Between September 1969 and January 1, 1986, 143 pelvic exenterations for recurrent cervical cancer were performed by the gynecologic oncologists at the University of Alabama at Birmingham. Of this group, 78 patients underwent total pelvic exenteration, 63 patients had anterior exenteration, and two had posterior exenteration. The overall operative mortality rate was 6.3%, mostly associated with total pelvic exenteration. The 5-year survival rates were 50% overall, 63% with anterior exenteration and 42% with total exenteration. Univariate and multivariate analyses were performed to identify clinical and histopathologic factors predictive of prolonged survival. Using three clinical factors (duration from initial radiation therapy to exenteration, size of the central mass, and presence of preoperative sidewall fixation), low-, intermediate-, and high-risk groups were constructed; the 5-year survival rates for these groups were 82, 46, and 0%, respectively. Inclusion of one histopathologic factor (margin status of the surgical specimen) added to the ability to predict 2- and 5-year survival rates. The best candidates for cure by pelvic exenteration were those with recurrent small (less than 3 cm), mobile central masses who were a year or longer from the time of their previous radiation therapy. Attempts to resect bulky pelvic recurrences that impinge on the pelvic sidewall, especially in the case of persistent or early recurrent disease (within 6 months), or continuation of exenterative procedures in women known to have nodal metastases or extrapelvic spread, are generally futile. For those women falling between the two extremes, sound clinical and operative judgment is imperative in regard to selecting the treatment offering the best quality of life.