utilization.
Our findings differ from those reported in a recent study that showed no substantial increase in HCRU among patients with FLT3 mutations who received maintenance therapy.35 These differences may reflect variations in patient populations as our study was restricted to patients who underwent allo-HCT, maintenance regimens, or changes in health care delivery practices. The study examined a medical record review cohort in which only 36.6% of patients received any maintenance after transplant (18.1% FLT3 inhibitors, 18.5% other, N = 1,208), whereas our study systematically captured all maintenance prescription fills. Our smaller maintenance therapy cohort reflects the real-world infrequency of posttransplant maintenance, potentially limiting power to detect HCRU differences compared with their larger sample. We also included cost outcomes, and the higher use we found may be driven by the need for closer monitoring and managing toxicities related to transplantation among those who have high-risk AML. Although several retrospective and small-sample prospective studies have suggested that posttransplant maintenance therapy with hypomethylating agents may lower relapse rates and improve survival, especially among those considered to have high-risk AML, other studies found no significant survival benefit to using hypomethylating agents as part of maintenance therapy.36–42 Inconsistency in the literature underscores that maintenance therapy effectiveness may be highly dependent on the patient population and regimens used. This highlights the need for further research to clarify the drivers of HCRU and to optimize the balance between clinical benefit and economic impact. To our knowledge, this is the first study to directly compare HCRU and costs between patients who receive maintenance therapy after allo-HCT and those who only underwent allo-HCT.
LIMITATIONS
Our study has several limitations inherent to claims-based research, such as coding errors or variations in billing practices. We were unable to capture key clinical indicators that may influence the decision to initiate maintenance therapy, including measurable residual disease, the type and intensity of conditioning regimen administered before allo-HCT, performance status, and the presence of mutations that negatively impact prognosis. These clinical factors may influence treatment selection and subsequent outcomes as they would not have been adequately controlled for through propensity score weighting. To align with clinical practice, patients who relapsed were excluded from the maintenance therapy cohort to ensure that the medications of interest were only administered to individuals in remission; in total, 3 patients with at least 12 months of continuous enrollment were removed on this basis. However, this approach may introduce selection bias by favoring the maintenance therapy group, as relapsed patients remained in the allo-HCT cohort and may incur higher HCRU and costs because of additional treatment and complications. Their exclusion from the maintenance group likely makes the comparative estimates of health care utilization and costs more conservative. Despite this conservative bias, our findings demonstrate that patients receiving maintenance therapy exhibit higher ED, OP, and pharmacy costs compared with those of allo-HCT–only patients, suggesting that maintenance therapy itself is a meaningful driver of increased resource use. When 3 patients who relapsed before maintenance therapy initiation were included as part of our sensitivity analysis, the results were directionally and statistically consistent with the primary analysis across nearly all outcomes. The reduced statistical significance of total PPPM costs and lower OP IRRs in the sensitivity analysis likely reflects the increased health care burden from relapsed patients in the maintenance therapy group, which introduces greater heterogeneity. Nevertheless, the overall results remain consistent with the primary analysis conclusions. Our study may also be subject to survivor bias, as we restricted the sample to patients with at least 12 months of follow-up who had already successfully undergone allo-HCT, potentially excluding patients who experienced early mortality. The 12-month follow-up period limits our ability to capture the full extent of HCRU and costs associated with maintenance therapy after allo-HCT. Given the median time to first maintenance fill of 244 (IQR = 108-343) days, our analysis likely captured only approximately 3-4 months of median maintenance exposure, which may underestimate the true resource utilization and costs for patients receiving prolonged maintenance therapy in clinical practice, where treatment can extend up to 2 years. Longer-term studies are needed to characterize the clinical and economic impact of maintenance therapy.
