Infants show improving survival after hematopoietic cell transplantation despite high rates of critical care admission

Highlights

• Infants undergoing HCT have a disproportionately higher risk of requiring critical care support.
• Overall survival of infants undergoing HCT who require critical care support is high and has improved in the last decade.

Introduction

Among pediatric patients, allogeneic hematopoietic cell transplantation (HCT) offers a potentially curative treatment for various malignant and non-malignant hematologic disorders (1–4). Infants represent a distinct subset compared to older pediatric patients undergoing HCT, with more unique indications, potentially exaggerated toxicity profiles, potential for more rapid clinical deterioration and limited data to predict outcomes among those who require critical care support (CC).

Non-malignant indications for infant HCT are often diagnosed through newborn screening programs and include inherited metabolic disorders such as Hurler Syndrome, infantile Krabbe disease (IKD) and inborn errors of immunity (5, 6). In some cases, pre-natal diagnosis may be possible, particularly, when there is an older affected sibling in the family (7). Time to HCT from diagnosis may be critically important. For example, HCT performed emergently ideally before 30 days of age, has a survival and functional benefit in presymptomatic infants with IKD (8). Acute infant leukemias, the most common malignancy-associated indication for HCT in infancy, typically present with aggressive disease and specific clinical, biological, cytologic, and cytogenetic characteristics such as KMT2A-r. Additionally, because juvenile myelomonocytic leukemia (9, 10) (JMML) is the most common myeloproliferative neoplasm observed in younger children, this represents a HCT indication that may be more commonly seen among infants. Similarly, inherited forms of hemophagocytic lymphohistiocytosis (HLH), presenting with immune dysregulation is more typically seen in infants under one year of age (4).

Infants receiving chemotherapy are at disproportionately higher risk for transplant related morbidity and mortality due to immature and developing organ systems, particularly underdeveloped liver metabolism, lower renal clearance (20%–40% of adult capacity), and higher body water content, leading to drug accumulation and heightened toxicity (11). Monitoring for clinical decompensation in infants also requires heightened vigilance compared to older children because infants can deteriorate rapidly and often present with subtle non-specific signs (12–14). Infants are primarily heart-rate dependent to maintain cardiac output and tachycardia may represent a critical early indicator, while bradycardia often signals imminent arrest. Hypotension is an extremely late and ominous finding among infants and respiratory monitoring for early intervention should assess signs of increased work of breathing, such as nasal flaring, grunting, and retractions (15). It is also important to assess behavioral cues in infants such as poor feeding, weak cry, inconsolability and/or floppiness, as they are unable to verbally express themselves.

Despite its curative potential, HCT may be associated with significant transplant-related morbidity and mortality, due to infection, endotheliopathies, graft-vs.-host disease (GVHD) and graft failure (7). A comprehensive study using data collected between 2009 and 2014 by the Virtual Pediatric Systems (VPS, LLC) and the Center for International Blood and Marrow Transplant Research (CIBMTR) focused on CC requirements in pediatric patients after HCT. Among all pediatric patients undergoing HCT during the study period 8.3% required CC admission within 100 days of HCT and survival to CC discharge was 85.7%. Among these patients that required CC support, overall survival at 1 year post-HCT was 52.5% (16). Early identification of pediatric patients at high risk for CC admission is crucial to preventing irreversible organ damage and improving clinical outcomes (16–19).

There is limited published literature regarding specific HCT outcomes of infants, less than 12 months of age (20). While these patients are often included in pediatric cohorts, they represent a unique subset where optimal toxicity mitigation strategies and prognostication especially when CC is needed remains sub-par (21, 22).

As such, this study aims to characterize outcomes among infants undergoing HCT with a focus on those that require CC support during the first 100 days post-HCT.

Methods

Study design and population

This IRB-exempt retrospective single center study was conducted to evaluate outcomes in infants undergoing allogeneic HCT. All infants (defined as those aged 0–12 months) who consecutively underwent HCT between 2003 and 2023 were included. Data was extracted from electronic medical records, including demographic and clinical variables, transplant-related factors, need for CC support and variables related to CC admission.

