Clinical Decision-Making in Hypersensitivity Pneumonitis

Diagnosis and Management

Evans R. Fernández Pérez, MD, MS; Tilman L. Koelsch, MD; Paolo M. Leone, MD; Steve D. Groshong, MD, PhD; David A. Lynch, MD; Kevin K. Brown, MD


Semin Respir Crit Care Med. 2020;41(2):214-228. 

In This Article

Estimating the Pretest Likelihood of Disease

The Clinical Context

Estimating disease likelihood before and after performing diagnostic testing through Bayesian reasoning is highly advisable when dealing with complex disorders like HP (Figure 1). Since the predictive value of a test is highly influenced by the background prevalence of the disease, having a rough sense of the epidemiology of HP is also key when conducting information-gathering during the diagnostic process. Thus, when establishing a pretest estimate of HP, the clinician must consider the specificity of the presenting findings, the number of independent features suggestive of HP, the individual's risk, and a general sense of the background prevalence of the disease. Since the likelihood of HP is generally uncertain at this step, diagnostic testing is performed to refine the differential. When evaluating an ILD, this almost always means a high-resolution computed tomographic (HRCT) scan of the chest. For instance, in a study performed to develop a diagnostic predictive model of chronic HP (CHP), the combination of patient age, a history of down feather and/or bird exposure, the presence of diffuse craniocaudal ground-glass opacity (GGO), and mosaic perfusion on HRCT had a specificity of 91% and a sensitivity of 48% for the diagnosis of CHP.[3] However, the chest HRCT (and any other subsequent tests) needs to be interpreted within the patient-specific clinical context by including the information, which is acquired during the clinical history and physical examination, and the background prevalence of HP in the population and setting of interest.

Figure 1.

Management varies according to disease behavior and treatment risk–benefit ratio and needs recalibration in patients with a working diagnosis or advanced disease (see the General Approach subsection). Management decisions can influence or interrupt the HP diagnostic process in subjects with a progressive fibrotic phenotype. Thus, the diagnostic HP approach may need recalibration in subjects with a working diagnosis or advanced disease according to disease behavior. BAL, bronchoalveolar lavage; MDD, multidisciplinary discussion; FH, family history; PMH, past medical history; ROS, review of systems; TT, testing threshold, TxT, treatment threshold; HP, hypersensitivity pneumonitis.

Background Prevalence of Hypersensitivity Pneumonitis

Variations in HP incidence and prevalence (Table 1) correlate with age, season, geography, occupation, and host risk factors. For example, in a retrospective study of 6,920 pigeon breeders identified in the records of the Danish Racing Pigeon Association, the adjusted hazard ratios of HP and other ILDs for pigeon breeders were 14.4 (95% confidence interval [CI]: 8.1–25.4) and 1.3 (95% CI: 1.0–1.7), respectively.[4] Also, in a prospective registry of 1,084 patients with new-onset ILD in India, the clinical diagnosis of all ILD cases validated by multidisciplinary discussion disclosed that almost half (47%) of the patients had a final diagnosis of HP attributable to domestic environmental exposure.[5] Conversely, published registries from the United States and Europe report a much lower frequency of HP among all ILD diagnoses. With this in mind, in a given setting, the information yield of a diagnostic test increases when applied to demographic groups known to be at higher risk. For instance, serum antigen-specific antibody testing may have a higher positive predictive value among people living in a farming community than among those living in areas with low HP prevalence. Similarly, the chance of finding HP is considerably smaller in a primary care practice compared with an ILD referral center, making the positive predictive value of any test significantly lower.

Antigen Identification and Exposure Source

HP associated with a typical exposure is, in relative terms, easy to recognize. The relationship of the development over days of dyspnea, cough, a diffusely abnormal chest imaging, and recent intense exposure, with moldy hay, silage, or grain in a farming community is generally quickly identified. A patient who has subacutely developed similar symptoms and who has been a bird-breeder for the past 18 months is also relatively easy to diagnose. Specific, unusual exposures are easier to appreciate (e.g., suberosis, cheese-washer's lung) than potential household molds. However, most of the morbidity and early mortality in HP is a result of unidentified or unappreciated antigen exposure. In some instances, exposure to the IA may have ceased before the recognition of the disease or evidence of disease progression, being an all-but-forgotten link between exposure and the lung disease.[6]

