on PTU for at least a few weeks can present with any combination of neutropenia or agranulocytosis, necrotic skin changes or gangrene, and pathologically-demonstrable eosinophilic or neutrophilic vasculitis. Similar to levamisole, patients present with violaceous skin changes, vasculitis in the skin, pinna or limbs, and DAH in a subset. Patients typically have elevated ANCA titers. Importantly, ANCA are rarely elevated in untreated Grave’s disease, and these antibodies are strongly associated with therapy with antithyroid drugs of any signi cant duration. Moderately elevated titers of both perinuclear and, less often, cytoplasmic, ANCA develop in 33–50% of patients chronically treated with propylthiouracil or benzylthiouracil. However, the majority of such patients have no symptoms and do not develop clinically evident disease. Propylthiouracil need not be discontinued in such cases, but ANCA titers should be monitored and the patient warned of the possible development of suggestive symptoms and signs, as acute AH or other forms of pulmonary, pleural, pericardial or extrathoracic manifestations of systemic vasculitis may develop after several months of treatment in up to 2.5% of them. ANCA with speci city to more than one lysosomal antigen constitute a distinctive serological pro le for drug-related ANCAs. The p-ANCA exhibit MPO or dual MPOand PR3 staining and anti-­ lactoferrin—neutrophil elastase—cathepsin and/or azurocidin speci city. Titers are also found to be much higher in drug-induced as opposed to naturally-occurring ANCA-­ related disease. Among 250 ANCA-positive cases, 30 (12%) had very high (>12-fold above the median) anti-MPO titers and all of these 30 had clinical features of vasculitis. Ten of these 30 had been exposed to hydralazine, three to propylthiouracil and ve to penicillamine, allopurinol, or sulfasalazine [94]. There was a strong association between the presence of anti-neutrophil elastase and/or antilactoferrin antibodies and exposure to one candidate toxic drug [95]. Therapy with antithyroid drugs, or hydralazine should be sought in cases of vasculitis coexisting with high ANCA titers [96]. From 25% to 60% of patients on chronic propylthiouracil therapy who happen to develop overt ANCA-­ related disease or vasculitis have pulmonary involvement, typically in the form of AH with or without histologically demonstrable capillaritis. Tissue granulomatous and necrotizing­ infammation has been reported. Although some form of renal involvement ranging from microscopic hematuria to necrotizing and/or crescentic glomerulonephritis is present in up to 75% of the patients, renal outcomes are better and mortality is lower in propylthiouracil-related as opposed to idiopathic ANCA-related renal disease. Terminal renal failure is noted in about 5%. Overall mortality is approximately 15%, with a few patients dying from uncontrollable DAH [3, 4]. Patients with the most severe involvement will necessitate induction therapy with corticosteroids, cyclophosphamide, and plasma exchange in a manner simi-

lar to idiopathic vasculitis. Cases with ANCAor mixed ANCA and ANA-­related autoimmune disease have occurred with the use of penicillamine, minocycline, or hydralazine. ANCA titers will fall in most patients upon drug discontinuance, but in a fraction, elevated levels will remain for years without necessarily evidence for disease.

Contrasting with idiopathic lupus, DAH is very unusual in drug-induced lupus (if it truly exists at all).

Anti-basement membrane-related AH (Goodpasture’s like) was considered an idiopathic condition. However, a study of 28 carefully-documented Goodpasture cases showed that 89% of the patients were either smokers, or gave a history of exposure to inhaled cocaine (in 36%), cannabis or heroin [97], raising the possibility that Goodpasture may be triggered by drugs or chemicals.

Such drugs as TNF antagonists, procainamide, levamisole, and interferon have been associated with circulating antiphos- pholipid/antisynthetase-antibodies or syndromes [3, 4].

Transfusion Reactions: TACO–TRALI

Acute noninfectious respiratory reactions following blood transfusion include anaphylaxis, transfusion-associated circulatory overload (TACO), and transfusion-related acute lung injury (TRALI). These entities carry a signi cant risk of respiratory failure.

