Reported cases of bacterial pneumonia in hospitalized patients have increased at a rate disproportionate to other common inpatient diagnoses. ^{Reference Prasad1} Causes of this trend are multifactorial—an increasingly frail patient population with impaired pulmonary hygiene, heightened sensitivity of contemporary diagnostic modalities, and lack of standardization in the diagnosis of pneumonia itself. Mounting evidence, however, suggests that the rigor of the case definition for what constitutes bacterial pneumonia has declined, contributing to overdiagnosis and antibiotic misuse. ^{Reference Burton, Price and Barr2–Reference Fridkin, Baggs and Fagan5}

Nowhere is the issue of pneumonia overdiagnosis more relevant or challenging than among hospitalized patients. Consider the following examples:

• At an academic center, a multidisciplinary review of 231 cases diagnosed and treated by intensive care unit (ICU) teams as ventilator-associated pneumonia (VAP) determined that antibiotics were not indicated in 59% of cases on day 1 of diagnosis. This number rose to 74% by day 3, accounting for 1,183 days of inappropriate antibiotics over 1 year. ^{Reference Nussenblatt, Avdic and Berenholtz6} Sputum culture collection on day 3 of treatment was independently associated with extension of antibiotic therapy.

• A single-center multidisciplinary review of surgical ICU patients determined that <25% of those empirically treated for postoperative pneumonia met standardized criteria for infection, accounting for 1,460 days of inappropriate antibiotics in <1 year. ^{Reference Swoboda, Dixon and Lipsett7}

• Within a matched cohort of mechanically ventilated patients with positive respiratory cultures who had minimal ventilator requirements and received antibiotics, “ultra-short” courses of 3 days of antibiotics yielded outcomes similar to longer courses of treatment, suggesting that some patients presumptively treated for VAP based on respiratory cultures are misdiagnosed. ^{Reference Klompas, Li, Menchaca and Gruber8}

• In a multicenter cohort of 12,273 hospitalized patients treated for possible community- or hospital-acquired pneumonia—many of whom had positive sputum cultures—2,149 (17.5%) had no cardinal signs of pneumonia at the time of diagnosis. ^{Reference Klompas, Ochoa and Ji9}

Our institutional experience similarly reveals a high prevalence of pneumonia misdiagnosis. We retrospectively reviewed the management of a cohort of hospitalized patients with positive respiratory cultures who were presumptively treated for bacterial pneumonia. ^{Reference Albin, Henig and Patel10} Each case and its associated imaging were evaluated by a blinded committee of infectious diseases physicians to determine prevalence of pneumonia misdiagnosis. Pneumonia misdiagnosis was defined as a case (1) not meeting and the Infectious Diseases Society of America/American Thoracic Society (IDSA/ATS) criteria of a “new lung infiltrate plus clinical evidence that the infiltrate is of infectious origin, which includes new onset of fever, purulent sputum, leukocytosis and/or decline in oxygenation” or (2) a case in which the reviewing clinician would have recommended against antibiotics for treatment of pneumonia. ^{Reference Kalil, Metersky and Klompas11} Among this cohort of 705 patients, 242 (34.3%) were determined to have pneumonia misdiagnosis.

Recognizing that a high burden of proof must be met to defer antimicrobial use for hospitalized patients with possible pneumonia, we sought to quantify unnecessary antibiotic use among this subset of patients with pneumonia misdiagnosis in as conservative a manner as possible. We considered antibiotic administration appropriate if patients had pathogenic organisms isolated from cultures outside of the respiratory tract, documented evidence of ventilator-associated tracheobronchitis, evidence of acute exacerbations of COPD or cystic fibrosis, or isolation of pathogenic organism from the respiratory tract among severely immunocompromised individuals. Despite such allowances, antibiotics were determined to have been administered unnecessarily in 151 patients (21.4% of the total cohort); these patients each received a median duration of 12 days of antibiotics. Respiratory cultures were obtained in a variety of clinical scenarios where their benefit is uncertain: following macroaspiration or mucus plugging episodes that rapidly resolved with supportive care, during surveillance bronchoscopies prior to elective tracheostomy or while patients were receiving extracorporeal membrane oxygenation, in patients with heavy secretion burdens during failed spontaneous breathing trials, and as part of a “pan-culture” workup for a new fever or leukocytosis in patients without localizing pulmonary features of infection.