In many cases, the exposure goes unrecognized because of (1) the absence of a thorough environmental history, (2) the lack of an appreciation for a temporal association between a specific exposure and HP development, (3) the inaccurate recollection of past exposures by the patient, (4) lack of knowledge of the potential significance of a specific exposure, or (5) lack of recognition of a new potential antigen(s) not previously described. That the IA remains unidentified does not mean that the disease is idiopathic. For example, in a study of 46 patients diagnosed with idiopathic pulmonary fibrosis (IPF) according to the American Thoracic Society/European Respiratory Society (ATS/ERS) 2011 criteria, after comprehensive reevaluation, 43% were subsequently diagnosed with FHP. Ten (63%) of the 16 cases with an identified exposure and 10 (33%) of the 30 cases without a history of specific concerning exposures were subsequently diagnosed with FHP.[7] Also, in a recent study of 95 patients with undiagnosed fibrotic ILD and an indeterminate-for-usual interstitial pneumonia (UIP) pattern on HRCT, bronchoalveolar lavage (BAL) fluid cellular analysis led to a reevaluation, and change in diagnosis in 14 (15%). In the majority of these cases (11/14; 79%), the initial diagnosis was altered from IPF to HP, and a previously unidentified IA exposure was found in seven patients.[8]

Keeping in mind these caveats, there are several key elements of a comprehensive exposure history (Figure 2). During the history, the clinician should have available a clinically relevant environmental and occupational questionnaire to guide the interview.[9] The questionnaire should be adapted to regional and local geography and customs. Grouping questions into three domains is often useful: the environmental history, the occupational profile, and exposure inventory.[10] The focus of the environmental history is to characterize home environment exposures that may be clinically significant, particularly uncontrolled indoor mold and mildew from water damage and dampness, condensation (e.g., windows), or standing water (e.g., swamp coolers, Jacuzzi tubs).[11] The occupational profile is a summary of the patient's occupational experiences (e.g., employers, occupations, job duties, dates worked) from which exposure to an IA or hazards and the degree of risk can be characterized. Lastly, the goal of the exposure inventory is to reconstruct, in more detail, the circumstances and the extent of potentially hazardous exposures.[10]

Figure 2.

(a) To help understand the magnitude, geographical boundaries, and patient proximity to the suspected exposure source. (b) Self-administered structured environmental and occupational questionnaires can help save clinic visit time and optimize recall. When done at home, the patient exposure can be reconstructed with the help of relatives or caregivers. Schedule periodic evaluations with the objective of updating prior probabilities, as more information is gathered about the exposure history and for early detection of the inciting antigen. (c) To help understand the type of exposure from the material safety data sheets and proximity of the home/workplace to industry or waste site (e.g., A searchable list of exposures and antigens can be found at websites such as and If a sentinel occupational case is suspected, consider early referral to occupational lung specialist. (d) Referral and consultation for exposure verification and for its adequacy in explaining the patient's exposure context may be considered on a case-by-case basis. Although there are no standardized validated home inspection protocols, when indicated, the environmental assessment can help provide remediation and source avoidance recommendations.

The likelihood of any specific IA exposure being responsible for causing HP can be informed in the context of causal inference: (1) strength, (2) consistency, (3) dose–response, (4) temporality, (5) reversibility (6) biological plausibility, (7) specificity, and (8) analogy.[12] However, the association between the possible IA exposure and disease may be viewed as convincing when several of the aforementioned features are present (e.g., appropriate temporal sequence alone is weak evidence for cause) and if all attempts to explain it away due to confounding have failed: were other causes overlooked and are there counterexamples of strong noncausal associations (e.g., persistent or progressive lung disease when away from the possible IA exposure or the presence of compelling comorbid conditions explaining the "HP" like presentation)?

While the screening and estimation of the magnitude, frequency, and duration of bird antigen exposure is commonly straightforward, the identification of hidden mold or bacteria associated with HP can be challenging and may require consultation with an occupational health professional, with a walk-through investigation of the home or worksite. Despite lack of standardization, environmental inspection and sometimes sampling of the implicated environment by an experienced industrial hygienist may help identify the source if there is suspicion for an unverified home mold exposure (i.e., musty smell but no visible mold growth) or workplace hazard.[13] This may be considered when HP is progressive without established risk factors for progression (such as pulmonary fibrosis) and when trying to distinguish the primary driving IA from other identifiable exposures (e.g., home[s] vs. workplace).