TACO (see also under iatrogenic cardiac pulmonary edema) has an incidence of 1–8% and develops when the recipient’s circulatory system is overwhelmed by the volume transfused or by the rate at which it is infused to the patient. Poor left ventricular reserve associated with underlying heart failure, aging or diabetes mellitus are risk factors for TACO to develop. TACO manifests with hydrostatic pulmonary edema and is often dif cult to separate from TRALI on clinical and/or imaging. In a recent survey among oncology patients [78], the incidence of TACO was 0.84 per 1000 transfusions, representing 6.6% of all transfusion reactions. This was higher compared with 1–6% in nononcology populations. Among notable risk factors for TACO, hematologic malignancies, receipt of cardiotoxic chemotherapy, preexisting oxygen use, hypertension, renal insuf ciency, daily use of corticosteroids, diuretics, beta-blockers, and elevated NT-proBNP were found. Elevated NT-pro-BNP may help differentiate TACO from TRALI.

TRALI (the term was coined in 1985 [82]) is a form of post-transfusion ARDS that ranks as the leading cause of death from hemotherapy in the US. In the majority of TRALI cases, a demonstrable immune-mediated syndrome is evident. TRALI may occur following transfusion of blood or components including platelets, solvent-detergent plasma, i.v. immunoglobulins (IVIG) or fresh frozen plasma (FFP). Among transfused compounds, platelets demonstrate higher risk. The clinical presentation includes pulmonary in ltrates, hypoxemia, and ARDS (or deterioration of preexisting

ARDS) [98] within 6–8 h of transfusion without evidence for heart failure or TACO [98]. Mechanical ventilation may be required for 12–96 h. Patient-related risk factors for TRALI include an older age, smoking, chronic alcohol abuse, an underlying infammatory condition and elevated IL8 levels, sepsis, mechanical ventilation and with potentially barotraumatic insuffation pressures, recent surgery or trauma or a positive fuid balance. Bloodand transfusion-related factors include the presence in the donor pool of at least one (generally multiparous) female donor, bearing a pathogenic antibody. There is an average vefold increased risk of TRALI if one donor has detectable anti-HLA I or II or anti-HNA3 antibodies at a high titer and with a high af nity for cognate antigens in the recipient. Other factors for TRALI include “shelf age” of the transfused blood, and maternal-to-child transfusion. Although TRALI occurs in 1/5000–7000 transfusions, in ICU, trauma or surgical care settings the incidence is up to 1–5% of those transfused. TRALI mitigation strategies have reduced the incidence in the recent years [82]. Signs and symptoms include dyspnea, hypoxemia, hypotension, fever, transient leukopenia due to granulocyte sequestration in the pulmonary circulation and moderate eosinophilia. A generally-held hypothesis is that TRALI occurs in the context of a “two-hit” process. Sepsis, infammation, mechanical ventilation or surgery causing pretransfusional pulmonary leukocyte sequestration is the predisposition, and transfusion of blood components containing an anti-HLA or HNA antibody is the trigger for full-­ blown TRALI. Accordingly, TRALI is more common in patients with sepsis or following surgery or gastrointestinal bleeding, where incidence can be as high as 15%. Transfer of complement-activating HLA class I or II, granulocyte-­ speci c or lymphocytotoxic antibodies from one or more donors presumably activates neutrophils causing leukosequestration and this is followed by precipitation of sharp-­ edged cholesterol crystals which mechanically injure the pulmonary venules, causing endothelial fenestration, fuid leakage, and pulmonary edema resulting in ARDS [99]. An antibody directed against a cognate antigen of the recipient is identi ed in the donor in about 75% (50–85%) of TRALI cases, which are then designated as immune TRALI. The remainder of cases are labeled nonimmune TRALI, with mechanisms that are less clearly de ned. Redox active lipids formed during blood storage may play a role, explaining why blood with longer shelf life has a propensity to cause TRALI more often than freshly-prepared samples. While the majority of TRALI patients recover in a few days, death from respiratory failure or multiple organ dysfunction occurs in 10–18%. Although recognition of TRALI can clearly guide safer donor selection, prevention has been suboptimal due to poor awareness of the syndrome outside the blood transfusion medicine community until recently. Examination of donor products must be carried out expeditiously, aiming at

the expeditious detection of an antibody, followed by removal of the implicated donor from the pool. In a retrospective study [100], out of ve patients who received multiple transfusions from the same donor, four suffered relapse and only two of eight severe reactions were reported to the blood safety authority. Risk reduction strategies include avoidance of unnecessary transfusions, transfusion of washed components, screening potential donors for antibodies, choosing products from male donors or of female donors without a history of pregnancy and testing negative for antibodies, more proximate donor selection, and increasing the number of donors per sample to dilute any possible antibody [101].

