Association of individual comorbidities with outcomes in allogeneic hematopoietic cell transplantation from unrelated adult donors versus unrelated cord blood: A study on behalf of the Donor/Source and Transplant Complications Working Groups of the Japanese Society for Transplantation and Cellular Therapy

We retrospectively evaluated the effect of 17 individual comorbidities, defined by the hematopoietic cell transplantation (HCT)‐specific comorbidity index, on non‐relapse mortality (NRM) and overall survival (OS) in 9531 patients aged between 16 and 70 years who underwent their first allogeneic HCT from 8/8 and 7/8 allele‐matched unrelated donors (8/8 and 7/8 MUDs) or single‐unit unrelated cord blood (UCB) between 2011 and 2020 using data from a Japanese registry database. In the multivariate analysis, infection (adjusted hazard ratio [HR], 1.62, 95% confidence interval [CI], 1.33–1.99 for 8/8 and 7/8 MUDs; adjusted HR, 1.33, 95%CI, 1.12–1.58 for UCB) and moderate/severe hepatic comorbidity (adjusted HR, 1.57, 95%CI, 1.04–2.38 for 8/8 and 7/8 MUDs; adjusted HR, 1.53, 95%CI, 1.09–2.15 for UCB) had a significant impact on NRM in both donor groups. Cardiac comorbidity (adjusted HR, 1.40, 95%CI, 1.08–1.80), mild hepatic comorbidity (adjusted HR, 1.22, 95%CI, 1.01–1.48), rheumatologic comorbidity (adjusted HR, 1.67, 95%CI, 1.11–2.51), renal comorbidity (adjusted HR, 2.44, 95%CI, 1.46–4.09), and severe pulmonary comorbidity (adjusted HR, 1.40, 95%CI, 1.11–1.77) were significantly associated with an increased risk of NRM but only in UCB recipients. Renal comorbidity had the strongest impact on poor OS in both donor groups (adjusted HR, 1.73, 95%CI, 1.10–2.72 for 8/8 and 7/8 MUDs; adjusted HR, 2.24, 95%CI, 1.54–3.24 for UCB). Therefore, unrelated donor selection should be taken into consideration along with the presence of specific comorbidities, such as cardiac, rheumatologic, renal, mild hepatic, and severe pulmonary comorbidities.


| INTRODUCTION
Allogeneic hematopoietic cell transplantation (HCT) is a potentially curative treatment modality for intractable hematological diseases.
However, non-relapse mortality (NRM) remains a significant problem for allogeneic HCT.The HCT-specific comorbidity index (HCT-CI), proposed in 2005 by a group in Seattle, weights 17 comorbidities and organ damages according to their hazard ratios (HRs) for a 2-year NRM and found that a history of prior solid tumors, heart valve disease, severe pulmonary comorbidity, and moderate to severe hepatic comorbidity were weighted the highest at 3 points. 1[4][5] NRM in allogeneic HCT might be dependent on patient age, comorbidities, organ damage, conditioning intensity, and donor type.
It has been shown that the impact of individual comorbidities and organ damage on NRM varies with the type of conditioning regimen. 2,5However, these reports only evaluated related and unrelated bone marrow and mobilized peripheral blood stem cell transplantations and did not include cord blood transplantation (CBT).Thus, the impact of individual comorbidities and organ damage on NRM by different donor types has not been evaluated.The impact of comorbidities and organ damage on transplant outcomes for each type of unrelated donor is expected to be a useful indicator for selecting unrelated donors.In this study, we evaluated the impact of individual comorbidities and organ damage on NRM and overall mortality in unrelated adult donor and unrelated cord blood (UCB) transplantations, respectively, using Japanese registry-based data.

| Study outcomes
The primary objective was to investigate the impact of individual comorbidities on NRM for each donor type.The secondary objective was to examine the effect of these comorbidities on overall survival (OS) after HCT for each donor type.

