Acute Respiratory Distress Syndrome following Hematopoietic Stem Cell Transplantation: One More Piece in the Puzzle

Critical illness across patients with cancer no longer means end of life. Over time, oncologic critical care outcomes have witnessed one of the greatest reductions in mortality when compared with other vulnerable populations (1). These improvements are attributable to advancements in cancer care, infectious disease, and critical care practices (2). Critical care mortality trajectories within oncology continue to improve with time as both fields venture further into precision medicine and bio-phenotyping (3–5).

Hematopoietic cell transplantation (HCT) is a potentially curative therapy for patients with malignant and nonmalignant hematologic disorders. However, patients undergoing HCT have remained especially vulnerable in the face of critical illness, particularly acute respiratory distress syndrome (ARDS) (6, 7). Indeed, historic guidelines had advised against intensive care unit admission for this population based on early data highlighting excessive mortality (8, 9). The reasons for the higher mortality are likely multifold and in part owing to mechanistic differences in ARDS compared with the general population. Although neutrophil response plays a central role in the development of ARDS, Braude and colleagues first described how neutropenic HCT recipients could develop ARDS (10), highlighting diverse pathophysiology underlying ARDS. ARDS in HCT recipients occur frequently outside of the neutropenic period with unique causes that occur at distinct time points posttransplant. Understanding these mechanistic differences within HCT as time posttransplant evolves and between HCT and non-HCT recipients is critical to ultimately improve outcomes.

As critical care and oncologic practices evolved over time, the HCT population similarly experienced improvements in mortality (5, 11). A greater understanding of ventilator-associated lung injury, the role of noninvasive oxygen strategies, infection identification, and targeted antimicrobial therapies were likely major factors impacting outcomes in this population. Unfortunately, however, this cohort has not experienced the same quantum of improvement compared with other oncology subtypes (3). In fact, recent data show a plateauing of critical care–associated mortality compared with other oncology subgroups in which mortality continues to improve (3).

In the face of this deceleration, as intensivists, we are left asking ourselves: Can we move the needle to further ameliorate outcomes? Is the mortality plateau because of 1) our limited understanding of the pathophysiology of HCT–ARDS, 2) the complex nature of their underlying hematologic malignancy, 3) the impact of immunosuppressive treatments, or 4) a greater susceptibility to the harms associated with critical care (i.e., ventilator-associated lung injury, secondary infections, etc.)? Do we need a complete paradigm shift in our categorization and approach to respiratory failure and ARDS in this cohort? The right answer likely involves a combination of all these factors. In the crusade to improve HCT–ARDS outcomes, dedicated and rigorous studies evaluating this complex subgroup are needed.

In this issue of AnnalsATS, Herasevich and colleagues (pp. 1004–1012) report the results of a nested case-control study that sought to describe risk factors of ARDS following a first-time HCT using variables available before transplant (12). Among 3,920 patients who underwent HCT, 4.5% developed ARDS within 1 year of transplant. The analytic cohort consisted of 164 ARDS cases matched to 492 non-ARDS control subjects. Among patients that developed ARDS, 65% were categorized as severe, and the majority (69%) occurred in the first 100 days following the transplant. The underlying etiology for ARDS was classified as infectious (54%), peri-engraftment (16%), diffuse alveolar hemorrhage (2%), or undetermined (17%). The authors identified several pretransplant variables that were associated with the development of ARDS: worse pulmonary function, poor performance status, abnormal laboratory tests (albumin, aspartate aminotransferase, hemoglobin, platelets, and leukocytes), specific chemotherapeutic agents, radiation therapy, higher need of blood products, and prior hospital admission (requiring mechanical ventilation). All-cause mortality at 1 year was 70% in the ARDS cohort.

These risk factors have face validity and seem to perform well in different types of HCT (allogeneic and autologous) and for different time periods (before and after 100 d following HCT). The authors highlight the notion that given the high mortality rates observed across HCT–ARDS, a focus on prevention might be desirable. This study sheds light on which patients are more likely to develop this devastating complication with factors that are easily identifiable prior to transplant. The extent to which these variables are mere predictors or are causally associated with the development of ARDS cannot be clearly depicted from this analysis; however, the study establishes the groundwork for future research evaluating the pathophysiologic processes that render them more susceptible to ARDS. It also identifies a high-risk cohort in which the risk of ARDS can be discussed together with other risk factors for transplant-related morbidity and mortality at the time of initial consultation. This may also potentially inform advanced care planning discussions (13). The authors should be commended for putting together one of the largest contemporary cohorts of patients with HCT and describing early factors associated with the subsequent development of ARDS.

The complex puzzle of respiratory failure following HCT has multiple edges that still need to be further characterized. Although Herasevich and colleagues’ work addresses the question of early identification of those at risk of ARDS, much work needs to be done in parallel to understand the mechanisms of ARDS in this population. The impact of immunosuppressive therapy in allogeneic transplant recipients and immunologic confounders such as graft versus host disease would also warrant further examination. Future research should further explore the physiologic links between these factors and ARDS and, perhaps most importantly, targeted interventions—when feasible—to decrease mortality.

Over the past 2 decades, significant efforts were made to achieve a common, simple, and pragmatic definition of ARDS to improve early identification (14). This, in turn, was meant to help with the adoption of life-saving evidence-based strategies that could be generally applied to a wide population (pressure and volume limited ventilation and prone positioning) (14, 15). However, have we gone too far in the oversimplification of ARDS—particularly across complex populations such as HCT? A recent observational study has even suggested that ARDS Berlin-severity categories defined by arterial oxygen tension/pressure to fraction of inspired oxygen thresholds may not carry the same prognostic weight in immunocompromised cohorts as it does in nonimmunocompromised patients (16). The HCT population has distinct time periods that correlate with distinct mechanisms of immunosuppression that render the recipient susceptible to unique causes of respiratory failure. Until we have a much clearer understanding of mechanisms, we may not be able to ameliorate mortality across those who develop ARDS. In this end, could the description of subphenotypes facilitate the depiction of the diverse respiratory syndromes occurring after HCT? Are there specific clinical and biologic factors that help differentiate infectious versus noninfectious complications? Do the same principles of management of ARDS apply to this specific population? How does neutropenic ARDS differ from nonneutropenic ARDS? And, finally, what is the impact that ARDS and mechanical ventilation have on short- and long-term outcomes in HCT recipients?

The study by Herasevich and colleagues is an important piece in the puzzle of respiratory failure following HCT and hopefully serves as a trigger for future research in the field. However, we must persist with a multipronged approach and continue to expand our understanding across the entire disease process. We must dive deeper back into pathophysiologic mechanisms of ARDS across this specific population in order to optimally understand the disease process driving mortality. In this case, less may not be more.