H ypersensitivity pneumonitis (HP) is an interstitial lung disease (ILD) characterized pathologically by varying degrees of inflammation and/or fi brosis that result from the inhalation of an antigen to which a person has been previously sensitized. The clinical expression of HP can vary and includes acute, sub acute, and chronic forms. Symptom onset that coincides with antigen exposure provides the fi rst diagnostic clue; however, despite an exhaustive search, a recognizable, temporal relationship (between exposure and symptoms) is frequently absent, particularly in patients who present with the chronic form of HP.Chronic HP, characterized clinically by ventilatory restriction and, most commonly, upper-zonepredominant fi brosis on high-resolution CT (HRCT) scan, is irreversible and often progressive. It is believed that continued exposure to an inciting antigen (IA) perpetuates progression and the development of fi brosis in HP, but this belief hinges more on scientific rationale than systematic research. In this study, we hypothesized that among patients with histologic and Background: The cornerstone of hypersensitivity pneumonitis (HP) management is having patients avoid the inciting antigen (IA). Often, despite an exhaustive search, an IA cannot be found. The objective of this study was to examine whether identifying the IA impacts survival in patients with chronic HP. Methods: We used the Kaplan-Meier method to display, and the log-rank test to compare, survival curves of patients with well-characterized chronic HP stratifi ed on identifi cation of an IA exposure. A Cox proportional hazards (PH) model was used to identify independent predictors in time-todeath analysis. Results: Of 142 patients, 67 (47%) had an identifi ed IA, and 75 (53%) had an unidentifi ed IA. Compared with survivors, patients who died (n 5 80, 56%) were older, more likely to have smoked, had lower total lung capacity % predicted and FVC % predicted, had higher severity of dyspnea, were more likely to have pulmonary fi brosis, and were less likely to have an identifi able IA. In a Cox PH model, the inability to identify an IA (hazard ratio [HR], 1.76; 95% CI, 1.01-3.07), older age (HR, 1.04; 95% CI, 1.01-1.07), the presences of pulmonary fi brosis (HR, 2.43; 95% CI, 1.36-4.35), a lower FVC% (HR, 1.36; 95% CI, 1.10-1.68), and a history of smoking (HR, 2.01; 95% C1, 1.15-3.50) were independent predictors of shorter survival. After adjusting for mean age, presence of fi brosis, mean FVC%, mean diffusing capacity of the lung for carbon monoxide (%), and history of smoking, survival was longer for patients with an identifi ed IA exposure than those with an unidentifi ed IA exposure (median, 8.75 years vs 4.88 years; P 5 .047). Conclusions: Among patients with chronic HP, when adjusting for a number of potentially infl uential predictors, including the presence of fi brosis, the inability to identify an IA was independently associated with shortened survival.
CHEST 2013; 144(5):1644-1651Abbreviations: D lco 5 diffusing capacity of the lung f...