Critical care in hematopoietic stem cell transplantation: Common complications and management

Abstract

This narrative review synthesizes contemporary evidence regarding critical care complications following hematopoietic stem cell transplantation (HSCT) and evaluates management strategies for improving outcomes in this complex population. We conducted a comprehensive literature search of MEDLINE/PubMed databases from January 2014 to December 2024, focusing on critical care complications, prognostic factors, and therapeutic interventions in HSCT recipients. Intensive care unit (ICU) admission affects 8.8%-40% of HSCT recipients, with contemporary 30-day survival reaching 57.7% in specialized centers. Respiratory failure predominates as the primary indication for ICU support. The Lung Injury Prevention Score for Bone Marrow Transplant demonstrates strong predictive accuracy for acute respiratory distress syndrome development. Novel therapeutic approaches show promise, including inhaled tranexamic acid protocols for diffuse alveolar hemorrhage showing promising hemostasis rates in cohort studies, though randomized controlled trial data are lacking. JAK inhibitors achieve improved response rates compared to best available therapy for steroid-refractory graft-versus-host disease (GVHD). The Mount Sinai Acute GVHD International Consortium algorithm provides validated biomarker-based prognostication. Post-transplant cyclophosphamide reduces acute GVHD incidence, while complement inhibition improves outcomes in transplant-associated thrombotic. Early recognition using validated scoring systems and integration of standard ICU protocols with specialized HSCT expertise are essential for optimizing outcomes. Despite advances, significant knowledge gaps remain regarding optimal management strategies for many complications, with most evidence derived from retrospective cohort studies.

Key Words: Hematopoietic stem cell transplantation; Intensive care unit; Respiratory failure; Graft-versus-host disease; Acute respiratory distress syndrome; Idiopathic pneumonia syndrome; Diffuse alveolar hemorrhage; Transplant-associated thrombotic microangiopathy

Core Tip: Critical care management of hematopoietic stem cell transplantation recipients has evolved from historically poor outcomes to contemporary 30-day intensive care unit (ICU) survival approaching 60% in specialized centers. Respiratory failure remains the predominant complication requiring ICU support. Integration of validated prognostic tools including the Mount Sinai Acute graft-versus-host disease International Consortium algorithm probability with targeted therapies and early intervention strategies represents the current standard of care for optimizing outcomes in this vulnerable population.

INTRODUCTION

Hematopoietic stem cell transplantation (HSCT) is a cornerstone therapeutic intervention for numerous malignant and non-malignant hematologic disorders, with over 50000 procedures performed annually worldwide and increasing utilization in patients over 70 years of age[1]. Despite significant advances in transplant techniques, conditioning regimens, and supportive care, HSCT remains a high-risk procedure associated with substantial morbidity and mortality, with recent real-world data demonstrating HSCT-comorbidity index scores ≥ 4 independently predicting both intensive care unit (ICU) admission and reduced overall survival[2]. A considerable proportion of patients, ranging from 8.8% to 40% depending on transplant type and center-specific factors, require ICU admission during the peri-transplant period, with temporal analysis revealing stable ICU utilization rates of 23% despite progressively older recipient populations increasing by 1.64 years annually[3].

Over the past two decades, the landscape of HSCT critical care has evolved dramatically. Historical data from 1996 documented dismal survival rates, with ICU mortality approaching 80%-90% for patients requiring mechanical ventilation[4]. These outcomes led to widespread pessimism regarding ICU care for HSCT recipients and resulted in many centers adopting restrictive ICU admission policies. However, contemporary studies have demonstrated marked improvements in survival, with recent multicenter analyses reporting 30-day survival rates of 57.7% overall and hospital survival rates exceeding 60% in cohorts utilizing advanced prognostic scoring systems, including dynamic Sequential Organ Failure Assessment (SOFA) score evaluation between days 1-3[5,6].

