BACKGROUND--In patients with limb girdle dystrophy the relative contribution of peripheral factors (respiratory muscle weakness, and lung and/or airway involvement) and central factors (blunted and/or inadequate chemoresponsiveness) in respiratory insufficiency has not yet been established. To resolve this, lung volumes, arterial blood gas tensions, respiratory muscle strength, breathing pattern and neural respiratory drive were investigated in a group of 15 patients with limb girdle dystrophy. An age-matched normal group was studied as a control. METHODS--Respiratory muscle strength was assessed as an arithmetic mean of maximal inspiratory (MIP) and expiratory (MEP) pressures. Breathing pattern was evaluated in terms of volume (ventilation VE, tidal volume VT) and time (respiratory frequency Rf, inspiratory time TI, expiratory time TE) components of the respiratory cycle. Neural respiratory drive was assessed as the mean inspiratory flow (VT/TI), mouth occlusion pressure (P0.1) and electromyographic activity (EMG) of the diaphragm (EMGd) and the intercostal parasternal (EMGp) muscles. In 10 of the 15 patients the responses to carbon dioxide (PCO2) stimulation were also evaluated. RESULTS--Most patients exhibited a moderate decrease in vital capacity (VC) (range 37-87% of predicted), MIP (range 23-84% of predicted), and/or MEP (range 13-41% of predicted). The arterial carbon dioxide tension (PaCO2) was increased in three patients breathing room air, while PaO2 was normal in all. Compared with the control group Rf was higher, and VT, TI and TE were lower in the patients. EMGd and EMGp were higher whilst VT/TI and P0.1 were normal in the patients. Respiratory muscle strength was inversely related to EMGd and EMGp. PaCO2 was found to relate primarily to VC and duration of illness, but not to respiratory muscle strength. During hypercapnic rebreathing delta VE/delta PCO2, delta VT/delta PCO2, and delta P0.1/delta PCO2 were lower than normal, whilst delta EMGd/delta PCO2 and delta EMGp/delta PCO2 were normal in most patients. A direct relation between respiratory muscle strength and delta VT/delta PCO2 was found. CONCLUSIONS--The respiratory muscles, especially expiratory ones, are weak in patients with limb girdle dystrophy. Reductions in respiratory muscle strength are associated with increased neural drive and decreased ventilatory output (delta VT/delta PCO2). The decrease in VC, together with the duration of disease, influence PaCO2. VC is a more useful test than respiratory muscle strength for following the course of limb girdle dystrophy.