Furthermore, the relatively small sample size limited the statistical power of our analysis, and we could not conduct subgroup analyses. Implementation of IPTW reduced the effective sample size from 373 to 105.3, reflecting the down-weighting of patients from the comparison cohort who were dissimilar to patients in the maintenance therapy cohort. The weighted sample size for the allo-HCT cohort (n = 43.74) reported in Supplementary Table 9 (1.1MB, pdf) represents the sum of normalized ATT weights for the allo-HCT–only group scaled to match the maintenance therapy group (n = 43) for covariate balance assessment. This reduction lowered the precision of our estimates, as reflected in wider CIs. Findings are most generalizable to the maintenance therapy cohort in our analysis. Our results may not be generalizable to uninsured populations as this study only included those with commercial insurance, those with Medicare supplemental coverage, or those who were enrolled in Medicaid. The exclusion of dependents and spouses and the focus on primary beneficiaries in our analysis may also introduce selection bias related to employment status and socioeconomic factors. Lastly, the claims-based costs used in the study reflect insurance reimbursement amounts rather than the actual costs incurred from delivering health care services and may not fully capture the true economic burden of care. The complexity of claims adjudication, including postpayment adjustments and the inclusion of patients with Medicare supplemental insurance, means that reported costs may not accurately reflect the overall Medicare spending.
Future research on maintenance therapy administered after allo-HCT should consider defining the index date as the initiation of maintenance therapy rather than the date of allo-HCT itself. Anchoring the index date to the start of maintenance therapy would more accurately capture the period during which patients are exposed to the intervention, thereby providing a clearer assessment of the clinical and economic impact attributable to maintenance therapy strategies. Moreover, future research should stratify HCRU and costs incurred before and after the initiation of maintenance therapy to better distinguish the impact of pre- vs postmaintenance care after allo-HCT on overall health care costs and utilization patterns.
Conclusions
This retrospective cohort study assessed HCRU and costs among commercially insured, Medicaid-, and Medicare-enrolled patients with AML who received maintenance therapy following allo-HCT. Our findings reveal that patients who receive maintenance therapy had greater HCRU across multiple domains, including office visits, IP admissions, ED visits, and OP services and significantly higher pharmacy costs compared with those who underwent allo-HCT alone. Although IP admissions were more frequent, mean length of stay was significantly shorter, and total health care costs were numerically higher but not statistically significant. These findings highlight the need to balance the clinical benefits of maintenance therapy with its associated health care resource demands and underscore the importance of longer-term studies to fully characterize its economic impact after allo-HCT.
Disclosures
Dr Kneitel was supported by an AbbVie-sponsored fellowship with the University of Washington. Dr Lee reports no disclosures. Dr Kreif reports no disclosures.
References
- 1.Siegel RL, Giaquinto AN, Jemal A. Cancer statistics, 2024. CA Cancer J Clin. 2024;74(1):12-49. doi: 10.3322/caac.21820 [DOI] [PubMed] [Google Scholar]
- 2.Vakiti A, Reynolds SB, Mewawalla P. Acute myeloid leukemia. Updated April 27, 2024. In: StatPearls. StatPearls Publishing; 2024. Accessed May 28, 2026. https://www.ncbi.nlm.nih.gov/books/NBK507875/ [PubMed] [Google Scholar]
- 3.Chennamadhavuni A, Lyengar V, Mukkamalla SKR, et al. Leukemia. Updated January 17, 2023. In: StatPearls. StatPearls Publishing; 2024. Accessed May 28, 2026. https://www.ncbi.nlm.nih.gov/books/NBK560490/ [PubMed] [Google Scholar]
- 4.National Cancer Institute . Cancer stat facts: leukemia — acute myeloid leukemia (AML). Surveillance, Epidemiology, and End Results Program. Accessed May 2, 2026. https://seer.cancer.gov/statfacts/html/amyl.html