| Definitions
NRM was defined as death from any cause without evidence of an underlying disease relapse.OS (the inverse of overall mortality) was defined as the time from HCT to death from any cause.Surviving patients were censored at the time of the last observation.The HCT-CI 1 and Eastern Cooperative Oncology Group performance status (PS) 8 were classified in accordance with published criteria.Disease type and status at HCT were classified into three groups: recipients in complete remission for hematological malignancies, those in nonremission for hematological malignancies, and those with nonmalignant hematological diseases.Myeloablative conditioning (MAC), as defined by the Center for International Blood and Marrow Transplant Research, is a regimen containing either total body irradiation fractionated doses of >8 Gy, oral busulfan doses of ≥9 mg/kg, intravenous busulfan doses of ≥7.2 mg/kg, or melphalan doses of >140 mg/m 2 .
Other regimens were classified as reduced-intensity conditioning (RIC). 9The degree of Human leukocyte antigen (HLA) matching was based on high-resolution levels at HLA-A, HLA-B, HLA-C, and HLA-DRB1 loci for HCT from unrelated adult donors or low-resolution levels at HLA-A, HLA-B, and HLA-DRB1 loci for HCT from UCB.

| Statistical analysis
Correlations and co-occurrences between individual comorbidities were evaluated according to their log (odds ratio) using Fisher's exact test.Cumulative incidence estimates were used to calculate the unadjusted cumulative incidence of NRM, which was compared using Gray's test.The Kaplan-Meier method was used to estimate the unadjusted probability of OS, which was compared using the log-rank test.In the multivariate analyses, a cause-specific Cox proportional hazards regression model was used for NRM and overall mortality (1-OS), and results are expressed as an HR and 95% confidence interval (CI).The following variables were considered in the multivariate analysis: individual comorbidities (present vs. absent, or severe vs. moderate vs. absent), age (<50 vs. ≥50 years), recipient sex (male vs. female), recipient cytomegalovirus serostatus (negative vs. positive vs. missing), PS (0-1 vs. 2-4), disease type and status at HCT (malignancies during remission vs. malignancies during non-remission vs. nonmalignancies), donor sex (male vs. female vs. missing), conditioning regimen (MAC vs. RIC), use of antithymocyte globulin (ATG) or antilymphocyte globulin (ALG) (without ATG/ALG vs. with ATG/ALG), and number of HLA mismatches (0 vs. 1 for unrelated adult donors, or 0, 1 vs. ≥2 for UCB).Patients undergoing 8/8 allele-MUDs or 7/8 allele-MUDs were combined into one group (unrelated adult donors) for all analyses because of the small number of events within each individual comorbidity.Statistical analyses were performed using EZR version 1.61 (Saitama Medical Center, Jichi Medical University), 10 a graphical user interface for the R 4.2.2 software program (R Foundation for Statistical Computing) and GraphPad Prism 9 for Mac OS X (GraphPad Software Inc, San Diego, CA).Two-sided p values are reported, and p < .05 was considered to be significant.

| Prevalence of individual comorbidities
Among the entire cohort, moderate pulmonary comorbidity was the most common (n = 1176, 12.3%), followed by infection (n = 826, 8.6%), mild hepatic comorbidity (n = 722, 7.5%), diabetes (n = 709, 7.4%), and prior solid tumor (n = 617, 6.4%).Although unrelated adult donor recipients and UCB recipients had the same five main comorbidities in common, the frequencies of the five main comorbidities were greater among UCB recipients (Figure 1A).When evaluating the correlation of these comorbidities with each other, the strongest correlation was observed for peptic ulcer and inflammatory bowel disease (IBD) (odds ratio [OR], 15.34), followed by heart valve disease and renal comorbidity (OR, 10.00), and then peptic ulcer and infection (OR, 7.82) among unrelated adult donor recipients.Among UCB recipients, the strongest correlation was observed for arrhythmia and cardiac comorbidity (OR, 8.44), followed by peptic ulcer and IBD (OR, 6.37), and then heart valve disease and cardiac comorbidity (OR, 5.83) (Figure 1B).

| Impact of individual comorbidities on NRM
The univariate analysis showed that the cumulative incidence of NRM was significantly distributed across the risk groups of HCT-CI scores in both unrelated adult donor recipients and UCB recipients (Supplementary Figure 1A).This result was maintained in the multivariate analysis (Supplementary Table 1).Among individual comorbidities, the univariate analysis showed that diabetes, infection, and severe pulmonary comorbidity were significantly associated with higher NRM in both unrelated adult donor recipients and UCB recipients  3A).
We also assessed the interaction between individual comorbidities and conditioning intensity for each donor type because the impact of some comorbidities on NRM has been reported to be dependent on the type of conditioning regimen. 2,5The increased effect of RIC on NRM was significant between patients with and without rheumatologic comorbidity in unrelated adult donor recipients (P for interaction = 0.054).In contrast, the decreased effect of RIC on NRM was significant between patients with and without diabetes in UCB recipients (P for interaction = 0.006) (Supplementary Table 3).