Acute Cellular Interstitial/Infltrative Lung Diseases

Many drugs can cause acute forms of interstitial lung disease with gas exchange characteristics of ARDS (Table 42.6).

Acute Cellular Nonspecifc Interstitial Pneumonia Pattern

Causal drugs include amiodarone, BCG therapy, bleomycin, crizotinib, m-TOR inhibitors (sirolimus, everolimus, and temsirolimus), fudarabine, gols, imatinib, immune checkpoint inhibitors, interferons, lefunomide, methotrexate, nitrosoureas, nitrofurantoin, and pemetrexed among 177 candidate drugs, of which 55 can cause a de nite pattern of cellular nonspeci c interstitial pneumonia on pathology [3, 4]. Recent evidence implicates new oral anticoagulants [102] and SARS COVID 19 vaccination [103, 104] as novel causes. Clinical presentation is with cough, fever, rapidly-­progressive diffuse pulmonary in ltrates with a predilection for denser images in the dependent regions of the lung, i.e., the bases or posterior areas in the erect or supine patient, respectively. Patients can show a prodromal phase of mild ILD where HRCT may disclose discreet haze, early ground glass or a mosaic pattern of attenuation which resembles that of hypersensitivity pneumonitis. The disease can then accelerate with little notice if the drug is continued and sometimes, even following withdrawal, cause respiratory failure requiring ventilatory support and/or ECMO. Imaging discloses a mixture of interand/or intra-lobular septal thickening, crazy-paving, consolidation and air bronchograms, volume loss, and associated pleural effusion. BAL typically shows increased lymphocyte counts, with a CD8+ or, less often, CD4+-dominant subtype pattern. A neutrophilor eosinophil-dominant BAL has also been reported [3, 4]. The BAL pro le is infuenced by timing of the test into the course of the disease, and whether patients have received corticosteroid therapy [105]. The BAL is useful to exclude coincidental or drug-­associated bacterial, viral, fungal, or parasitic infection. It may be dif-cult to separate acute drug-induced-pulmonary in ltrates from Pneumocystis jiroveci pneumonia. It may be especially dif cult to separate true pneumocystis pneumonia from acute drug-induced ILD with pneumocystis colonization when a rt-PCR signal for Pneumocystis is detected and in the

absence of positive stains for organism. Pathology discloses interstitial infammation with dense interstitial mononuclear in ltrate and some degree of interstitial edema [106]. A risk/ bene t analysis for lung biopsy or cryobiopsy is not available, and most physicians will use the BAL data as a surrogate marker and proceed with empiric drug withdrawal with or without corticosteroidand anti-pneumocystis therapy (the latter for those patients who have risk factors: underlying malignancy, connective tissue disease, corticosteroid therapy, recent irradiation, low circulating CD4+ lymphocytes or an rt-PCR signal). Outcome of this form of drug-­ induced condition is good, typically with resolution of all signs and symptoms following drug discontinuance and corticosteroid therapy. Corticosteroid dosage is adjusted to response with tapering over a few weeks or months. High-­ dose methylprednisolone i.v. boluses have become popular in a certain literature from Asia to treat acute drug-induced ILD, but the merit of this form of therapy vs. a more conventional i.v. or oral dosage is not established. Rechallenging the patient with the culprit drug may lead to ILD relapse, but this is not seen in every patient, and is not advisable, as death may follow rechallenge with the drug. Pulmonary brosis following resolution of acute ILD is quite unusual, except in those patients with preexisting rheumatoid lung or idiopathic pulmonary brosis, who may suffer a permanent decrement in lung function after transient drug-induced deterioration and drug withdrawal.

