OBJECTIVES To assess the efficacy and risks of laparoscopic para-aortic lymph node sampling compared with standard laparotomy in staging cervical carcinoma. METHODS From August 1993 through July 1994, 38 patients with biopsy-proven invasive cervical carcinoma (24 early and 14 advanced invasive cancers) were entered into the study. This was a prospective study of laparoscopic para-aortic lymphadenectomy in invasive cervical carcinoma, with patients serving as their own controls. Video laparoscopic lymph node sampling was performed. In patients with early invasive cancer, the nodes were sent for frozen section and, if negative, laparotomy was performed to look for any residual nodes. Radical hysterectomy was performed immediately if residual nodes were negative. Patients with either nodal metastasis on frozen section or with advanced cancer underwent para-aortic lymphadenectomy only. The operative technique was also evaluated. RESULTS Laparoscopy required an average of 77 min (S.D. 40), with an average blood loss of 116 ml (S.D. 321). The average number of nodes was 15 (S.D. 7). At subsequent laparotomy the average number of residual nodes found was 0.4 (S.D. 0.9) and none showed metastasis. One vena cava laceration and one ureteral injury required immediate repair, and two patients were too obese to undergo laparoscopy. CONCLUSIONS Laparoscopic para-aortic lymph node sampling is a less invasive, reliable method for staging invasive cervical carcinoma and can substitute for traditional open procedures. The incidence of risks with this method appears to be low.