Background The significance of consistent small changes in PEF related to work, or records with diurnal variation within the normal range, are disputed. We related PEF changes at diagnosis to longitudinal changes in FEV1 in workers remaining exposed to their causative agent.

Methods 163 consecutive workers attending a tertiary referral clinic were identified. All had a clinical history suggestive of occupational asthma and a diagnostic serial PEF record (Oasys score >2.50). Average annual rate of FEV₁ decline was calculated using simple linear regression for each worker with at least 12 months of FEV1 measurements. PEF responses to work exposure were made from the initial diagnostic PEF record. Subjects were split into tertiles according to their PEF responses. Mean work-rest PEF was >35 l/min in the highest tertile and <20 l/min in the lowest. Corresponding values for diurnal variation (DV) % predicted were ≥16% (high) and <10% (low).

Results Ninety workers had FEV1 measurements for at least 12 months during exposure to the original causative agent (median 2.9 years, IQR 1.8–4.8 years). The mean annual loss of FEV1 was increased at 101 (SEM 17.7) ml/year, unrelated to agent; high MW (10), isocyanates (16), metals (13), colophony (9), biocides (10), metal working fluid (7) and others (23). There were no significant differences related to continued smoking or treatment. There were large between individual differences with annual declines <50 ml in 35/90 workers. There were no statistically significant differences between those with high rest-work PEF changes (−126 SEM 44 ml/year), and low changes (95 SEM 22.1 ml/year), or between those with high DV (98 SEM 36.5 ml/year) and low DV (142 SEM 30.3 ml/year).

Conclusion Some workers with occupational asthma have normal FEV1 decline during continued exposure, but the average is around 3× predicted. Workers with small work-related PEF changes or low DV do as badly as those with larger values. The classification of FEV1 decline in an individual worker requires longer measurements than we have available; the group mean changes should be more reliable. Those who are able to remain exposed are likely to be those with least severe disease.