Outcomes, exposure, and variables of interest

The primary outcome of this study was death, and the primary exposure of interest was CC admission within 100 days post-transplant. Additional variables of interest included: demographics (sex, race and age at diagnosis); transplant characteristics (age at transplant, HCT indication, histocompatibility of donor cells, preparative regimen intensity, and donor type); CC admission characteristics (indication for CC admission, use of non-invasive respiratory support (NIV), and invasive mechanical ventilation (IMV) and other clinical factors including time to neutrophil and platelet engraftment, infection, donor chimerism, graft failure, relapse, and incidence of acute GVHD).

Statistical analysis

Descriptive statistics were used to summarize patient demographics and clinical characteristics. Continuous variables were reported using the median and interquartile range (25th and 75th percentiles) or range (minimum, maximum). Categorical variables were summarized using frequency counts and percentages based on non-missing values.

To evaluate the association between death and various factors—including demographics, transplant characteristics, CC admission, and other clinical variables—univariable Cox proportional hazards regression analyses were conducted. Independent variables included: age <6 months, sex, disease type (malignant vs. non-malignant), transplant type (related cord blood vs. unrelated cord blood), donor type (marrow vs. cord), neutrophil engraftment, histocompatibility, days to neutrophil engraftment (first of 3 consecutive days with an absolute neutrophil count (ANC) of 0.5 × 10⁹/L), use of NIV and IMV, duration of IMV, acute GVHD (aGVHD), chronic GVHD (cGVHD), relapse, graft failure and transplant period (2003–2013 vs. 2013–2023).

Two Cox regression models were conducted. The first model evaluated the association between individual covariates and overall survival among all patients. The second model evaluated the association between individual covariates and overall survival among patients who required critical care (CC) admission. In both models, time to death was followed until the end of observation period. Hazard ratios (HRs) and corresponding 95% confidence intervals (CIs) were reported. Kaplan–Meier (KM) plots were used to illustrate patient survival probabilities.

All statistical tests were two-sided, with a significance level of α = 0.05. Analyses were performed using SAS version 9.4 (SAS Institute Inc., Cary, NC).

Ethical considerations

This IRB exempt study was conducted in compliance with the ethical standards of the Declaration of Helsinki. Patient confidentiality was maintained throughout the study.

Results

Demographic and HCT characteristics

During the study period, 107 infants underwent allogenic HCT. Demographic and clinical characteristics are summarized in Table 1. The cohort was predominantly male (61.7%) and the vast majority had non-malignant indications for HCT (90.7%). The most frequent underlying diagnoses were inborn errors of metabolism or leukodystrophies (64.5%), followed by inborn errors of immunity (25.2%). Malignant indications for HCT included acute myeloid leukemia (3.7%), acute lymphoblastic leukemia (2.8%) and juvenile myelomonocytic leukemia (2.8%).

Table 1. Demographic and clinical characteristics of infants that underwent HCT (n = 107).

Continuous variables are reported as the median and interquartile range (25th and 75th percentiles), and categorical variables are summarized using frequency counts and percentages in parenthesis.

HCT, Hematopoietic cell transplantation.

The median age at diagnosis for HCT indication was 1.35 (Q1: 0.07, Q3: 4.83) months. Pre-natal diagnosis was made in 6.5% of patients due to previous family history/screening. The median age at HCT was 5.81 (Q1: 3.45, Q3: 10.55) months, with a median time from diagnosis to HCT of 3.12 (Q1: 2.04, Q3: 4.89) months. Most infants received myeloablative preparative regimens (87.9%) that were chemotherapy based (no patients received total body irradiation (TBI)). The most common donor graft source was umbilical cord blood (UCB) (90.7%), most commonly from unrelated UCB donors (90.7%).