Several factors have contributed to this paradigmatic shift in outcomes. The increased utilization of reduced-intensity conditioning regimens has decreased early treatment-related mortality while expanding transplant eligibility to older and more comorbid patients, although myeloablative conditioning with total-body irradiation remains a significant predictor of acute respiratory distress syndrome (ARDS) development[7,8]. Advances in antimicrobial prophylaxis and treatment have significantly reduced the incidence of life-threatening infections, particularly cytomegalovirus (CMV) pneumonia, with letermovir prophylaxis reducing clinically significant CMV infection despite emerging resistance mutations[9]. Enhanced understanding of transplant-specific complications has facilitated earlier recognition and targeted interventions[10-12]. Figure 1 summarizes the contemporary landscape of HSCT critical care, including epidemiological outcomes, temporal patterns of complications, and emerging therapeutic advances.

Figure 1 Contemporary landscape of critical care complications following hematopoietic stem cell transplantation. HSCT: Hematopoietic stem cell transplantation; ICU: Intensive care unit; GVHD: Graft-versus-host disease; TA-TMA: Transplant-associated thrombotic microangiopathy; TXA: Tranexamic acid; DAH: Diffuse alveolar hemorrhage; MAGIC: Mount Sinai Acute GVHD International Consortium.

This narrative review was conducted to synthesize current evidence regarding critical care complications following HSCT. We performed a comprehensive literature search using MEDLINE/PubMed, EMBASE, and Cochrane databases covering the period from January 2014 to December 2024. Search terms included combinations of "hematopoietic stem cell transplantation", "bone marrow transplantation", "intensive care unit", "critical care", "respiratory failure", "sepsis", "graft-versus-host disease", and specific complications. We included English-language publications reporting on adult and pediatric HSCT recipients requiring critical care support. Priority was given to multicenter studies, randomized controlled trials, systematic reviews, and large cohort studies. Case reports were included only when describing novel therapeutic approaches. Exclusion criteria included non-English publications, abstract-only publications, and studies focused exclusively on outpatient management. This narrative approach was chosen to provide a comprehensive overview of the evolving field rather than a systematic quantitative analysis.

PROGNOSTIC FACTORS AND OUTCOME PREDICTION

Prognostic scoring systems

Accurate prognostication in critically ill HSCT recipients remains challenging owing to the complex interplay of transplant-specific factors, critical illness severity, and individual patient characteristics with heterogeneous outcomes. The SOFA score has demonstrated superior discriminative ability compared to other scoring systems in HSCT recipients, with recent validation in consecutive patient cohorts[5]. Dynamic SOFA scores between days 1-3 show area under the receiver operating characteristic curve (AUROC) of 0.765 for predicting mortality, superior to APACHE II (AUROC 0.722) and surprisingly outperforming the Prognostic Index for Intensive Care After Allogeneic Stem Cell Transplantation (PICAT) (AUROC 0.5687) in contemporary cohorts.

PICAT represents the first transplant-specific ICU prognostic tool validated across multiple centers[5]. PICAT incorporates transplant-specific factors including time from transplant, graft-versus-host disease (GVHD) status with grade III-IV conferring additional points, platelet count < 50000/μL, bilirubin > 2 mg/dL, and organ dysfunction to stratify patients into three risk groups with hospital mortality rates of 34%, 61%, and 91% for low, intermediate, and high-risk groups respectively. However, PICAT's complexity requiring 15 variables and requirement for transplant-specific data limit its widespread adoption with only 41% of centers routinely using it.

The Mount Sinai Acute GVHD International Consortium (MAGIC) has developed and validated biomarker-based prognostic tools that significantly enhance risk stratification[13-18]. The MAGIC algorithm probability (MAP), derived from two serum biomarkers [suppressor of tumorigenicity 2 (ST2) and regenerating islet-derived 3α (REG3α)], measures damage to gastrointestinal crypts during GVHD and predicts non-relapse mortality with high accuracy[13]. Recent validation studies demonstrate that MAP predicts treatment response and long-term outcomes better than clinical assessment alone, with the ability to identify distinct risk groups among both responders and non-responders to therapy[14,15].

