The role of short-term tests of reversibility in selecting patients with COAD for long-term nebuliser therapy is uncertain. In a double-blind placebo-controlled crossover study we have examined the correlation between short-term reversibility and response to a home nebuliser. We studied 20 patients with severe COAD (mean age 66, mean FEV1 0.81 l) and little reversibility (less than 20% increase in FEV1 post-inhaled salbutamol 200 micrograms and less than 25% increase in peak expiratory flow rate, PEFR, on oral steroids). PEFR, spirometry, lung volumes and airways conductance were recorded before and 1 h after a mixture of nebulised ipratropium 0.5 mg and fenoterol 1.25 mg. Patients then recorded twice-daily PEFR at home while they received nebulised ipratropium plus fenoterol, or saline placebo, four times a day for three week blocks using a double-blind cross over protocol. Mean PEFR on home nebuliser rose from 164 l m-1 (placebo) to 196 l m-1 (ipratropium plus fenoterol), paired t-test P = 0.0001. Correlation coefficients between short-term response for PEFR, spirometry and lung volumes, and improvement in home PEFR on nebulised ipratropium plus fenoterol, were all poor (R = -0.37-0.35, P = 0.83-0.11). We conclude that in severe COAD, reversibility tests of PEFR, spirometry and lung volumes do not correlate with response to a home nebuliser. Home measurements of PEFR are probably the best objective method of assessing response to a home nebuliser in such patients.