Active compression-decompression cardiopulmonary resuscitation (ACD-CPR) has been introduced to improve outcome of CPR after cardiac arrest. Usually, ACD-CPR is performed with the rescuer kneeling beside the patient (ACD-B), but ACD-CPR with the rescuer in standing position (ACD-S) has been taught and applied in some centres in addition to conventional ACD-CPR (ACD-B). The aim of this randomised and cross-over study was to evaluate the new technique of ACD-S and to compare it with conventional ACD-B. Twelve professional rescuers (aged 30.8 +/- 7.9 years) applied both methods of ACD-CPR on a manikin. We obtained the following results. (1) Duration of CPR performance was comparable for ACD-S (13.2 +/- 7.1 min) and ACD-B (15.5 +/- 10.2 min, P = 0.48). (2) Pain in the upper extremity and pain in the vertebral column were the main reasons for break-off by the rescuers. Exhaustion was judged to be similar during ACD-S (5.3 +/- 2.3) and ACD-B (6.2 +/- 2.1; on a rating scale with 1 = no and 9 = complete exhaustion). (3) Oxygen consumption was significantly higher during ACD-S (P < 0.005), whereas heart rate and lactate levels did not differ. (4) Decompression forces were lower than compression forces. The averaged decompression forces in both methods were similar during the first 2 min and the last min. Compression forces decreased in ACD-S from 55.1 to 48.9 kp (P = 0.002) and in ACD-B from 52.8 to 47.0 kp (P = 0.069). We conclude that ACD-CPR in standing position can be considered equal to ACD-B in view of maximal duration of CPR, exhaustion of the rescuers and decompression forces. The decrease of compression forces in ACD-S and ACD-B as well as the difference between compression forces in ACD-S and ACD-B seem to be of no clinical relevance, and exhaustion was judged to be similar despite oxygen consumption being higher in ACD-S than in ACD-B.