Biomarkers for risk stratification

Several biomarkers with immediate clinical applicability have emerged as potentially useful adjuncts to clinical scoring systems for risk stratification in critically ill HSCT recipients. The MAGIC biomarkers ST2 and REG3α have shown particular promise, with the combination generating estimated probabilities of 6-month non-relapse mortality that consistently outperform clinical parameters alone[16,17]. When measured at treatment initiation, these biomarkers separate patients into distinct risk groups with dramatically different outcomes, and dynamic changes in MAP after treatment validate it as a response biomarker[13].

Procalcitonin, while primarily used for infection diagnosis, demonstrates prognostic value when elevated > 2 ng/mL with increased mortality risk, particularly in patients with suspected bacterial sepsis where levels > 10 ng/mL indicate significantly increased mortality[19,20]. Serial procalcitonin measurements with > 80% decrease at 72 hours predicted survival with 89% positive predictive value. Ferritin levels provide important prognostic information, particularly in patients with hyperinflammatory syndromes affecting 15%-30% of critically ill HSCT recipients[21].

EPIDEMIOLOGY AND RISK FACTORS FOR ICU ADMISSION

Incidence and temporal patterns

The incidence of ICU admission following HSCT varies considerably based on transplant type, conditioning intensity, and institutional practices. Recent multicenter studies reported ICU admission rates of 8.8% to 20% for all HSCT recipients, with significantly higher rates among allogeneic transplant recipients (16%-30%) compared to autologous transplant recipients (3.3%-8%)[3,5,22]. Table 1 summarizes the epidemiology, timing patterns, primary indications, organ support requirements, and survival outcomes of ICU admission following HSCT. These rates have remained relatively stable over the past decade despite progressively older patient populations, suggesting that while survival has improved through enhanced patient selection and pre-ICU management protocols, the fundamental risk of critical illness persists.

Table 1 Epidemiology and outcomes of intensive care unit admission following hematopoietic stem cell transplantation.

Parameter	Autologous HSCT	Allogeneic HSCT	Overall
ICU admission rate	3.3%-8%	16%-30%	8.8%-23%
Timing of ICU admission
Early (Days 0-30)	Conditioning toxicity, engraftment syndrome	Conditioning toxicity, SOS, early infections	25%-40% of admissions
Intermediate (Days 31-100)	Infections	Acute GVHD, infections, TA-TMA	30%-45% of admissions
Late (> 100 days)	Late effects	Chronic GVHD, opportunistic infections, BOS	20%-35% of admissions
Primary indications
Respiratory failure	30%-50%	50%-70%	40%-70%
Sepsis/septic shock	25%-40%	35%-50%	30%-50%
Neurological complications	5%-10%	10%-20%	8%-15%
Organ support requirements
Mechanical ventilation	25%-35%	35%-55%	38%
Vasopressor support	20%-30%	30%-50%	25%-45%
Renal replacement therapy	5%-10%	15%-25%	10%-20%
Survival outcomes
ICU survival (30-day)	70%-85%	55%-70%	57.7%
Hospital survival	65%-80%	45%-65%	> 60%
1-year survival	55%-70%	30%-50%	35%-55%

HSCT: Hematopoietic stem cell transplantation; ICU: Intensive care unit; GVHD: Graft-versus-host disease; TA-TMA: Transplant-associated thrombotic microangiopathy.

The temporal distribution of ICU admissions follows predictable patterns related to transplant recovery and immune reconstitution phases. Early admissions (days 0-30) are predominantly related to conditioning regimen toxicity, including sinusoidal obstruction syndrome occurring at a median of 21 days, engraftment syndrome, and early infectious complications during profound neutropenia[19,20,23]. The intermediate period (days 31-100) is characterized by infectious complications as immune reconstitution occurs and acute GVHD develops[9,24]. Late admissions (> 100 days) are typically associated with chronic GVHD complications, opportunistic infections, and late organ toxicities[25,26].

RESPIRATORY COMPLICATIONS

Table 2 provides an overview of major complications requiring ICU support in HSCT recipients, including respiratory, infectious, and systemic complications, with their incidence, timing, diagnostic criteria, management approaches, and mortality rates.