| Impact of individual comorbidities on OS
The univariate analysis showed that the probability of OS was significantly distributed across the risk groups of HCT-CI scores in both unrelated adult donor recipients and UCB recipients (Supplementary Figure 1B).This result was maintained in the multivariate analysis (Supplementary Table 4).Among individual comorbidities, the univariate analysis showed that prior solid tumor, diabetes, infection, renal comorbidity, severe pulmonary comorbidity, and moderate/severe hepatic comorbidity were significantly associated with lower OS in both unrelated adult donor recipients and UCB recipients (Supplementary Figure 2, Supplementary Figure 3, Supplementary Table 5).Arrhythmia, cardiac comorbidity, and mild hepatic comorbidity were significantly associated with lower OS but only in UCB recipients, whereas cerebrovascular comorbidity and peptic ulcer were significantly associated with lower OS but only in unrelated adult donor recipients (Supplementary Figure 3, Supplementary Table 5).
Regarding the early or late phase of OS following HCT, the 100-day OS was substantially influenced by arrhythmia, cardiac, hepatic, and renal comorbidities and infection.1][22] Previous studies demonstrated that the cumulative burden of the HCT-CI score failed to predict NRM 20,22 or OS 21,22 among CBT recipients.However, those previous studies included a relatively small number of patients who received CBT, which could have resulted in a lower impact of the cumulative burden of the HCT-CI score on NRM and OS in CBT.Indeed, our study, which had a relatively large cohort size, clearly demonstrated that the probability of NRM and OS was significantly distributed across the risk groups of HCT-CI scores not only in unrelated adult donor recipients but also in UCB recipients.Therefore, the cumulative burden of the HCT-CI score could predict NRM and OS, regardless of unrelated donor type.
Regarding individual comorbidities, certain comorbidities, such as renal comorbidity and infection, had a strong impact on adverse outcomes among both unrelated adult donor recipients and UCB recipients.Renal comorbidity had the strongest impact on OS in both unrelated donor recipient groups, but the impact on NRM was significant only for CBT recipients.However, renal comorbidity was a very rare comorbidity in both unrelated adult donor recipients (0.61%) and Non-relapse mortality (NRM) for individual comorbidities in unrelated adult donor hematopoietic cell transplantation (HCT) and unrelated cord blood (UCB) HCT.Unadjusted cumulative incidence curves for NRM in unrelated adult donor HCT and UCB HCT with or without a cardiac comorbidity (A), hepatic comorbidity (B), infection (C), rheumatologic comorbidity (D), renal comorbidity (E), or pulmonary comorbidity (F).Univariate group comparisons were conducted using Gray's test for NRM.[Color figure can be viewed at wileyonlinelibrary.com]UCB recipients (0.86%) because renal comorbidity was defined as serum creatinine ≥2.0 mg/dL by HCT-CI, 1 a level that is too severe for patients to receive allogeneic HCT.Recent studies have shown that an estimated glomerular filtration rate (eGFR) of <60 mL/min, which is a more liberal definition of renal dysfunction, was associated with an increased risk of NRM. 2,4,23-25Shouval et al. showed that not only moderate to severe renal comorbidity (defined as eGFR <60 mL/min) but also mild renal comorbidity (defined as eGFR 60-89.9mL/min) was strongly associated with an increased risk of NRM, 4 suggesting that renal comorbidity evaluated by eGFR could improve the stratification of the risk of NRM after allogeneic HCT.Infection also had a significant impact on NRM and OS in both unrelated donor recipient groups, which is consistent with previous studies. 2,3In general practice, most physicians would prefer to select unrelated adult donors rather than UCB among patients with carryover infection before HCT.
However, the HRs of infection on overall mortality were unexpectedly higher in unrelated adult donor recipients compared to UCB recipients.The exact mechanisms underlying the effects of infection on increased overall mortality in unrelated adult donor recipients compared to UCB recipients remain to be fully elucidated.However, infection was defined previously as a requirement for continuous antimicrobial treatment during day 0 by HCT-CI. 1 Therefore, this definition of infection might not be truly reflected in carryover infection before HCT.Indeed, given the expected slower neutrophil recovery HCT 26 because of early infectious complications, 27,28 relating to the use of potentially nephrotoxic drugs, and the development of preengraftment syndrome.0][31] All these situations could affect the poor prognosis of patients with certain comorbidities, such as cardiac comorbidity, renal comorbidity, mild hepatic comorbidity, and severe pulmonary comorbidity.Moreover, rheumatologic comorbidity might be associated with a history of immunosuppressive treatment before and during HCT, which could contribute to higher risks of infectious complications but only after UCB transplants because CBT is primarily impeded by delayed neutrophil recovery and immune reconstitution, thereby increasing the susceptibility to infectious complications. 32 found that the early or late phase of NRM and OS following HCT was different depending on the types of comorbidities.Indeed, the 100-day NRM was strongly affected by cardiac, hepatic, pulmonary, and renal comorbidities and infection, irrespective of donor type.