Appraisal of these data reveals a common feature: the presence of positive respiratory cultures incenting pneumonia misdiagnosis and treatment. A positive culture from the respiratory tract informs the diagnostic probability of pneumonia, but by itself is not synonymous with infection. Respiratory bacterial colonization with pathogenic organisms has been reported in nearly 25% of hospitalized patients, 40% of patients with selected comorbidities, and >50% of patients requiring intensive care. ^{Reference Johanson, Pierce and Sanford12–Reference Jongerden, Speelberg and Satizábal15} Among patients requiring mechanical ventilation, respiratory colonization of endotracheal material is common; it is reported in up to 90% of tracheostomy-dependent patients. ^{Reference Feldman, Kassel and Cantrell16,Reference Lepainteur, Ogna and Clair17} Sampling of the distal respiratory tract represents an attractive option to differentiate proximal airway colonization from clinically relevant infection, but a positive culture obtained via bronchoalveolar lavage by itself adds little to the diagnostic accuracy of VAP, with reported specificity as low as 42%. ^{Reference Klompas18} Indeed, advances in culture-independent microbiologic techniques have demonstrated that the lungs, once considered sterile, in fact house diverse communities of bacteria, the relative composition of which may influence health and disease. ^{Reference Dickson, Erb-Downward and Freeman19}

This scenario—high prevalence of asymptomatic colonization of a body site once considered sterile, indiscriminate collection of cultures precipitating antibiotic misuse, and practice patterns equating positive cultures with infection—parallels that of asymptomatic bacteriuria (ASB). We believe that this comparison is instructive and that stewardship efforts for pneumonia should be patterned to a greater extent on validated strategies proven to mitigate overtreatment of ASB. ^{Reference Nicolle, Gupta and Bradley20} Adoption of the term “ASB” into clinical vernacular has facilitated recognition of its prevalence and spotlighted the distinction between bacteriuria and clinically relevant urinary tract infection. Accordingly, we propose similar adoption of the term “asymptomatic bacterisputia” (ASBS) into antibiotic stewardship heuristics.

There are conceptual differences between asymptomatic bacteriuria and bacterisputia—notably, in how we characterize an asymptomatic state for each condition. For bacteriuria, clear symptoms (eg, dysuria) register the condition as symptomatic and indicate the need for antibiotic treatment. The clinical scenario in which bacterisputia occurs, however, is often more muddled. Isolated clinical symptoms (productive cough, dyspnea) and radiographic features (infiltrates) suggesting a diagnosis of bacterial pneumonia overlap substantially with noninfectious mimics of bacterial pneumonia in hospitalized patients (eg, COPD exacerbations, aspiration pneumonitis, decompensated heart failure, acute respiratory distress syndrome, atelectasis, and viral respiratory infections). Thus, the idea that individual signs or symptoms should register bacterisputia as symptomatic and therefore trigger antibiotic treatment is problematic.

Instead, a more instructive way to frame this issue is not to ask which individual respiratory symptoms should trigger the treatment of bacterisputia but rather to identify concrete clinical scenarios in which bacterisputia should not require treatment. The relevant parallel to ASB in this context is altered mental status in functionally or cognitively impaired individuals with bacteriuria. This factor was classically felt to be a constitutional manifestation of UTI, even in the absence of fevers or leukocytosis, but recent work suggests that bacteriuria in this context is often an incidental finding for which antimicrobial treatment should be deferred. ^{Reference McKenzie, Stewart, Bellantoni and Finucane21,Reference Mayne, Bowden, Sundvall and Gunnarsson22} Bacterisputia is often similarly discovered during a work-up for an isolated respiratory complaint or constitutional sign in a patient who otherwise lacks compatible aggregate clinical or radiographic features of pneumonia. In these scenarios, while the patient is indeed “symptomatic,” symptoms are not causally dependent upon bacterisputia.