Table 2 Major complications requiring intensive care unit support: Diagnosis and management.

Complication	Incidence	Timing	Diagnostic criteria	First-line management	Established/emerging therapies	Mortality
ARDS	> 15% (allo)	Variable	Berlin criteria + LIPS-BMT score	Low TV ventilation, PEEP optimization	Prone positioning, ECMO	50%-70%
Idiopathic pneumonia syndrome	3.7%	Median 19 days	Multilobar infiltrates, no infection	High-dose steroids	Etanercept (established, multiple trials); JAK inhibitors (emerging, limited data)	Variable
Diffuse alveolar hemorrhage	3%-10%	Median 30 days	Progressive bloody BAL, bilateral infiltrates	Supportive care	Inhaled TXA ± rFVIIa protocols show promise; RCT data limited	56% at day 100
Neutropenic sepsis	Up to 50%	Days 0-30	Fever + ANC < 500	Broad-spectrum antibiotics within 1 hour	Combination therapy for severe sepsis/shock	42.2%
CMV pneumonia	< 6% with prophylaxis	Days 30-100	High viral burden in BAL with clinical-radiologic correlation	Ganciclovir/valganciclovir + IVIG	Resistance monitoring, maribavir for resistant	60%-80%
Invasive aspergillosis	5%-15% (allo)	Variable	Galactomannan + CT findings	Voriconazole or isavuconazole	Combination therapy under investigation	50%-80%
Acute GVHD	39% grade II-IV	Median day 34	Clinical ± biopsy; MAGIC biomarkers (ST2, REG3α)	High-dose steroids	JAK inhibitors (ruxolitinib FDA-approved)	> 50% severe
Sinusoidal obstruction syndrome	5%-60%	Median day 13	EBMT 2023 criteria	Supportive care	Defibrotide (FDA-approved)	20%-80% severe
TA-TMA	10%-40% (allo)	Median 86 days	Clinical + lab criteria	Withdraw/reduce CNI	Complement inhibition (eculizumab)	30%-60%
Engraftment syndrome	10%-20% auto, 35% allo	During engraftment	Fever, weight gain, infiltrates, rash	Supportive care	Steroids for severe cases	5%-15%

HSCT: Hematopoietic stem cell transplantation; ICU: Intensive care unit; SOS: Sinusoidal obstruction syndrome; GVHD: Graft-versus-host disease; TA-TMA: Transplant-associated thrombotic microangiopathy; BOS: Bronchiolitis obliterans syndrome; ARDS: Acute respiratory distress syndrome; LIPS-BMT: Lung Injury Prevention Score for Bone Marrow Transplant; TV: Tidal volume; PEEP: Positive end-expiratory pressure; ECMO: Extracorporeal membrane oxygenation; BAL: Bronchoalveolar lavage; TXA: Tranexamic acid; rFVIIa: Recombinant activated factor VII; RCT: Randomized controlled trial; ANC: Absolute neutrophil count; CMV: Cytomegalovirus; IVIG: Intravenous immunoglobulin; CT: Computed tomography; MAGIC: Mount Sinai Acute GVHD International Consortium; EBMT: European Society for Blood and Marrow Transplantation; CNI: Calcineurin inhibitor; FDA: Food and Drug Administration.

Acute respiratory failure and ARDS

Respiratory failure represents the predominant indication for ICU admission following HSCT, affecting 40%-70% of critically ill transplant recipients[6]. ARDS affects more than 15% of allogeneic HSCT recipients, with the Lung Injury Prevention Score for Bone Marrow Transplant (LIPS-BMT) score incorporating 22 variables achieving predictive accuracy (AUROC 0.85) for ARDS development and 0.87 for mechanical ventilation requirement[7]. The pathogenesis of ARDS in HSCT patients differs from classical ARDS, involving direct lung injury from conditioning regimens, immune-mediated inflammation, and infectious triggers.

