Rectal fistulas after prostate brachytherapy

To compare the rectal and prostatic radiation doses for a prospective series of 503 patients, 44 of whom developed persistent rectal bleeding, and 2 of whom developed rectal-prostatic fistulas.

The 503 patients were randomized and treated by implantation with 125I vs. 103Pd alone (n = 290) or to 103Pd with 20 Gy vs. 44 Gy supplemental external beam radiotherapy (n = 213) and treated at the Puget Sound Veterans Affairs Medical Center (n = 227), Schiffler Cancer Center (n = 242) or University of Washington (n = 34). Patients were treated between September 1998 and October 2001 and had a minimum of 24 months of follow-up. The patient groups were treated concurrently. Treatment-related morbidity was monitored by mailed questionnaires, using standard American Urological Association and Radiation Therapy Oncology Group criteria, at 1, 3, 6, 12, 18, and 24 months. Patients who reported Grade 1 or greater Radiation Therapy Oncology Group rectal morbidity were interviewed by telephone to clarify details regarding their rectal bleeding. Those who reported persistent bleeding, lasting for >1 month were included as having Grade 2 toxicity. Three of the patients with rectal bleeding required a colostomy, two of whom developed a fistula. No patient was lost to follow-up. The rectal doses were defined as the rectal volume in cubic centimeters that received >50%, 100%, 200%, or 300% of the prescription dose. The rectum was considered as a solid structure defined by the outer wall, without attempting to differentiate the inner wall or contents.

Persistent rectal bleeding occurred in 44 of the 502 patients, 32 of whom (73%) underwent confirmatory endoscopy. In univariate analysis, multiple parameters were associated with late rectal bleeding, including all rectal brachytherapy indexes. In multivariate analysis, however, only the rectal volume that received >100% of the dose was significantly predictive of bleeding. Rectal fistulas occurred in 2 patients (0.4%), both of whom had received moderate rectal radiation doses and extensive intervention for rectal bleeding.

Partly on the basis of data from others and data presented here, we believe that the incidence of rectal fistulas can be much lower than in our series. High rectal radiation doses should be avoided a priori, to minimize the likelihood of rectal bleeding, and hence the likelihood that invasive procedures will be performed.