Acute Eosinophilic Pneumonia

Acute eosinophilic pneumonia (AEP) has been reported with exposure to 62 drugs or agents, mainly antibiotics (minocycline, daptomycin), chloroquine, antidepressants, infiximab, mesalazine, NSAIDs, sertraline, cannabis, cocaine, e-­cigarette vapor, heroin, drug excipients, tobacco smoke, incense and radiation therapy [3, 4]. Drugs causing AEP overlap with the 206 drugs capable of causing the more classic and less severe eosinophilic pneumonia. However, the boundary between the two conditions can be blurred clinically. AEP and eosinophilic pneumonia may be the same disease running a different clinical course, with AEP representing the upper end of the severity spectrum. AEP is an acute febrile illness that culminates with hypoxemic respiratory failure, diffuse white-out, pleural effusion, and ARDS. Mechanical ventilation is required in the majority of affected patients, and a few require ECMO. BAL eosinophils above 25% are typical and gures above 50% are common. Blood eosinophilia can be in the normal range in patients who have progressed rapidly or who have received a course of corticosteroids prior to admission. Massive BAL eosinophilia obviates the need for a con rmatory lung biopsy. For the diagnosis to be considered, parasitic or other infections, notably with Strongyloides stercoralis must be ruled out, also because corticosteroid therapy in this context can be

extremely hazardous [107]. Blood/BAL eosinophils in the absence of an infection portend potentially reversible disease. Histopathologic analysis reveals tissue eosinophils on a background of rich mononuclear cell in ltrate and acute and/ or organizing diffuse alveolar damage. Hyaline membranes are unusual [61, 108, 109]. Discontinuation of the drug or of exposure to cigarettes, marijuana or e-cigarette smoke/vapor are the mainstays of management. Corticosteroid therapy is used in the majority or patients. Outcome is generally good and fatalities are the exception.

Eosinophilic granulomatous with polyangiitis (EGPA; a.k.a. Churg-Strauss’) may develop during treatments with 1 of 20 separate drugs [3, 4], notably leukotriene receptor antagonists (LTRA) or omalizumab. To get a full understanding of causation, one has to take into account the increased background rate of EGPA in the asthmatic patient, whether corticosteroid therapy has been tapered or withheld in those who are started on LTRA. There are indisputable cases of LTRA and drug-induced EGPA [110, 111], and withdrawal of LTRA or omalizumab should be considered in the patient who presents with EGPA of new onset.

Acute Granulomatous Interstitial Lung Disease

Forty drugs and families of drugs, including topical BCG therapy in the urinary bladder, interferon, methotrexate, TNF antagonists, ICI, and illicit drugs can cause ILD with a granulomatous component [112]. Although generally mild, this pattern can cause acute respiratory failure or ARDS [3, 4]. On imaging, granulomatous ILD has an established reputation for causing a miliary pattern or haze. In severe cases, micro-nodules tend to form a rapidly-progressive coalescent white-out pattern. Other identi able causes for granulomatosis should be carefully excluded [113, 114], and the BAL fuid and lung tissue should be examined for bacterial, mycobacterial, fungal, and parasitic infection. The quest for diagnosis is both patientand country-dependent as certain microorganisms (e.g., Mycobacteria, Histoplasma, Cryptococcus, Blastomyces, and Coccidioidomycosis) have a predilection for speci c geographical areas or are more common in patients with a given underlying condition (e.g., Pneumocystis and CTD on immunosuppressive drugs, radiation therapy, long-term corticosteroid therapy). Noninfectious causes of pulmonary granulomas include environmental agents (bird droppings, beryllium, hot-tub), aspiration of food particulates, sarcoidosis, ANCA-related GPA, and rheumatoid arthritis. Central necrosis of the granulomas makes infection a more likely but not an absolute diagnostic consideration [114, 115]. This is seldom is seen in methotrexate lung [116].

A fraction of patients on long-term methotrexate (incidence estimates have dropped from 3% to 0.5% nowadays, down to a point where undisputable cases have become rare) will develop an acute pulmonary reaction without

any convincing features for an infection [117]. Lymphocytes are increased in the BAL in some but not all patients. A con rmatory lung biopsy is now rarely performed. Ruling out pneumocystis pneumonia by immunofuorescence and molecular techniques is vital as pneumocystis pneumonia may assume a non-necrotizing granulomatous pulmonary reaction. A review of the pathologic features of methotrexate pneumonitis revealed granuloma formation in 35%, giant cells in 26.5%, tissue eosinophils in 18%, or diffuse alveolar damage in 8% [116]. An ARDS picture in methotrexate lung can be caused by dense underlying interstitial lung disease with or without granuloma formation and superimposed DAD, pulmonary edema or AH [3, 4, 116].