Idiopathic pneumonia syndrome

Idiopathic pneumonia syndrome (IPS) is a distinct form of noninfectious lung injury characterized by widespread alveolar damage without identifiable pathogens following comprehensive microbiological evaluation[12,27-31]. The incidence of IPS is 3.7% within 120 days of allogeneic HSCT, with median onset at 19 days post-transplant. The pathogenesis involves complex interactions between conditioning regimen toxicity and alloreactive T-cell responses mediated through tumor necrosis factor-α pathways[27,28].

Contemporary management of IPS has evolved with both established and emerging therapeutic approaches. High-dose corticosteroids (methylprednisolone 2 mg/kg/day) remain first-line therapy, although response rates have historically been limited to 15%-30%[12]. Etanercept, a soluble TNF receptor fusion protein, has shown significant efficacy in multiple studies. Early pilot studies demonstrated rapid response (median 7 days) with day 28 survival of 73% when combined with corticosteroids[28]. A subsequent randomized, double-blind, placebo-controlled trial (BMT CTN 0403) comparing corticosteroids plus etanercept vs placebo was terminated early due to slow accrual but showed trends toward improved outcomes[29]. Retrospective comparative studies have shown significantly improved survival with etanercept plus corticosteroids compared to corticosteroids alone, with 28-day survival of 88.2% vs 36.4% and 2-year survival of 18% vs 9.1%[30]. Pediatric studies (ASCT0521) demonstrated 71% overall response rate with etanercept therapy[31].

JAK inhibitors, particularly ruxolitinib, represent a promising emerging therapy for IPS, though current evidence is limited to small cohorts and requires validation in larger studies[12]. The combination of ruxolitinib with corticosteroids has shown encouraging results in limited case series, targeting key inflammatory pathways involved in IPS pathogenesis.

Diffuse alveolar hemorrhage

Diffuse alveolar hemorrhage (DAH) represents a serious pulmonary complication of HSCT with incidence rates of 3%-10%[32-36]. Recent multicenter studies have refined our understanding of DAH risk factors and management strategies. Umbilical cord blood transplants carry significantly higher DAH risk compared to peripheral blood or bone marrow grafts, while delayed neutrophil or platelet engraftment represents a major risk factor across all graft sources[32]. DAH typically occurs at median 30 days post-transplant, often coinciding with neutrophil engraftment[33].

Contemporary management has evolved with evidence from multiple studies showing that systemic corticosteroids may be associated with worse outcomes in some cohorts[32,34]. Inhaled hemostatic protocols have emerged as promising alternatives. A standardized two-step protocol utilizing inhaled tranexamic acid (250-500 mg every 6 hours via nebulizer) has shown cessation of pulmonary hemorrhage in 48% of cases as first-line therapy[33,34]. For non-responders, addition of inhaled recombinant activated factor VII (35-50 μg/kg/dose) achieved hemostasis in an additional 41% of patients[35]. In pediatric cohorts, inhaled tranexamic acid achieved 95% hemostasis rate with no major adverse events recorded[34]. A systematic review of management strategies emphasized the need for randomized controlled trials to establish optimal treatment protocols[36]. These approaches require careful monitoring for thrombotic complications, though current studies report no increase in thrombotic events.

INFECTIOUS COMPLICATIONS

Neutropenic sepsis and septic shock

Infectious complications represent a leading cause of ICU admission and mortality in HSCT recipients, reflecting profound immunosuppression inherent to the transplant process[20,37-42]. Neutropenic sepsis affects up to 50% of HSCT recipients with significant associated mortality. Contemporary international guidelines from the German Society of Hematology and Medical Oncology provide evidence-based recommendations for management[37,38]. The pathogenesis involves loss of neutrophil-mediated bacterial clearance combined with mucosal barrier disruption from conditioning regimens.

Contemporary management emphasizes immediate empirical antimicrobial therapy within one hour of sepsis recognition, achieving significant mortality reduction[37-39]. First-line therapy typically consists of anti-pseudomonal beta-lactam antibiotics, with piperacillin-tazobactam or meropenem representing preferred agents[37]. Meta-analyses show that combination treatment with aminoglycosides increases renal toxicity without improving efficacy in neutropenic patients with bacteremia, though retrospective studies suggest potential benefit in severe sepsis and septic shock[38]. The management of neutropenic septic shock follows standard sepsis bundles with modifications for immunocompromised hosts, including aggressive fluid resuscitation targeting mean arterial pressure ≥ 65 mmHg and norepinephrine as first-line vasopressor[37,38].