Early HCT outcomes

The majority of infants (88.8%) achieved absolute neutrophil count (ANC) recovery at a median of 18 (range: 4–75) days post-HCT. Respiratory distress (46.7%), severe sepsis (17.8%), pericardial effusion (10.3%), grade 3–4 GVHD (12.1%), sinusoidal obstructive syndrome (SOS) (12%), diffuse alveolar hemorrhage (DAH) (5.6%), graft failure (5.6%), central nervous system bleeding (2.8%), and gastrointestinal bleeding (3.7%) were the most common clinically significant complications which occurred within the first 100 days post-HCT. The majority (56.7%) of infants that developed hypoxia within 100 days post-HCT ultimately required CC support.

At one-year post-HCT, the most common reported complications were hypertension (34.6%), gastrointestinal symptoms (19.6%), adrenal insufficiency (18.7%), autoimmune hemolytic anemia (AIHA) (15.9%) and developmental delay (15.9%). There were no clear differences in these one-year post-HCT complication rates between infants who required CC support and those that did not.

Critical care support in the first 100 days post HCT

Within the first 100 days post-HCT, 41% of infants required critical care support with a median length of CC admission of 8.5 (Q1 = 4.5, Q3 = 25.5) days. The median time from HCT to CC admission was 26.5 (Q1 = 11, Q3 = 51) days. Characteristics of patients requiring CC support are summarized in Table 2. The most common indications for CC admission were respiratory distress (77.3%) and sepsis (11.4%). All patients with primary graft failure (13.6%) and one of three patients with secondary graft failure required CC support.

Table 2. Characteristics of critically Ill infants undergoing HCT (n = 44).

Continuous variables are reported as the median and range (minimum and maximum), and categorical variables are summarized using frequency counts and percentages in parenthesis.

BiPAP, BiLevel Positive Airway Pressure; BBO, Blow-by Oxygen; CPAP, Continuous Positive Airway Pressure; HCT, Hematopoietic Cell Transplantation; HFNC, High-Flow Nasal Cannula; IMV, Invasive Mechanical Ventilation; NIV, Non-invasive Ventilation.

Almost half of CC admitted patients (45.5%) required respiratory support (oxygen supplementation via nasal cannula) prior to CC admission. During CC admission, 36.4% of patients required non-invasive ventilatory (NIV) support, while 56.8% of patients ultimately required invasive mechanical ventilation (IMV). Patients received IMV over a median 14 (Q1:6, Q3:24) days. Only 2 patients received NIV prior to IMV during CC admission. Among this cohort, 4.5% (n = 2) patients received continuous renal replacement therapy, 1 patient underwent placement of a tracheostomy.

Overall survival

At a median follow-up of 4.4 (IQR: 0.6–9.6) years, the overall survival (OS) rate estimated from the KM curve was 67.5% (95% CI: [57.5%, 75.6%]), corresponding to 38 deaths out of 107 infants (Figure 1A). The estimated survival rate at 100 days post-HCT was and 85.9% [95% CI: (77.6%–91.2%)], corresponding to 15 deaths, and the estimated survival rate at 1 year post-HCT was 73.5% [95% CI: (64.0%–80.9%)], corresponding to 28 deaths. Causes of death were attributed to infection (27%), multi-organ dysfunction (20%), pulmonary hemorrhage (17%), progressive disease/relapse (17%), respiratory failure (17%) and cardiac compromise (2%).

