Nitrofurantoin-induced pulmonary injury

INTRODUCTION — Nitrofurantoin is an antibacterial agent frequently used in the management of urinary tract infections (UTIs). This antimicrobial drug is generally used for treatment of acute cystitis and for prophylaxis in patients with recurrent UTIs. Pulmonary toxicity due to nitrofurantoin has two main presentations: an acute onset approximately nine days after a short course of therapy and a chronic onset developing after several months or years of nitrofurantoin therapy.

The clinical manifestations, diagnosis, and treatment of nitrofurantoin toxicity will be reviewed here. The evaluation and diagnosis of interstitial lung disease in general is discussed separately. (See "Approach to the adult with interstitial lung disease: Clinical evaluation" and "Approach to the adult with interstitial lung disease: Diagnostic testing".)

EPIDEMIOLOGY — The vast majority of patients (85 to 95 percent) who present with pulmonary reactions to nitrofurantoin are women [1,2]. This is related to the greater susceptibility of women to recurrent urinary tract infections, so that women are exposed to long-term nitrofurantoin more often than men.

The median age among those presenting with the acute or chronic forms of the illness is approximately 60 and 70 years, respectively [2,3]. A retrospective cohort study in adults aged 65 and older found an increased risk of lung injury with chronic, compared with acute, use of nitrofurantoin (adjusted relative risk 1.53, 95% CI 1.04-2.24) [4].

The acute form of the illness is the most common pulmonary reaction to nitrofurantoin, occurring in approximately 1 in 5000 patients after first exposure [5,6]. In one study, among 66 patients with pulmonary reactions, 80 percent had an acute presentation and 20 percent had a subacute or chronic presentation [3]. Respiratory impairment severe enough to require hospitalization occurs in approximately 1 of 750 long-term users [7].

The prevalence of nitrofurantoin-induced pulmonary injury decreased from a high in the 1960s to a low in the 1980s. However, this condition is on the increase again, reflecting increased popularity as a urinary antiseptic [8-13].

PATHOLOGY — Acute and chronic pulmonary reactions to nitrofurantoin have different histopathologic patterns [3,14].

●In acute pulmonary reactions, histopathologic examination of the lung typically reveals mild interstitial inflammation (sometimes with eosinophilia), reactive type II pneumocytes, focal hemorrhage, and (in some cases) eosinophils [3]. Less common features include small organizing microthrombi, an alveolar exudate with macrophages and, rarely, granulomas.

●In contrast, diffuse interstitial pneumonitis is observed in chronic reactions [14-16], and some reports have noted vascular sclerosis, thickening of the alveolar septa, and interstitial inflammation [3]. A pattern of organizing pneumonia (also known as bronchiolitis obliterans organizing pneumonia) has also been reported [17-21], as have chronic eosinophilic pneumonia [22], and desquamative interstitial pneumonia with accumulation of macrophages in the alveoli [10,23]. (See "Cryptogenic organizing pneumonia" and "Chronic eosinophilic pneumonia" and "Idiopathic interstitial pneumonias: Classification and pathology".)

PATHOGENESIS — The acute and chronic forms of lung injury result from different pathogenic mechanisms. The acute form has features of a hypersensitivity reaction (type I or III). By comparison, the chronic form may be either a cell-mediated or toxic response. In vitro studies suggest that toxic products produced by drug metabolites may cause injury in the presence of oxygen and lung microsomes. This mechanism of lung damage may be similar to that observed with the herbicide paraquat [24,25]. (See "Paraquat poisoning".)

CLINICAL MANIFESTATIONS — Pulmonary reactions to nitrofurantoin do not appear to be dose related, and can present with a range of symptoms. Concomitant hepatic toxicity has been described [3,26-28].

Acute hypersensitivity — The mean onset of symptoms of acute hypersensitivity pneumonitis due to nitrofurantoin is 8.7 days after initiation of nitrofurantoin, based on a series of 66 patients [3]. However, when symptoms develop after repeat exposure, the onset is much shorter (often within 24 hours) [5,15].

The most frequently reported symptoms are fever exceeding 37.9ºC (82 percent), dyspnea (60 percent), irritating cough (43 percent), and rash (20 percent) [2,3]. Chest pain and cyanosis may also occur [29].

Inspiratory crackles (especially over the basal lung fields) are heard in most patients [3,5].

Chronic pneumonitis — Symptoms of chronic pulmonary reactions develop after several months (eg, six months to several years) of low dose treatment [1,2,16]. The most frequently reported symptoms are dyspnea (73 percent), dry cough (63 percent), and fatigue (37 percent). Symptoms are usually less intense than in acute reactions [3]. Fever is uncommon in patients with the chronic form of the disease [15].

Dry crackles are reported in patients with chronic reactions [3].

