This study evaluated the effectiveness of aerosolized cyclosporine as rescue therapy for refractory acute rejection in lung-transplant patients that is unresponsive to conventional therapy. Over 2 yr, nine allograft recipients with histologic evidence of persistent acute rejection and worsening pulmonary function were enrolled. Twenty-two patients with similar degrees of unremitting rejection served as historical controls. Aerosolization of cyclosporin A (300 mg in 4.8 ml propylene glycol) using an AeroTech II jet nebulizer was instituted daily for 12 consecutive days followed by a maintenance regimen of 3 d/wk. Cyclosporine and tacrolimus blood and plasma levels were maintained within therapeutic ranges throughout this trial. Efficacy was assessed by histologic grade of rejection, interleukin-6 (IL-6) mRNA expression by graft bronchoalveolar lavage cells, and pulmonary function testing before and during cyclosporine therapy. In seven patients, results were correlated to deposition of cyclosporine aerosol in the allograft(s) as measured by radioisotopic techniques. At a mean of 37 d after initiation of aerosolized cyclosporine, graft histology improved in eight of the nine patients. Cellular IL-6 mRNA expression decreased significantly in seven patients (mean IL-6/actin +/- SD, 40.96 +/- 118 versus 0.33 +/- 0.57 [p = 0.038]). Pulmonary function (FEV1), which had decreased posttransplant (over a mean of 347 d of observation) from a best value of 1.98 +/- 0.8 L to 1.59 +/- 0.6 L (p = 0.0077), improved over time (152 d) to a posttransplant value of 1.90 +/- 0.8 (p = 0.025). In the control subjects, FEV1 inexorably declined over a comparable period of observation (best posttransplant value 2.36 +/- 0.86 to 1.32 +/- 0.53, p < 0.0001). There was a strong correlation between cyclosporine deposition in the allograft and improvement in FEV1 (r = 0.900, p < 0.01). Fewer cycles of pulsed corticosteroids (1.4 +/- 0.9 versus 0.2 +/- 0.4, p = 0.011) and anti-thymocyte globulin 0.8 +/- 0.4 versus 0, p = 0.018) and reduced doses of oral prednisone (10.8 +/- 3.1 versus 6.1 +/- 4.2 mg/d, p = 0.026) were observed during treatment with aerosolized cyclosporine. Episodes of pneumonia also were reduced significantly during aerosol therapy (2.6 versus 0.95 episodes/100 d, p = 0.029). Nephrotoxicity and hepatotoxicity did not occur, and no patients withdrew from the study. Aerosolized cyclosporine appears to be safe and effective therapy for refractory acute rejection, but confirmation by a larger, randomized trial is necessary. The correlation observed between deposition of cyclosporine aerosol and physiologic improvement of lung function suggests that there is a dose-response relationship between the concentration of cyclosporine in the allograft and immunologic tolerance.