BCNU-associated pneumonitis: portrait of a toxicity

Pulmonary toxicity has historically been a frequent complication of high-dose therapy with autologous stem cell transplant (ASCT), particularly with conditioning regimens containing 1,3-bis[2-chloroethyl]-1-nitrosurea (BCNU, or carmustine). Numerous studies in lymphomas, malignant gliomas and breast cancer have established a strong correlation between BCNU dose and pulmonary toxicity. The incidence of pulmonary toxicity is <  10% when cumulative BCNU doses are <  800 mg/m² as a single agent [1], but in combination with other cytotoxic drugs, particularly cyclophosphamide, the tolerated dose is lower. Several studies have demonstrated that doses of 600 mg/m² or higher are associated with an increased risk of pneumonitis [2–4], and for many years 450 mg/m² has been considered the upper dose limit for BCNU and is the standard dose used in the CBV regimen (cyclophosphamide, BCNU, etoposide).

Despite decades of experience with high-dose BCNU in the ASCT setting, many aspects of BCNU-associated lung injury remain unclear, including the precise safe dose limit, the predisposing risk factors and the pathophysiology. Thus, studies attempting to elucidate these questions, such as the one by Lane and colleagues in this issue of Leukemia and Lymphoma [5], are welcomed. Before initiating any retrospective analysis on this condition, however, it is important to define its characteristics, in view of the diverse spectrum of lung complications encountered after hematopoietic transplant. A recent statement [6] from the American Thoracic Society is helpful in this regard. The most common non-infectious lung injury in the ASCT setting shares several common features with the idiopathic pneumonia syndrome (IPS) that occurs after allogeneic hematopoietic transplant: pulmonary symptoms (dyspnea or cough), scattered or diffuse pulmonary infiltrates on imaging, evidence of abnormal pulmonary physiology and, most commonly, diffuse alveolar damage on histology, all in the absence of infectious, cardiac, renal or iatrogenic causes [7,8]. This ASCT-associated condition has been described as “delayed pulmonary toxicity syndrome” (DPTS) [9], and is distinct from IPS since it occurs later (median of day 45 vs. 19), has a much better prognosis (<  10% mortality rate, compared with 60–80%), is associated with specific drugs (BCNU, cyclophosphamide and cisplatin) and is highly responsive to corticosteroid therapy.

The study by Lane and colleagues retrospectively examines the incidence and risk factors of pneumonitis in a population of patients with lymphoma treated at two centers with high-dose CBV and ASCT [5]. The authors appropriately designate pulmonary symptoms and diffuse radiologic infiltrates in the absence of infection as the cornerstones of their definition of pneumonitis. A more rigorous definition might also have included a decrease in D_LCO (diffusing capacity of the lung for carbon monoxide), although the more minimalist classification may be adequate, since the selected patients did exhibit a 31% median reduction in D_LCO. The strengths of the study include well-defined inclusion criteria and an analysis limited to a relatively homogeneous group of patients treated with a uniform conditioning regimen, thus minimizing potential confounding factors.

The most salient observation from this study is that even within the circumscribed body surface area (BSA)-based dose of 450 mg/m², higher absolute BCNU doses appear to carry an increased risk of pneumonitis. Specifically, there is a suggestion of a threshold effect at BCNU doses of 1000 mg or more. If this is a true phenomenon, it may explain some of the apparent heterogeneity in risk thresholds determined by BSA-based dosing in other studies. In addition to BCNU dose, the authors confirmed previous observations that prior mediastinal radiation is a risk factor for delayed pulmonary toxicity following ASCT [4]. Contrary to previous reports, female gender was not a risk factor for BCNU-induced pneumonitis in the current study [10,11]. One surprise in the study by Lane and colleagues was that younger age appeared to correlate with increased pneumonitis risk, a finding with no readily apparent explanation and contrary to other published reports.

The delayed median onset of symptoms (50 days post-ASCT) is beyond the typical window of 30–40 days post-ASCT when most patients are discharged from transplant centers, and thus the majority of cases of DPTS are diagnosed by the referring oncologist. This study therefore serves as a useful reminder of the frequency (22%) of this complication, which is likely underappreciated by providers at transplant referral centers. The potential for DPTS should be discussed during the consent process with all patients planned for BCNU-containing regimens. Additionally, referring providers should be alerted to watch for DPTS symptoms and be prepared to institute corticosteroid therapy once lower respiratory tract infection has been excluded.

The larger follow-up study planned by this group in conjunction with the Center for International Blood and Marrow Transplant Research may answer some of the questions raised by this study, including whether younger age is truly a risk factor, and more important, whether the 1000 mg threshold effect is reproducible in a larger population receiving other conditioning regimens. If the threshold is confirmed, we should strongly consider capping the BCNU dose in the CBV regimen at 1000 mg or using alternative regimens such as BEAM (BCNU, etoposide, Ara-C, melphalan; which uses a BCNU dose of 300 mg/m²), particularly in patients previously treated with mediastinal irradiation. While the results from Lane and colleagues are reassuring in that the vast majority of patients who develop CBV-related pneumonitis recover, 3% of all patients receiving CBV died of this complication, which may be largely avoidable with proper limits on BCNU dosing. Additionally, more information about long-term recovery of lung function in these patients is needed.

Finally, a critical unanswered question is: what is the underlying biological mechanism? The direct association with BCNU dose suggests tissue injury, yet the median onset of 6–7 weeks is beyond the typical window of acute tissue injury and inflammation. The delayed onset, temporally correlated with lymphocyte recovery, as well as the responsiveness of the condition to corticosteroids, suggests a possible immune-mediated mechanism. Further research is needed to elucidate the biology of this condition, which may lead to new preventive or therapeutic options.

Potential conflict of interest:

Disclosure forms provided by the authors are available with the full text of this article at www.informahealthcare.com/lal.