For patients with drug-induced lupus or antineutrophil cytoplasmic antibody (ANCA)-associated vasculitis, skin rash (eg, palpable purpura) may be present [30,31].

EVALUATION — The evaluation of suspected nitrofurantoin lung toxicity is designed to review the association of symptoms and signs with nitrofurantoin use, explore other exposures, exclude other diseases with a similar presentation, and assess the severity of respiratory impairment.

Laboratory findings — As with any patient being evaluated for new onset interstitial lung disease, initial laboratory testing is focused on assessing possible contributions from infection, heart failure, and rheumatic disease. The evaluation of acute and chronic presentations of nitrofurantoin lung toxicity is similar, except for a greater focus on infection in acute presentations and a greater focus on rheumatic disease in patients with a gradual onset of symptoms.

A complete blood count and differential is useful to look for eosinophilia and indirectly for evidence of infection. For acute presentations routine bacterial, atypical bacterial, and viral infections are high in the differential diagnosis. In addition to blood cultures, other testing may include serology for influenza, community respiratory viral infections, mycoplasma, and Legionnaire’s disease. A brain natriuretic peptide can help to exclude heart failure. (See "Clinical evaluation and diagnostic testing for community-acquired pneumonia in adults".)

Among patients with acute reactions, the most commonly reported laboratory findings are eosinophilia (69 to 83 percent), leukocytosis (34 to 52 percent), and an elevated erythrocyte sedimentation rate (over 50 mm/hour, 44 percent) [2,3].

By comparison, findings noted in the chronic group include elevated serum gamma globulins (80 percent), eosinophilia (44 percent), elevated serum transaminases (37 percent), and increased erythrocyte sedimentation rates [2,3]. A variety of positive serologies have been reported, including antinuclear antibody [32], antismooth muscle antibody [32], and antineutrophil cytoplasmic antibody (ANCA; typically myeloperoxidase) [3,10,31]. Positive antinuclear antibody was noted in 25 percent of acute cases and 60 percent of chronic cases in one study [2,33]. In a separate report, positive rheumatoid factor tests were found in isolated patients in the acute and chronic groups, but were only assessed in a few patients [2].

Pulmonary function tests — The main role of pulmonary function testing (eg, spirometry, lung volumes, diffusing capacity for carbon monoxide, and six-minute walk test) is to assess the severity of respiratory impairment. Most patients exhibit some degree of pulmonary dysfunction on pulmonary function testing.

●In one study, for example, all patients had a moderate impairment of the gas transfer factor (47 to 78 percent of the reference value) when examined within two weeks after admission to the hospital [3]. No reduction in forced expiratory volume in one second (FEV₁) or vital capacity was observed [3].

●In a second report, half of patients with chronic disease had a restrictive pattern of pulmonary dysfunction with reduced lung volumes [1].

Imaging — A conventional chest radiograph is frequently obtained to evaluate patients with complaints of dyspnea and cough, particularly when associated with crackles on physical examination and/or a decrease in pulse oxygen saturation. High resolution chest tomography (HRCT) is obtained to clarify the pattern and extent of abnormalities.

Abnormalities are observed on chest radiograph in the vast majority of patients (image 1). In one study, one-third of patients were noted to have diffuse parenchymal changes consisting of streaky or mottled shadowing located predominantly in the lower lung zones [3]. Parenchymal changes were almost always bilateral. Pleural effusions, which were predominantly unilateral, were accompanied by parenchymal changes.

In a separate study, all patients with chronic reactions had pulmonary parenchymal opacities on chest radiography [2]. Pleural effusions are rare in patients with the chronic form of the disease.

The CT findings of chronic nitrofurantoin-induced lung disease were evaluated in a retrospective review of 18 patients [1]. Bilateral areas of ground glass attenuation with no dominant pattern of zonal distribution were the most common pattern. In contrast to patients with idiopathic pulmonary fibrosis, irregular linear opacities were uncommon. Honeycombing is rarely present [9,34], and pleural effusions are uncommon [1].

Other patterns that are described in rare reports include chronic eosinophilic pneumonia and organizing pneumonia.

●In a case report, HRCT showed ground glass opacities with an upper lung zone peripheral pattern similar to chronic eosinophilic pneumonia (CEP), and lung biopsy confirmed CEP [22].

●A small number of reports describe a late onset of imaging and histopathology findings consistent with organizing pneumonia [17-20]. In these patients, the HRCT showed bilateral asymmetric areas of consolidation and patchy ground glass opacities [17,19]. (See 'Pathology' above.)

Bronchoalveolar lavage — The main roles of bronchoalveolar lavage are to evaluate for eosinophilia and to exclude infection, alveolar hemorrhage, and lymphangitic carcinomatosis. (See "Role of bronchoalveolar lavage in diagnosis of interstitial lung disease".)