Multiple prognostic scoring systems have been developed for risk stratification. The Multinational Association for Supportive Care in Cancer risk index and Clinical Index of Stable Febrile Neutropenia help stratify patients into high and low-risk groups[40,41]. The quick SOFA score shows poor sensitivity (0.51-0.60) but higher specificity (0.72-0.83) for mortality prediction in neutropenic patients, while SIRS criteria demonstrate high sensitivity (0.86-0.88) but poor specificity (0.25-0.29)[37].

NON-INFECTIOUS SYSTEMIC COMPLICATIONS

GVHD

Acute GVHD affects 32%-60% of allogeneic HSCT recipients with grade II-IV incidence of 39% and represents a major cause of ICU admission when severe[13-18,24,43]. The MAGIC consortium has established standardized diagnostic and staging criteria that improve consistency across centers[18]. The pathogenesis involves donor T-cell recognition of recipient tissue antigens leading to multiorgan inflammation primarily affecting skin, gastrointestinal tract, and liver.

The MAGIC algorithm incorporates biomarkers ST2 and REG3α to generate probability scores (MAP) that predict non-relapse mortality and treatment response better than clinical assessment alone[13,14]. Recent studies demonstrate that MAP measured at treatment initiation and dynamically during therapy provides superior prognostication, identifying distinct risk groups with significantly different outcomes in both first-line and second-line therapy settings[15,16]. The MAGIC composite response integrating biomarkers with clinical parameters further enhances outcome prediction[17].

First-line treatment consists of high-dose corticosteroids, though response rates are suboptimal. Corticosteroid-refractory aGVHD carries poor prognosis. JAK inhibitors, with ruxolitinib demonstrating improved response rates compared to best available therapy in clinical trials (REACH2 and REACH3), represent the current standard for second-line therapy[24,25]. The higher day 28 response rates and survival observed with ruxolitinib compared to other therapies were limited to patients with low MAP scores, highlighting the importance of biomarker-guided therapy[15].

Sinusoidal obstruction syndrome

Sinusoidal obstruction syndrome (SOS), also known as veno-occlusive disease, is a potentially life-threatening complication affecting 5%-60% of HSCT recipients with wide variability based on conditioning intensity[10,23]. The incidence is higher in pediatric populations (20%-30%) compared to adults. The pathogenesis involves toxic injury to hepatic sinusoidal endothelial cells from conditioning regimens, leading to sinusoidal obstruction and portal hypertension.

Clinical manifestations typically develop within the first 21 days post-transplant, though late-onset cases occur. The classical triad includes painful hepatomegaly, weight gain, and hyperbilirubinemia. Diagnosis relies on clinical criteria, with the 2023 European Society for Blood and Marrow Transplantation refined criteria improving diagnostic accuracy[23]. Defibrotide represents the only Food and Drug Administration (FDA)-approved therapy for severe SOS, demonstrating improved survival in clinical trials[10]. Supportive care focuses on careful fluid management and managing complications of portal hypertension.

Transplant-associated thrombotic microangiopathy

Transplant-associated thrombotic microangiopathy (TA-TMA) affects 10%-40% of allogeneic HSCT recipients with increasing recognition as automated screening improves detection[11,44]. The pathogenesis involves complement dysregulation and endothelial injury, leading to microangiopathic hemolytic anemia, thrombocytopenia, and organ dysfunction. Clinical manifestations include acute kidney injury, hypertension, and neurological symptoms.

Diagnosis requires integration of laboratory findings with clinical evidence of organ dysfunction. Several diagnostic criteria have been proposed, though lack of consensus complicates diagnosis. Management focuses on addressing precipitating factors and providing organ support. Discontinuation or dose reduction of calcineurin inhibitors represents a key intervention, though this must be balanced against GVHD risk. Complement inhibition with eculizumab shows promise in recent studies, particularly in pediatric cohorts with appropriate dosing achieving 88% complete response[11].