CC admission was associated with an increased risk of death, with a 3.59-fold higher hazard compared with patients who were not admitted to CC [HR = 3.59, 95% CI (1.85, 6.98), p < 0.001]. In the first Cox regression model including all infants, delayed or absent neutrophil engraftment was significantly associated with mortality. Patients who achieved neutrophil engraftment had a 69% lower hazard of death compared with those who did not achieve engraftment [HR = 0.31, 95% CI: (0.14, 0.67), p = 0.003]. Relapse was associated with a substantially increased risk of death, with a 5.23-fold higher hazard compared with patients without relapse [HR = 5.23, 95% CI (1.57, 17.42), p = 0.007]. Other factors—including age, sex, indication for HCT (malignant vs. non-malignant), donor source, indication for CC admission, histocompatibility, and development of graft-vs.-host disease (GVHD)—were not significantly associated with mortality. Among CC admitted patients, those who required CC within the first 30 days of HCT had an overall survival (OS) rate estimated from the KM curve of 39.9% [95% CI: (18.5%, 60.5%)], corresponding to 13 deaths, while patients who required CC after 30 days of HCT had 34.0% [95% CI: (8.2%, 62.8%)], corresponding to 11 deaths (logrank p-value = 0.41). The estimated survival rate at 100 days post-HCT was 56.5% [95% CI: (34.3%, 73.8%)] in the early group and 81.0% [95% CI: (56.9%, 92.4%)] in the late group (logrank p-value = 0.053). The estimated survival rate at 1 year post-HCT was 52.2% [95% CI: (30.5%, 70.0%)] in the early group and 56.1% [95% CI: (32.5%, 74.3%)] in the late group (logrank p-value = 0.45).

Among patients that required CC support (n = 44), 32 infants (32/44 = 72.7%) survived to CC discharge. At one-year post-HCT, survival rate of infants that required CC support was 54.2% [95% CI: (38.4%, 67.6%)], corresponding to 20 deaths as compared to 87.1% [95% CI: (75.9%, 93.3%)] survival rate of infants who did not require CC support, corresponding to 8 deaths out of 63 (Figure 1B).

In the second Cox regression model including infants who required CC admission, recovery of ANC during CC compared with before CC admission [HR = 4.8, 95% CI (1.94, 11.87), p < 0.001], IMV use [HR = 10.04, 95% CI (2.96, 34.02), p < 0.001], and duration of IMV [HR = 1.03, 95% CI (1.01, 1.05), p < 0.001] were significantly associated with an increased risk of death. NIV use during CC [HR = 0.18, 95% CI (0.05, 0.60), p = 0.005] and early indication for CC admission (2013–2023) [HR = 0.41, 95% CI (0.17, 0.96), p = 0.041] were significantly associated with a decreased risk of death.

Discussion

This current study adds important and more specific prognostic guidance for the care of critically ill infants undergoing HCT. A retrospective analysis by the Center for International Blood and Marrow Transplant Research (CIBMTR) and the Virtual Pediatric Systems (VPS), report 8% cumulative incidence of CC admission within the first 100 days post-HCT among 6995 pediatric patients (≤21 years) who underwent first allogeneic HCT between 2008 and 2014 across 69 centers in the United States or Canada (16). In this current report, limited to infants undergoing HCT, a substantially higher rate of CC admission was observed (41%) compared to cohorts across the entire pediatric age spectrum (16, 23–25). These findings are important as it highlights specific resource utilization needs required to adequately anticipate and support infants undergoing HCT. Informed consent for HCT should discuss the likelihood of CC admission and when possible, meeting the CC team and/or touring the CC area may help alleviate caregiver anxiety when CC is needed (26). To assist in early recognition of clinically deteriorating infants and because normal ranges vary even among infant ages, transplant units that infrequently care for infants may benefit from review of age and patient specific normative ranges for early warning system recognition of at-risk patients, prior to each admission (27, 28).

In the CIBMTR/VPS cohort, 85% of HCT patients across the pediatric age spectrum survived to first CC discharge and 52% were alive at 1-year post-HCT (16). In this current study, 72.7% of infants survived to CC discharge and 54.5% were alive at 1-year post-HCT. While infants undergoing HCT have disproportionately higher rates of CC admission, they appear to have similar CC outcomes as older pediatric patients (16). While prognostication of infants undergoing HCT remains difficult, it is important that they be supported as they show similar survival as older pediatric patients who require CC support.