- 5.American Cancer Society . Key statistics for Acute myeloid leukemia (AML). Accessed May 2, 2026. https://www.cancer.org/cancer/types/acute-myeloid-leukemia/about/key-statistics.html
- 6.Wang T, Chen S, Chen J, et al. Allogeneic hematopoietic stem cell transplantation improved survival for adult core binding factor acute myelogenous leukemia patients with intermediate- and adverse-risk genetics in the 2017 European Leukemia Net. Transplant Cell Ther. 2021;27(2):173.e1-9. doi: 10.1016/j.jtct.2020.10.010 [DOI] [PubMed] [Google Scholar]
- 7.Sterling C, Webster J. Harnessing the immune system after allogeneic stem cell transplant in acute myeloid leukemia. Am J Hematol. 2020;95(5):529-47. doi: 10.1002/ajh.25750 [DOI] [PubMed] [Google Scholar]
- 8.Cornelissen JJ, Versluis J, Passweg JR, et al. ; HOVON; SAKK Leukemia Groups . Comparative therapeutic value of post-remission approaches in patients with acute myeloid leukemia aged 40-60 years. Leukemia. 2015;29(5):1041-50. doi: 10.1038/leu.2014.332 [DOI] [PubMed] [Google Scholar]
- 9.Manobianco SA, Rakiewicz T, Wilde L, Palmisiano ND. Novel mechanisms for post-transplant maintenance therapy in acute myeloid leukemia. Front Oncol. 2022;12:892289. doi: 10.3389/fonc.2022.892289 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Kreidieh F, Abou Dalle I, Moukalled N, et al. Relapse after allogeneic hematopoietic stem cell transplantation in acute myeloid leukemia: an overview of prevention and treatment. Int J Hematol. 2022;116(3):330-40. doi: 10.1007/s12185-022-03416-7 [DOI] [PubMed] [Google Scholar]
- 11.Al-Shaibani E, Novitzky-Basso I, Mattsson J, Kim DDH. Post-transplant maintenance therapy in acute myeloid leukemia after allogeneic hematopoietic stem cell transplantation harmonizing multiple therapeutic modalities including targeted therapy, immunotherapy and cellular therapy. Int J Hematol. 2023;118(1):1-17. doi: 10.1007/s12185-023-03614-x [DOI] [PubMed] [Google Scholar]
- 12.DeFilipp Z, Chen YB. How I treat with maintenance therapy after allogeneic HCT. Blood. 2023;141(1):39-48. doi: 10.1182/blood.2021012412 [DOI] [PubMed] [Google Scholar]
- 13.Parks K, Aslam MF, Kumar V, Jamy O. Post-transplant maintenance therapy in acute myeloid leukemia. Cancers (Basel). 2024;16(11):26. doi: 10.3390/cancers16112015 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Nayak RK, Chen YB. Maintenance therapy for AML after allogeneic HCT. Front Oncol. 2022;12:895771. doi: 10.3389/fonc.2022.895771 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Azacitidine . RED BOOK Online. Merative Micromedex. DynaMedex; 2025. Accessed June 10, 2025. https://www.micromedexsolutions.com
- 16.Daurismo . RED BOOK Online. Merative Micromedex. DynaMedex; 2025. Accessed June 10, 2025. https://www.micromedexsolutions.com
- 17.Idhifa . RED BOOK Online. Merative Micromedex. DynaMedex; 2025. Accessed June 10, 2025. https://www.micromedexsolutions.com
- 18.Mylotarg . RED BOOK Online. Merative Micromedex. DynaMedex; 2025. Accessed June 10, 2025. https://www.micromedexsolutions.com
- 19.Nexavar . RED BOOK Online. Merative Micromedex. DynaMedex; 2025. Accessed June 10, 2025. https://www.micromedexsolutions.com
- 20.Rydapt . RED BOOK Online. Merative Micromedex. DynaMedex; 2025. Accessed June 10, 2025. https://www.micromedexsolutions.com
- 21.Tibsovo . RED BOOK Online. Merative Micromedex. DynaMedex; 2025. Accessed June 10, 2025. https://www.micromedexsolutions.com
- 22.Vanflyta . RED BOOK Online. Merative Micromedex. DynaMedex; 2025. Accessed June 10, 2025. https://www.micromedexsolutions.com
- 23.Venclexta . RED BOOK Online. Merative Micromedex. DynaMedex; 2025. Accessed June 10, 2025. https://www.micromedexsolutions.com
- 24.Xospata . RED BOOK Online. Merative Micromedex. DynaMedex; 2025. Accessed June 10, 2025. https://www.micromedexsolutions.com
- 25.Merative MarketScan Research Databases . Accessed May 28, 2026. https://www.merative.com/documents/merative-marketscan-research-databases
- 26.US Department of Health and Human Services . Protection of human subjects. 45 CFR §46. Published Oct 1, 2005. Accessed May 18, 2025. https://www.govinfo.gov/app/details/CFR-2005-title45-vol1/CFR-2005-title45-vol1-part46