Use of the term ASBS rather than “respiratory colonication” has the potential (1) to highlight the disconnect between microbiologic findings and antibiotic-responsive clinical syndromes and (2) to situate respiratory stewardship efforts within a framework established for ASB. Using this framework as a foundation, ^{Reference Nicolle, Gupta and Bradley20} we encourage the following initiatives:

1. 1) Comprehensive epidemiologic research is needed to better define the prevalence and microbial ecology of respiratory tract colonization in specific patient subgroups and the associations between ASBS and physiologic and clinical outcomes. Special attention should be given to patients with structural lung disease other than cystic fibrosis, impaired oropharyngeal airway defenses, prolonged/chronic dependence on mechanical ventilation, immunocompromising conditions and critical illness. In these populations, bacterisputia is common and is often of uncertain clinical significance. This research should utilize traditional microbiologic techniques as well as emerging culture-independent modalities so that findings can be interpreted in the context of our growing understanding of the respiratory microbiome. Like ASB work, such studies should (1) inform the pretest probability of culturing organisms from the respiratory tract of patients with diverse clinical characteristics and (2) inform investigations into potential benefits and harms of antimicrobial treatment of ASBS.

2. 2) Clinical studies and quality improvement initiatives addressing the appropriate collection of respiratory cultures are needed, particularly establishing the safety of deferring respiratory cultures in low-yield clinical scenarios, thereby averting triggers for unnecessary antibiotic use. Potential low-yield scenarios for obtaining respiratory cultures include during surveillance bronchoscopies, following macroaspiration events, in the setting of mild, transient respiratory decompensations in patients with mechanical airways, in cases of established viral respiratory infections (where biphasic clinical decompensation is not present), and as part of a “pan-culture” workup for fever (particularly relevant for mechanically-ventilated patients without changes in respiratory status).

3. 3) Multidisciplinary initiatives are needed to improve provider contextualization of positive respiratory cultures through emphasis on data other than qualitative culture results. Interventions by microbiology labs in culture workflows for urinary specimens requiring the presence of pyuria on urinalyses prior to performance of urine cultures, deferring automatic reporting of subquantitative or polymicrobial culture growth results—have been efficacious in limiting treatment of ASB. ^{Reference Wiley, Jacob and Burd23} Similar variables pertaining to respiratory specimens (eg, quantitative respiratory culture thresholds, culture/gram stain concordance, and the degree of alveolar neutrophilia obtained from distal pulmonary sampling) inform the likelihood of the pneumonia diagnosis and can be used to “nudge” appropriate antibiotic prescribing behaviors. ^{Reference Klompas18,Reference Walter, Ren and Yacoub24}

4. 4) Provider education and oversight initiatives performed by antimicrobial stewardship programs aimed at refining institution-specific microbiologic diagnostic practices for respiratory infections are needed. These may include decision support systems, prospective audit and feedback, and clinical protocols with interventions at the preanalytic, analytic, and postanalytic phases of the diagnostic testing pathway. ^{Reference Nicolle, Gupta and Bradley20,Reference Wiley, Jacob and Burd23,Reference Barlam, Cosgrove and Abbo25}

5. 5) Advancements in reliable pneumonia diagnostic tools are needed. A major driver of antibiotic use, even among patients with low pretest probability of pneumonia, is the lack of a reliable biomarker, imaging modality, or clinical indicator to confidently rule out a serious bacterial infection of the lung parenchyma. This factor incentivizes antibiotic use out of caution. Future studies should evaluate incorporation of novel imaging techniques and biomarkers into the pneumonia diagnostic pathway and should prospectively investigate the safety of antibiotic cessation when applying these techniques.

Positioning pneumonia stewardship efforts within a framework established by ASB represents an opportunity to improve patient care and curb unnecessary antibiotic use. Adoption of the term ASBS alone will not solve the problem of pneumonia overdiagnosis, but it may provide clinicians with a familiar heuristic, potentially facilitating efforts to improve management decisions and encouraging research into the responsible procurement and interpretation of respiratory cultures.