The diagnosis of drug-induced lung injury is one of exclusion, as the histologic findings are nonspecific, and the manifestations may overlap with underlying rheumatologic disease or radiation treatment. Although there is variability depending on the drug, there is no consistent relationship with exposure, both dose and interval between onset of lung symptoms and onset of therapy. Furthermore, many patients are on multiple medications, and it may be hard to discern the specific toxic agent. A presumed diagnosis is made only when infections, heart disease, and collagen vascular disease are excluded, and the lung findings resolve with cessation of drug. In only rare instances is the diagnosis confirmed by rechallenge with the offending agent to confirm the association between exposure and pulmonary manifestations.

The radiologic findings are generally those of diffuse disease, with symmetric reticular densities, nodules, and ground-glass opacities, depending on the pattern of lung injury. Nodular opacities occur with the organizing pneumonia pattern and with hypersensitivity-like reaction. Rarely, asymmetric large nodules can mimic metastatic tumor.

The major histologic findings are those of acute lung injury, interstitial inflammation, organizing pneumonia, and nonspecific interstitial pneumonia (Table 59.1). Less commonly, there is an eosinophilic pneumonia. Diffuse alveolar hemorrhage is generally diagnosed by the finding of red blood cells in bronchoalveolar lavage fluid and is an uncommon manifestation of drug-induced lung injury (see Chapter 32).

Because tissue sampling is usually performed by transbronchial biopsy, there is limited tissue, and the histologic findings may not be representative. In addition, there is a spectrum of findings spanning acute alveolar injury, chronic interstitial inflammation, and organizing pneumonia, which may represent manifestations of the same process at different phases of injury. Furthermore, nonspecific interstitial pneumonia is difficult to diagnose without open lung biopsy and is often diagnosed on smaller biopsies based on the findings in chronic interstitial inflammation with reactive pneumocytes (a finding that may be present in organizing acute lung injury). Therefore, transbronchial biopsy should exclude an infectious or neoplastic cause for the patient’s lung disease and document specific histologic findings that indicate acute lung injury (intra-alveolar fibrin with hyaline membranes) or organizing phase of injury (organizing pneumonia). Interstitial inflammation with or without reactive pneumocytes should be reported as a nonspecific finding, with a comment that correlation with imaging is necessary to suggest a diagnosis of nonspecific interstitial pneumonia. There is little documentation of histologic findings in end-stage fibrotic lung disease due to drugs, but the imaging data in these rare patients suggest a nonclassifiable type of fibrosis that is typical neither of nonspecific interstitial pneumonia nor of usual interstitial pneumonia.

Pulmonary function tests generally show decreased lung volumes and diffusing capacity for carbon monoxide. Treatment is drug cessation with or without prednisone.¹

Acute lung injury secondary to drug exposure typically occurs within weeks of treatment and may clinically mimic acute pulmonary edema. The patient typically presents with dyspnea, cough, and decreased diffusing capacity for carbon monoxide, with variable fever. Imaging by computed tomography shows bilateral opacities, usually more prominent in the mid- and lower lung fields, with diffuse, ground glass. Transbronchial biopsies show acute lung injury, with hyaline membranes, thickened alveolar septa with chronic interstitial inflammation, and reactive type II pneumocytes. There may be organization with loose interstitial fibrosis that may overlap with an organizing pneumonia pattern (Figs. 59.1, 59.2, 59.3). It is important to document a lack of organisms by stains for acid-fast bacilli, fungi, and bacteria. The acute exudative phase, which occurs during the first week, is characterized by intra-alveolar fibrin and alveolar septal edema, and the later phase by interstitial fibroblasts are reactive alveolar pneumocytes. In reported fatal cases, histologic findings are typically intense interstitial inflammation with early organization.^2,3

Expansion of interstitium by chronic inflammation is a common finding in drug reactions and overlaps with nonspecific interstitial pneumonia, if there is mild interstitial fibrosis, and organizing diffuse alveolar damage, if there is marked pneumocyte atypia and interstitial edema.⁴ Interstitial inflammation is also common in organizing pneumonia and in itself is not a specific pattern of lung injury. If the only finding in a small lung biopsy, concomitant bronchiolar or peribronchiolar inflammation should be noted and a differential diagnosis based on exclusion of other findings and computed tomographic studies presented. There may be scattered eosinophils in the interstitium in drug reactions, as has been reported in organizing pneumonia secondary to nitrofurantoin.⁵

Organizing lung injury may form nodules with relatively uninvolved surrounding lung (“organizing pneumonia”), may be seen in diffuse chronic inflammation typical of nonspecific interstitial pneumonia, or may manifest as multiple areas in organizing diffuse alveolar damage. The process is more often interstitial than peribronchiolar, although the term bronchiolitis obliterans organizing pneumonia is often (confusingly) applied.⁶ Organizing pneumonia is a common finding in drug toxicity (Figs. 59.4 and 59.5), including that secondary to methotrexate, gemcitabine, amiodarone, penicillamine, and sirolimus.^6,7,8,9,10

The most common type of fibrosing lung disease associated with drug toxicity is nonspecific interstitial pneumonia. The histologic features on small biopsy, however, consist of chronic interstitial inflammation, mild interstitial fibrosis, and reactive pneumocyte hyperplasia, which are features that overlap with organizing diffuse alveolar damage and

which may be seen to some extent in organizing pneumonia. In most cases, the diagnosis of nonspecific interstitial pneumonia is made by imaging. More advanced stages of lung fibrosis are uncommon in drug toxicity and have been reported especially in children treated with chemotherapeutic agents such as carmustine.¹¹ Reports of imaging findings in fibrotic lung disease document dense scars with cystic change and honeycombing, which are not features of nonspecific interstitial pneumonia.^12,13

Eosinophilic pneumonia is a relatively uncommon manifestation of drug toxicity. Conversely, eosinophilic lung disease is caused by drugs in a relatively small proportion of patients (see Chapter 29). Patients typically present with progressive dyspnea, dry cough, and occasionally fever. Peripheral eosinophilia and elevated serum IgE levels are common. Chest radiographs show homogeneous opacities that typically have a peripheral and upper lobe distribution. Biopsies show findings of eosinophilia pneumonia, namely, intra-alveolar fibrin with interstitial and intra-alveolar eosinophils. Organizing pneumonia is a common additional histologic feature. Beta-lactam antibiotics, those that contain sulfa, and daptomycin are relatively common drugs associated with eosinophilic pneumonia.¹⁴

Granulomas have been described in the interstitial infiltrates secondary to nitrofurantoin, sulfasalazine, and sirolimus lung toxicity and may be poorly formed, or discrete and well formed, mimicking sarcoidosis.^15,16,17

Prominent nodules on computed tomographic scans are seen in lung toxicity when the histologic pattern is primarily organizing pneumonia.¹⁸ Less commonly, nodules that can mimic metastatic malignancies on imaging occur as fibrotic masses as a reaction to bleomycin or cyclophosphamide toxicity.^19,20 Inflammatory nodules may be seen in amiodarone toxicity,²¹ and Wegener-like necrotizing granulomas have been described as a reaction to sulfasalazine.²²