In contrast, the 100-day NRM was less affected by rheumatologic comorbidities, diabetes, and prior solid tumors, but these comorbidities more strongly affected the 2-year NRM.Clearly, rheumatologic comorbidity and prior solid tumors were associated with an increased risk of NRM but only after CBT in the univariate analysis.The reason for the association between later NRM and certain comorbidities in CBT recipients only is unclear, but it might partly be due to the potential and existing organ dysfunction caused by the treatment history for rheumatologic disease and solid tumors.Importantly, certain comorbidities exerted different effects on NRM between unrelated adult donor recipients and UCB recipients.However, further validation using independent cohorts is needed before recommending a change in practice for unrelated donor selection by the presence of specific comorbidities.
The prognostic impact of individual comorbidities might be dependent on conditioning intensity.Fein et al. showed that certain comorbidities were associated with excess risk of NRM for specific types of conditioning regimens. 2They also showed that certain comorbidities, such as cardiac comorbidity, severe pulmonary comorbidity, diabetes, and infection, were associated with a higher risk of NRM but only within the MAC setting. 2 Clearly, in our present study, the proportion of MAC regimens was almost similar between unrelated adult donor recipients and UCB recipients, and the data in the multivariate analysis were adjusted for conditioning intensity.
However, it is unclear whether certain comorbidities affected NRM for specific types of conditioning regimens, irrespective of unrelated donor type.Moreover, the use of intensified conditioning regimens is preferable for CBT, 33,34 which could also have affected the results of our study.Therefore, further study is needed to clarify the prognostic impact of individual comorbidities according to specific types of conditioning regimens between donor types.
Our study had several limitations.First, a major limitation of our study was that it was a retrospective registry-based study.Therefore, the physician might have selected the donor based on each patient's comorbidity.Second, psychiatric comorbidity, obesity, and heart valve disease were not associated with an increased risk of NRM or OS in either unrelated donor recipient group in the univariate and multivariate analyses.Despite the large size of the cohort, the number of events for some rare comorbidities was too small to evaluate the impact of individual comorbidities on transplant outcomes.This implies that our data should be regarded with caution in terms of the influence of outcomes, especially for rare comorbidities.Several studies have used the term composite comorbidity (e.g., composite cardiac comorbidity including cardiac comorbidity, arrhythmia, and heart valve disease defined by HCT-CI) to overcome the under-reporting of low events for some rare comorbidities, as has been described elsewhere. 4,5However, the strength of the current study was the large number of adult patients undergoing allogeneic HCT from unrelated adult donors or UCB, which provided a good opportunity to evaluate the prognostic impact of individual comorbidities.
In summary, our registry-based study demonstrated that the risk of NRM was increased by some individual comorbidities, such as cardiac comorbidity, rheumatologic comorbidity, renal comorbidity, mild hepatic comorbidity, and severe pulmonary comorbidity, among CBT recipients, but these effects were not observed among unrelated adult donor recipients.Some individual comorbidities, such as infection and moderate/severe hepatic comorbidity, had similar effects on NRM, irrespective of unrelated donor type.Although further validation using independent cohorts is needed before changing the practice of unrelated donor selection by the presence of certain comorbidities, our results suggest that unrelated adult donors rather than UCB should be selected for patients with certain comorbidities, such as cardiac comorbidity, rheumatologic comorbidity, renal comorbidity, mild hepatic comorbidity, and severe pulmonary comorbidity.