The results obtained with bronchoalveolar lavage (BAL) in patients with nitrofurantoin-induced pulmonary injury are nonspecific and include lymphocytosis, neutrophilia, and eosinophilia [24,35-37].

Lung biopsy — Lung biopsy is typically not necessary due to the high rate of improvement with drug withdrawal. In selected cases, especially when there is no substantial improvement after discontinuation of nitrofurantoin and other diagnoses (eg, infection, heart failure) have been taken into consideration, a lung biopsy may be appropriate. Depending on availability, lung biopsy may be accomplished by transbronchial cryobiopsy or video-assisted thoracoscopic surgery [38]. (See 'Differential diagnosis' below and "Role of lung biopsy in the diagnosis of interstitial lung disease".)

When obtained, lung biopsy of nitrofurantoin pulmonary toxicity typically reveals interstitial inflammation (often eosinophilic) in acute disease, or nonspecific interstitial pneumonia in chronic disease [1], although other patterns such as organizing pneumonia and chronic eosinophilic pneumonia have been reported. (See 'Pathology' above.)

DIAGNOSIS

Acute pneumonitis — The diagnosis of acute lung toxicity from nitrofurantoin is based on the presence of characteristic symptoms, duration of time between drug exposure and symptom onset, and response to drug discontinuation [2,3]. Common alternative diagnoses to be excluded are heart failure, community acquired pneumonia (eg, bacterial, atypical, or viral), and exacerbations of asthma or chronic obstructive pulmonary disease (COPD). (See "Approach to the adult with interstitial lung disease: Clinical evaluation" and "Approach to the adult with interstitial lung disease: Diagnostic testing".)

In the acute form, symptoms should improve within 24 to 48 hours after discontinuation of nitrofurantoin therapy. Failure of resolution should lead to consideration of other diagnostic possibilities and additional testing. (See 'Differential diagnosis' below.)

Chronic lung toxicity — Chronic lung toxicity can usually be recognized by a compatible history of exposure to nitrofurantoin, a restrictive pattern in lung function testing, and bilateral reticular and ground glass opacities on imaging studies. Infection, heart failure, rheumatic disease, and chronic hypersensitivity pneumonitis should be excluded. For patients with more severe respiratory impairment, bronchoscopy with bronchoalveolar lavage may be indicated to evaluate other possibilities, while awaiting response to drug discontinuation. It may take weeks to months for the symptoms and radiographic changes to regress after cessation of nitrofurantoin.

DIFFERENTIAL DIAGNOSIS — The differential diagnosis of acute nitrofurantoin pulmonary toxicity includes infection, lymphangitic tumor, heart failure, acute eosinophilic pneumonia, acute hypersensitivity pneumonitis, and acute interstitial pneumonia. (See "Idiopathic acute eosinophilic pneumonia" and "Hypersensitivity pneumonitis (extrinsic allergic alveolitis): Clinical manifestations and diagnosis" and "Acute interstitial pneumonia (Hamman-Rich syndrome)".)

The differential diagnosis of chronic nitrofurantoin pulmonary toxicity includes the idiopathic interstitial pneumonias, chronic hypersensitivity pneumonitis, sarcoidosis, and rheumatic diseases with lung involvement. However, the presence of autoantibodies can be a manifestation of the reaction to nitrofurantoin and may not indicate a separate rheumatic disease. (See 'Laboratory findings' above and "Drug-induced lupus".)

Depending on the severity of disease, degree of concern about infection, alveolar hemorrhage, or tumor, presence of extrapulmonary manifestations of rheumatic disease, bronchoscopy with bronchoalveolar lavage, and/or a transbronchial or surgical lung biopsy may be needed to evaluate alternative diagnoses [22]. (See "Role of bronchoalveolar lavage in diagnosis of interstitial lung disease" and "Role of lung biopsy in the diagnosis of interstitial lung disease" and "Idiopathic interstitial pneumonias: Classification and pathology".)

TREATMENT — Discontinuation of nitrofurantoin therapy is the cornerstone of therapy for both acute and chronic presentations of nitrofurantoin lung toxicity. For patients with an acute presentation, regression of symptoms and radiographic abnormalities is generally rapid, over days to weeks [2,3,14]. However, for chronic presentations, weeks to months may be required for full recovery. Rechallenge with nitrofurantoin is not advised due to the high likelihood of recurrence [29,33,39].

Oral glucocorticoids are sometimes given as a therapeutic measure [1,3], but the benefit has never been proven. We typically do not use glucocorticoids for treatment of nitrofurantoin lung toxicity, unless the patient has severe respiratory compromise, and only after infection has been carefully excluded. For patients with severe respiratory impairment (eg, dyspnea and hypoxemia at rest), a prednisone-equivalent dose of 0.5 to 1.0 mg/kg (up to 60 mg) per day is given with tapering as tolerated over several weeks to months.