- 27.Pandya BJ, Chen CC, Medeiros BC, et al. Economic and clinical burden of relapsed and/or refractory active treatment episodes in patients with acute myeloid leukemia (AML) in the USA: a retrospective analysis of a commercial payer database. Adv Ther. 2019;36(8):1922-35. doi: 10.1007/s12325-019-01003-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Quan H, Li B, Couris CM, et al. Updating and validating the Charlson comorbidity index and score for risk adjustment in hospital discharge abstracts using data from 6 countries. Am J Epidemiol. 2011;173(6):676-82. doi: 10.1093/aje/kwq433 [DOI] [PubMed] [Google Scholar]
- 29.van Walraven C, Austin PC, Jennings A, Quan H, Forster AJ. A modification of the Elixhauser comorbidity measures into a point system for hospital death using administrative data. Med Care. 2009;47(6):626-33. doi: 10.1097/MLR.0b013e31819432e5 [DOI] [PubMed] [Google Scholar]
- 30.Gasparini A. comorbidity: An R package for computing comorbidity scores. J Open Source Softw. 2018;3(23):648, doi: 10.21105/joss.00648 [DOI] [Google Scholar]
- 31.US Bureau of Labor Statistics . Measuring price change in the CPI: medical care. Accessed May 28, 2025. https://www.bls.gov/cpi/factsheets/medical-care.htm
- 32.Morgan SL, Todd JL. A diagnostic routine for the detection of consequential heterogeneity of causal effects. Sociol Methodol. 2008;38(1):231-81. doi: 10.1111/j.1467-9531.2008.00204.x [DOI] [Google Scholar]
- 33.Greifer N. Using weightit to estimate balancing weights. GitHub. Published April 10, 2025. Accessed June 9, 2025. https://ngreifer.github.io/WeightIt/articles/WeightIt.html
- 34.Senapati J, Kadia TM, Ravandi F. Maintenance therapy in acute myeloid leukemia: advances and controversies. Haematologica. 2023;108(9):2289-304. doi: 10.3324/haematol.2022.281810 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Griffin JD, Song Y, Yang H, Freimark J, Shah MV. Post-transplant maintenance therapy in patients with FLT3-mutated acute myeloid leukemia: real-world treatment patterns and outcomes. Eur J Haematol. 2021;107(5):553-65. doi: 10.1111/ejh.13692 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Jabbour E, Giralt S, Kantarjian H, et al. Low-dose azacitidine after allogeneic stem cell transplantation for acute leukemia. Cancer. 2009;115(9):1899-905. doi: 10.1002/cncr.24198 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.de Lima M, Giralt S, Thall PF, et al. Maintenance therapy with low-dose azacitidine after allogeneic hematopoietic stem cell transplantation for recurrent acute myelogenous leukemia or myelodysplastic syndrome: a dose and schedule finding study. Cancer. 2010;116(23):5420-31. doi: 10.1002/cncr.25500 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Pusic I, Choi J, Fiala MA, et al. Maintenance therapy with decitabine after allogeneic stem cell transplantation for acute myelogenous leukemia and myelodysplastic syndrome. Biol Blood Marrow Transplant. 2015;21(10):1761-9. doi: 10.1016/j.bbmt.2015.05.026 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Ma Y, Qu C, Dai H, et al. Maintenance therapy with decitabine after allogeneic hematopoietic stem cell transplantation to prevent relapse of high-risk acute myeloid leukemia. Bone Marrow Transplant. 2020;55(6):1206-8. doi: 10.1038/s41409-019-0677-z [DOI] [PubMed] [Google Scholar]
- 40.Maples KT, Sabo RT, McCarty JM, Toor AA, Hawks KG. Maintenance azacitidine after myeloablative allogeneic hematopoietic cell transplantation for myeloid malignancies. Leuk Lymphoma. 2018;59(12):2836-41. doi: 10.1080/10428194.2018.1443334 [DOI] [PubMed] [Google Scholar]
- 41.Oshikawa G, Kakihana K, Saito M, et al. Post-transplant maintenance therapy with azacitidine and gemtuzumab ozogamicin for high-risk acute myeloid leukaemia. Br J Haematol. 2015;169(5):756-9. doi: 10.1111/bjh.13248 [DOI] [PubMed] [Google Scholar]
- 42.Xuan L, Liu Q. Maintenance therapy in acute myeloid leukemia after allogeneic hematopoietic stem cell transplantation. J Hematol Oncol. 2021;14(1):4. doi: 10.1186/s13045-020-01017-7 [DOI] [PMC free article] [PubMed] [Google Scholar]