We studied the effects of elective hip surgery, performed under either spinal (SA, n = 10) or general anesthesia (GA, n = 10), on breathing pattern and gas exchange. Measurements were made with respiratory inductive plethysmograph and indirect calorimetry in two positions before and after surgery. The method of anesthesia had no effect on the severity of postoperative hypoxemia. Reduced arterial oxygenation (PaO2; P less than 0.001, SA from 12.5 +/- 2.37 kPa to 10.5 +/- 1.38 kPa, GA from 12.5 +/- 2.95 kPa to 10.5 +/- 1.75 kPa) despite increased alveolar ventilation (P less than 0.01; from 2.30 +/- 0.37 l/min to 2.39 +/- 0.43 l/min in SA, 2.27 +/- 0.56 l/min to 2.57 +/- 0.35 l/min in GA) and reduced arterial carbon dioxide partial pressure (PaCO2; SA from 5.20 +/- 0.22 kPa to 4.95 +/- 0.33 kPa, P less than 0.01, GA from 5.07 +/- 0.36 kPa to 4.72 +/- 0.41 kPa, P less than 0.05) indicated maldistribution of ventilation and perfusion. Changes in breathing pattern and gas exchange and differences between the groups were minimal. Minute ventilation, tidal volume and mean inspiratory flow remained unchanged in both groups. The contribution of rib cage to tidal volume increased postoperatively in the supine position (P less than 0.001; SA from 32.6% +/- 10.3 to 46.3% +/- 7.5, GA from 36.5 +/- 16.4 to 48.5% +/- 15.4). CO2 production, oxygen consumption and energy expenditure remained unchanged. The postoperative changes in breathing pattern are related to the operation, not to the type of anesthesia and do not explain the alterations in gas exchange.