Management of acute drug-induced granulomatous lung disease is similar to other forms of DIRD and consists of drug withdrawal, exclusion of an infection (particularly in patients exposed to corticosteroids, TNF-alpha antagonists, biologicals or BCG therapy), corticosteroid therapy and supportive care including mechanical ventilation. Mortality can be as high as 16% overall. Relapse occurs in 25–50% of patients, and death may follow in as many as half of patients who relapse after rechallenge with methotrexate [118].

Treatments of bladder carcinoma with topical BCG can lead to an acute pulmonary granulomatous reaction in about 3% of those so treated. In most patients, the disease is self-­ limited but fever and acute respiratory failure can develop in some patients. When tissue hypersensitivity is present, corticosteroid therapy is indicated. In a few patients, pulmonary granulomas correspond to infection with M. bovis BCG and molecular imprints can be found in lung or in other tissues. In such patients, additional chemotherapy against M. bovis is indicated, in addition to corticosteroid therapy [119].

Treatments with TNF antagonists can produce a systemic granulomatous reaction mimicking sarcoidosis, with an indolent course in most patients [3, 4]. Corticosteroid therapy may be indicated. Review of prior interferon gamma release assay is indicated to exclude antecedent infecting exposure to M tuberculosis. A granulomatous pulmonary reaction during treatments with biological therapy should prompt the diagnostic consideration of tuberculosis, miliary tuberculosis or any granulomatous pulmonary infection; all conditions to which TNF antagonists (particularly infiximab), or other biologicals including ICI predispose (Full list in Pneumotox [3, 4]). Tuberculosis can develop despite a negative pretherapy IGRA or preemptive chemoprophylaxis for LTBI, suggesting that treatment with these agents expose not only to reinfection but also to new infection [120].

A few drugs and substances of abuse can cause a foreign body pulmonary reaction that can be detected on pathology [121].

Acute Organizing Pneumonia (OP), Bronchiolitis Obliterans Organizing Pneumonia (BOOP), or Acute Fibrinous Organizing Pneumonia (AFOP) Patterns

Many agents and exposures, including 116 drugs, radiation therapy, excipients, and abused substances can cause organizing pneumonia [3, 4]. Collectively, drugs account for up to a third of biopsy-proven BOOP cases [39]. Approximately 9% of drug-induced/iatrogenic OP cases are lethal from uncontrollable respiratory failure. This is similar to what happens in cryptogenic OP (COP). The quality of evidence for drug causality in OP in the literature is inconsistent and often limited. Authors may attribute the term organizing pneumonia to a certain imaging pattern, with insuf cient evidence. Interobserver reproducibility of imaging patterns is suboptimal [122]. The merit of using pathology descriptors to categorize imaging patterns [29, 30, 123] is unclear [86], inasmuch as radiographic patterns may change with time in a given patient, or may overlap [86], and roentgenographic appearances correlate poorly with pathology [31]. The former Fleischner Society terms [33] stay valid, and the imaging section in Pneumotox [3, 4] offers a classi cation derived from the latter [33].

Ascertaining drug causality is problematic because OP can be idiopathic, occur as a manifestation of a recent infection, upstream bronchial obstruction, or autoimmune disorders including connective tissue disease, infammatory bowel disease, or infammatory myopathies. In hematologic malignancies, OP can occur following stem-cell transplantation. The diagnosis of therapy-related OP should be entertained in the patient with migratory pulmonary opacities or diffuse in ltrates on sequential imaging, in some, a con rmatory histopathologic diagnosis, with exposure to a compatible drug and abatement of signs and symptoms following drug discontinuance be it with or without corticosteroid therapy, and absence of relapse over a prolonged follow-up period.

Acute brinous organizing pneumonia or AFOP is a pathologic pattern that has been reported in association with 19 distinct drugs [3, 4]. This severe OP variant is characterized by a dominant pattern of intra-alveolar brin and brin balls on a background of organizing pneumonia [124]. Organizing pneumonia and AFOP cases may represent the resolving phase of drug-induced acute lung injury/ DAD. Compared to OP, AFOP carries a worse prognosis. Firm evidence that drugs cause AFOP is often elusive. One out of 17 patients described by Beasley was being treated with amiodarone. Two AFOP cases occurred in patients treated with a statin drug. Other causal drugs include bleomycin, mTOR inhibitors, ICI, and trimethoprim-­ sulfamethoxazole [3, 4]. Readministration of any suspect medication after recovery from an AFOP episode should be carefully discussed, lest fatal relapse may supervene.