An experimental human model of metal fume fever

To examine the pathogenesis of metal fume fever in humans by studying functional, cellular, and biochemical responses after exposure to zinc welding fume.

Clinical experimental study.

We studied 14 welders recruited through public advertisements.

Participants welded galvanized steel.

We measured lung volumes, airflow, diffusing capacity for carbon monoxide, and airway reactivity at baseline as well as either 6 or 20 hours after welding. We carried out bronchoalveolar lavage either 8 hours (early follow-up, 5 participants) or 22 hours (late follow-up, 9 participants) after welding, assaying the fluid for total and differential cell counts and bronchoalveolar lavage supernatant concentrations of interleukin-1 and tumor necrosis factor (TNF).

Changes in pulmonary function and airway reactivity were minimal. Cumulative zinc exposure and polymorphonuclear leukocyte count in bronchoalveolar lavage fluid at late (r = 0.87; P less than 0.01) and early (r = 0.93; P less than 0.05) follow-up were positively correlated. Among the late follow-up group, the mean proportion of polymorphonuclear leukocytes was 37% (range, 19% to 63%), a statistically greater proportion than the 9% (range, 2% to 21%) seen among the early follow-up group (P less than 0.05). We did not detect TNF or more than a trace amount of interleukin-1 in the bronchoalveolar lavage supernatant.

Zinc oxide welding fume was associated with a marked dose-dependent increase in the number of polymorphonuclear leukocytes recovered in bronchoalveolar lavage fluid 22 hours after exposure but was not associated with a clinically significant change in pulmonary function or airway reactivity. Although we did not identify increases in either interleukin-1 or TNF levels in bronchoalveolar lavage fluid, cytokines or a cytokine-like mechanism may mediate the syndrome of metal fume fever.