Individual case reports have demonstrated rapid improvement after initiation of glucocorticoids [9,14,15,24,40], but resolution without glucocorticoids is common and even severe pulmonary toxicity may resolve without glucocorticoids [21,40].

Follow up is needed to ascertain whether other abnormalities, such as peripheral blood eosinophilia, elevated transaminases, and autoantibody positivity, have resolved.

PROGNOSIS — The prognosis is good if the condition is recognized early and nitrofurantoin is stopped [14]. Most patients with an acute pulmonary reaction recover within 15 days [3]. One study found that 47 percent of patients were symptom-free within one day, and 88 percent within three days [2].

Among patients with chronic reactions, recovery took from two weeks to three months, although mild fibrotic changes persisted in two-thirds of patients [3]. Resolution of advanced-appearing interstitial disease on CT has been reported following discontinuation of the drug [9,40].

Despite the generally good prognosis, fatalities have been reported [2,3,34,41-43]. As an example, among 392 nitrofurantoin-associated respiratory adverse effects reported to the Medicines and Healthcare products Regulatory Authority (MHRA) from 1963 to 2010, 11 had a fatal outcome [43]. Additional cases of fatal nitrofurantoin-induced pulmonary injury have been described separately [2,3,41,44]. Interactions with concomitant drugs, eg, methotrexate, may lead to irreversible fibrosis [45].

SUMMARY AND RECOMMENDATIONS

●Clinical manifestations – Nitrofurantoin-induced pulmonary toxicity has two main presentations (see 'Clinical manifestations' above):

•Acute reactions – Fever, dyspnea, and cough approximately nine days after starting therapy; rash is occasionally seen

•Chronic toxicity – Insidious onset of dyspnea, dry cough, and fatigue in the absence of fever

●Evaluation – Laboratory findings are generally nonspecific. Peripheral blood eosinophilia is common among patients with acute reactions. Chronic reactions can also be associated with elevations in serum transaminases and erythrocyte sedimentation rate. Imaging frequently shows diffuse opacities. Chronic disease is often associated with restriction on pulmonary function testing. (See 'Evaluation' above.)

●Diagnosis

•Acute reactions – The diagnosis of an acute reaction to nitrofurantoin can be established by the combination of exposure to nitrofurantoin, the characteristic time course and pattern of symptoms, and compatible laboratory tests. Prompt resolution following discontinuation of nitrofurantoin further supports the diagnosis. (See 'Diagnosis' above.)

•Chronic toxicity – The diagnosis of chronic pulmonary toxicity is based on the history of long-term nitrofurantoin therapy, the absence of another explanation for the patient’s symptoms and radiographic abnormalities, and response to drug cessation. (See 'Diagnosis' above.)

●Differential diagnosis

•Acute reactions – The differential diagnosis of acute nitrofurantoin pulmonary toxicity includes infection, heart failure, acute eosinophilic pneumonia, acute hypersensitivity pneumonitis, acute idiopathic interstitial pneumonia, and less commonly lymphangitic tumor. (See 'Differential diagnosis' above.)

•Chronic toxicity – The differential diagnosis of chronic nitrofurantoin pulmonary toxicity includes the idiopathic interstitial pneumonias, rheumatic diseases with lung involvement, chronic hypersensitivity pneumonitis, sarcoidosis, and alveolar hemorrhage. (See 'Differential diagnosis' above and "Approach to the adult with interstitial lung disease: Diagnostic testing" and "Approach to the adult with interstitial lung disease: Clinical evaluation".)

●Treatment – Discontinuation of nitrofurantoin is required for both acute and chronic presentations of nitrofurantoin lung toxicity because continued use is associated with progressive respiratory insufficiency. We do not typically use other treatments (eg, glucocorticoids or immunosuppressants) in the absence of severe respiratory compromise. Nitrofurantoin should not be reintroduced due to the high rate of recurrence. (See 'Treatment' above.)

For patients with substantial respiratory insufficiency (eg, dyspnea and hypoxemia at rest), we suggest initiation of oral glucocorticoids (Grade 2C) after ruling out infectious etiologies. The usual dose is the equivalent of prednisone 0.5 to 1.0 mg/kg (up to 60 mg) per day with tapering as tolerated over several weeks to months. (See 'Treatment' above.)

●Prognosis – Generally, symptoms and signs of acute nitrofurantoin lung toxicity improve within a few days, while chronic toxicity improves over weeks to months. Some patients have persistent radiographic abnormalities; rare cases are fatal. (See 'Treatment' above and 'Prognosis' above.)

ACKNOWLEDGMENT — The UpToDate editorial staff acknowledges Jochen Schumacher, MD, who contributed to an earlier version of this topic review.