CONTEMPORARY ADVANCES AND FUTURE DIRECTIONS

The landscape of HSCT critical care continues to evolve rapidly, driven by advances in transplantation techniques and novel therapeutic approaches[45-51]. Post-transplant cyclophosphamide has emerged as an effective GVHD prophylaxis strategy, demonstrating reduced acute GVHD rates, although requiring careful monitoring for delayed engraftment[52].

Cell therapy innovations

Mesenchymal stromal cells (MSCs) have shown substantial promise for steroid-refractory GVHD treatment[45-51]. Multiple studies demonstrate overall response rates of 40%-72% in steroid-refractory acute GVHD, with complete response in approximately 30%-50% of patients[45,46]. A recent study of 86 patients with grade III-IV steroid-refractory aGVHD treated with umbilical cord-derived MSCs showed 52.3% overall response at day 28, with 27.9% achieving complete remission[46]. Long-term follow-up studies report sustained safety, with 2-year survival of 60% in recent trials of induced pluripotent stem cell-derived MSCs[51]. Meta-analyses confirm acceptable safety profiles with no increase in infection or relapse rates[47,48].

The FDA has now approved at least seven chimeric antigen receptor T-cell products for hematologic malignancies, including ABECMA, BREYANZI, CARVYKTI, KYMRIAH, TECARTUS, YESCARTA, and AUCATZYL[53]. These represent expanding therapeutic options for relapsed/refractory malignancies, though they carry their own risks of cytokine release syndrome and neurotoxicity requiring ICU management.

Regulatory T-cell therapy for GVHD prevention and treatment is under investigation, with early studies showing reduction in acute GVHD incidence to < 10% in haploidentical transplants when used prophylactically[13]. These cellular therapies represent a paradigm shift in managing transplant complications.

Artificial intelligence and precision medicine

The integration of artificial intelligence and machine learning into HSCT critical care represents a transformative development[43,54]. Predictive modeling for ARDS development has advanced with tools like LIPS-BMT enabling proactive interventions. Machine learning algorithms incorporating transcriptomic data can identify patients at high risk for specific complications, enabling targeted prevention strategies. Real-time risk stratification using continuous physiologic monitoring combined with AI analytics may enable earlier identification of clinical deterioration.

LIMITATIONS

This narrative review has several limitations that should be acknowledged. The evidence base consists predominantly of retrospective cohort studies with inherent selection bias and heterogeneity between centers. The paucity of randomized controlled trials in this population reflects both ethical and practical challenges of conducting such studies in critically ill HSCT recipients. Management approaches vary considerably between institutions, limiting generalizability of outcomes data. Additionally, rapid evolution of transplant techniques and supportive care means that historical comparisons may not reflect current practice. The complexity of this patient population, with multiple concurrent complications and interventions, makes it difficult to isolate the impact of specific therapeutic approaches. Future prospective multicenter studies are needed to validate emerging therapies and establish standardized management protocols.

CONCLUSION

Critical care management of HSCT recipients has evolved from historically poor outcomes to contemporary survival rates that justify aggressive supportive care in appropriately selected patients. The epidemiology reflects ongoing changes in transplant practices, with respiratory failure remaining the predominant indication for ICU admission. Early recognition of complications using validated tools including the MAGIC algorithm probability and prompt initiation of supportive care are essential. The integration of standard ICU protocols with transplant-specific considerations has improved outcomes. Novel therapeutic approaches including etanercept for IPS, inhaled hemostatic therapies for DAH, and cellular therapies for GVHD offer promise for further improving outcomes, though many require validation in larger studies. Several key research priorities emerge including development of transplant-specific prognostic tools, comparative effectiveness research for specific complications, and quality of life research focusing on ICU survivors. The future lies in continued integration of transplant immunobiology with critical care medicine, supported by emerging technologies and precision medicine approaches.