Spirometry and pulmonary mechanics were measured pre- and postoperatively in 37 patients undergoing bilateral lung volume reduction surgery (LVRS). The relative contributions of changes in compliance (CL), recoil pressures (PTLC), small airway conductance (Gu), and airway closing pressures (Ptm') to changes in expiratory flows were examined with a Taylor series expansion of the Pride- Permutt model of flow limitation. The resulting variational expression, deltaVmax = GudeltaPel + PeldeltaGu - GudeltaPtm' - Ptm'deltaGu - deltaGudeltaPtm', was then used to describe how the peak flow rate (Vmax) depends on preoperative Gu, P TLC, Ptm', and on changes (delta) in these parameters after surgery. After LVRS, both FEV(1) and Vmax increased significantly ( DeltaFEV(1) = 28 +/- 44%; DeltaVmax = 78 +/- 132%), and changes in FEV(1) and Vmax correlated closely (r = 0.74, p < 0.001). Among responders (DeltaFEV(1) > or = 12%; n = 19; DeltaFEV(1) = 60 +/- 38%), PTLC increased (8.8 +/- 2.8 to 12.2 +/- 4.7 cm H2O) and the time constant for expiration (tau = CL/Gu) decreased (2.67 +/- 0.62 to 2.35 +/- 0.55 s), while Ptm', CL, and Gu did not change. GudeltaPel, the change in recoil weighted by preoperative conductance upstream of the flow-limiting site, accounted for 72% of the improvement in Vmax. Among nonresponders ( DeltaFEV(1) = -6 +/- 15%, n = 18), tau increased significantly, contributing to a decline in FEV(1)/FVC ratio. PeldeltaGu decreased (-0.25 +/- 0.68, p = 0.013), accounting for all of the decline in Vmax. This analysis suggests that (1) improvement in expiratory flows after LVRS is largely due to increases in recoil pressure; (2) large improvements in FEV(1) can occur without changes in Gu or Ptm', arguing that LVRS has little effect on airway resistance or closure; and (3) large changes in PTLC can occur without changes in CL, supporting arguments of Fessler and Permutt (Am J Respir Crit Care Med 1998;157:715-722) that "resizing of the lung to chest wall" is the primary mechanism by which LVRS improves lung function.