A Practical Clinical Model for Diagnosing Occupational Asthma in Workers Exposed to Low-Molecular-Weight Agents: Model Development and Validation.

Abstract

The specific inhalation challenge (SIC) is the reference standard for diagnosing occupational asthma (OA) but is not widely used globally. We aimed to develop a more practical clinical model to diagnose OA in workers exposed to low-molecular-weight (LMW) agents. We conducted a diagnostic study using clinical interview variables and non-SIC tests as predictors. OA was defined by positive SIC. Retrospective data from Quebec and British Columbia were used to develop logistic models. External validation included centers with routine SIC (Finland, Poland) and expert-confirmed OA (Ontario, Turkey, England). The clinical interview model included male sex, isocyanate exposure, work-related rhinoconjunctivitis, smoking status, and exposure duration <10 years. The clinical interview model can correctly discriminate a positive from a negative SIC in 65% of the cases (area under the receiver operating characteristic curve [AUC] = 0.65). Adding diagnostic tests to the clinical interview model improved the AUC to 0.73 for the nonspecific bronchial hyperreactivity (NSBHR), 0.80 for the serial peak expiratory flow (PEF), and 0.81 for NSBHR plus serial PEF. Combining the clinical interview with serial PEF was the model of choice, showing strong internal validity (shrinkage 0.93) and adequate calibration (Hosmer-Lemeshow P > .05). In Finland/Poland, AUCs were 0.61 for the clinical interview alone and 0.72 with serial PEF; in England, 0.73 and 0.81; and in Ontario/Turkey, 0.60 and 0.68, respectively. Calibration was adequate in all centers. A novel model, comprising clinical features and serial PEFs, can predict positive SIC caused by LMW agents. It can guide referrals or diagnosis when SIC is unavailable or unnecessary.