In a multicenter study investigating the kinetics of neutrophil recovery among 125 children undergoing HCT who required IMV, patients who recovered neutrophils at any time during their CC course had better survival than patients who never recovered neutrophils (29). Importantly, patients who had already achieved neutrophil recovery at the time of IMV fared worse than patients who recovered on the ventilator (29). In this current study, neutrophil recovery was in fact an important prognostic indicator among critically ill infants, with significant increase in mortality with each additional day to neutrophil recovery. While no definitive conclusions may be drawn regarding prolonged IMV support, findings from this study may assist multi-disciplinary providers in goals of care discussions. In a retrospective multicenter study of pediatric HCT recipients with acute respiratory failure, patients ventilated longer than 15 days had higher odds of death. Among infants in this current study, each additional day of IMV was associated with a 3% increased risk of mortality. While goals of care are dynamic and should be frequently assessed, this information may help guide discussions, in particular, after prolonged IMV. Rowan et al, reported higher mortality among pediatric HCT recipients who received NIV prior to IMV. Among infant HCT recipients in this single center study we did not find this see the same association (30). Patients who required CC admission between 30 and 100 days post HCT appeared to have better 100 -day survival (81% vs. 56.5%) as compared to those who required CC admission within the first thirty days, but this finding was not statistically significant (logrank p-value = 0.053). This observation should be studied in a larger prospective cohort and likely reflects differing pathophysiology driving underlying critical illness and host immune status. Although our cohort included a relatively small proportion of patients with malignant diseases (9.3%), relapse remained a significant predictor of inferior survival, with relapsed patients experiencing substantially worse outcomes than those who did not relapse (HR 5.23).Markedly significant improvements in survival were observed among infants who underwent HCT more recently and CC admitted compared to before 2013, with a 59% lower risk of early mortality. Indeed, recent implementation of more comprehensive newborn screening programs have improved time to diagnosis and time to HCT which presumably has led to decreased pre-HCT co-morbidities among infants (6, 31, 32). Likewise, advancements in pre-HCT management continue to positively impact prognosis among infants with leukemia (33). Further, in the last decade, there have been remarkable advancements in (i) HCT technique, donor selection and GVHD management (ii) anti-microbial regimens and (iii) CC and advanced life support technologies (34). This point highlights the importance of multi-disciplinary rounds and care conferences to ensure all care teams are up-to-date on expected outcomes in the context of a rapidly changing treatment landscape (35).

While high rates of CC admission were observed among infants undergoing HCT, CC survival is similar to older pediatric patients and at one-year post-HCT and no clear differences in late post-HCT complications were observed between infants who required CC support and those that did not (16). While longer term follow-up is needed among survivors, these findings suggest that despite high rates of CC admission, infants should be supported as aggressively as older pediatric HCT patients.

This study is limited by its retrospective design and lack of longer-term available data. As many patients returned to their local centers, comprehensive follow—up assessments were not available to investigate longer-term functional outcomes. Likewise, more granular analysis of CC outcomes was not possible, especially among patients who underwent HCT in earlier decades prior to current electronic medical record systems. Additionally, the availability of newborn screening at our center and use of UCB as a common donor source in infants may limit the generalizability of these findings, across global health-systems. Finally, given the small sample size, multivariable analysis was limited and we acknowledge the limited sample size as a potential constraint on the stability and power of the adjusted models.

While infants undergoing HCT have a disproportionately higher risk of requiring CC, OS is high and continues to improve, likely due to advancements in HCT techniques, CC and supportive care. Multi-disciplinary goals of care with the patients' caregivers, should be guided by chances of recovery from critical illness, which appear similar to older pediatric patients. Future prospective studies among infants undergoing HCT should guide strategies to reduce transplant related toxicity, need for CC and improve CC outcomes and OS.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Avinash Prakash, All India Institute of Medical Sciences Nagpur, India

Reviewed by: Hiremath Sagar, Narayana Health, India

Indira Jayakumar, Apollo Proton Cancer Centre, India