AUTHOR CONTRIBUTIONS
Takaaki Konuma designed the research, analyzed the data, performed the statistical analysis, and wrote the manuscript.All the other authors contributed to data collection.All authors approved the final version.

ACKNOWLEDGMENTS
We thank all of the physicians and staff at the centers who provided the clinical data to the Transplant Registry Unified Management

2 | METHODS 2 . 1 |
Data source and study cohort Clinical data were extracted from the Transplant Registry Unified Management Program of the Japanese Data Center for Hematopoietic Cell Transplantation (JDCHCT) and the Japanese Society for Transplantation and Cellular Therapy (JSTCT). 6,7Patients aged between 16 and 70 years who underwent their first allogeneic HCT from 8/8 allele-matched unrelated donors (MUDs), 7/8 allele-MUDs, or single-unit UCB between 2011 and 2020 in Japan were included in this retrospective analysis.Eligibility criteria included the presence of data on the individual status of HCT-CI and survival.Patients with a previous history of autologous HCT were also excluded.Finally, 9531 patients who met these criteria were enrolled in this study.This study was approved by the Donor/ Source and Transplant Complications Working Groups of the JSTCT and by the institutional review board of the Institute of Medical Science, The University of Tokyo (2022-55-1226), where the study was conducted.

F I G U R E 1
Prevalence and co-occurrence of individual comorbidities according to unrelated donor type.(A) Prevalence of individual comorbidities according to unrelated donor type.(B) Heatmap of the correlations between individual comorbidities in unrelated adult donor hematopoietic cell transplantation (HCT) and unrelated cord blood (UCB) HCT.IBD, inflammatory bowel disease.[Color figure can be viewed at wileyonlinelibrary.com]The purpose of this retrospective study was to compare the impact of individual comorbidities on NRM and OS after allogeneic HCT using unrelated adult donors or UCB.Our study clearly demonstrated that cumulative comorbidity burden (evaluated by an HCT-CI score 0 vs. 1-2 vs. ≥3) stratified NRM and OS in both unrelated adult donor recipients and UCB recipients.Regarding individual comorbidities, cardiac comorbidity, rheumatologic comorbidity, renal comorbidity, mild hepatic comorbidity, and severe pulmonary comorbidity were significantly associated with an increased risk of NRM but only in UCB recipients in the multivariate analysis.Cardiac comorbidity, IBD, and severe pulmonary comorbidity were significantly associated with OS but only in UCB recipients in the multivariate analysis.Renal comorbidity had the strongest impact on OS in both unrelated adult donor recipients and UCB recipients, but only NRM was significantly different in CBT recipients, whereas infection had a significant impact on NRM and OS in both unrelated donor recipient groups.Importantly, certain individual comorbidities had a significant effect on NRM and OS between unrelated adult donor recipients and UCB recipients.

F I G U R E 3
Individual comorbidities-associated risk for non-relapse mortality (NRM) and overall mortality by unrelated donor type.Forest plots for the adjusted hazard ratio (HR) and 95% confidence interval (CI) for the presence of individual comorbidities on NRM (A) and overall mortality (1-OS) (B) in the multivariate analysis.IBD, inflammatory bowel disease.[Color figure can be viewed at wileyonlinelibrary.com] after UCB transplantations, more physicians might start intravenous antibiotic therapy early for UCB recipients.Moreover, the proportion of malignancies during non-remission among disease type and status was higher in UCB recipients, which means that more patients received a conditioning regimen without neutrophil recovery.These findings may contribute to the lesser effects of infection on overall mortality in UCB recipients compared to unrelated adult donor recipients.Our data clearly demonstrated that the effect of certain individual comorbidities, such as cardiac comorbidity, rheumatologic comorbidity, renal comorbidity, mild hepatic comorbidity, and severe pulmonary comorbidity, on NRM was significant only in UCB recipients.The exact mechanisms underlying the differential effects of these comorbidities on NRM between unrelated adult donor recipients and UCB recipients remain to be fully elucidated.UCB recipients could have an increased susceptibility to fluid overload during the early phase of