Background and purpose: There is still a concern that the use of HDR brachytherapy might result in an increase of late tissue damage. This retrospective study evaluates the incidence and severity of late complications in patients with carcinoma of the cervix who underwent combined external beam radiation (EBR) and Ir-192 HDR brachytherapy and attempts to identify pretreatment and treatment parameters correlating with late complications. Material and methods: Between 1985 and 1992, 161 patients with carcinoma of the cervix (FIGO stages IB–IVB) received EBR to the pelvis (ave. max. dose 48.8 Gy) followed by 1–6 Ir-192 HDR placements (median 2). Doses to point A ranged from 8.5 to 38.7 Gy (median 17 Gy). Parameters examined included age, diabetes, obesity, history of inflammatory bowel disease or diverticulitis, prior surgery, hemoglobin level, FIGO stage, EBR dose, technique and daily dose fraction, number of HDR treatments and total dose to point A, maximum doses to bladder and rectum delivered by brachytherapy and cumulative dose to point A. Median follow-up for all patients was 37 months. Complications were rated using an in-house scoring system and according to the French-Italian Glossary (FIG). Results: Actuarial 5-year survival was 93%, 57%, 46%, and 0% for stages IB, II, IIIB, and IV, respectively. Of 161 patients, 11% developed moderate and 3.7% severe sequelae (FIG: 2.5%, 3.7%). Since some patients experienced more than one complication, the overall incidence was 13.6% and 4.9% (FIG: 3.1%, 4.9%) with respective 5-year actuarial rates of 14% and 5% for moderate, and 2% and 8% for severe bowel and genitourinary tract complications (FIG: 3.5%, 0, and 2%, 8%). All severe bowel complications occurred within 1.5 years whereas urinary tract sequelae continued to develop throughout the follow-up period. FIGO stage was associated with a significant increase in late sequelae ( P=0.015). Analysis of the remaining pretreatment and treatment parameters failed to reveal any statistically significant correlation with moderate or severe sequelae. Conclusion: In our series using HDR brachytherapy, complication and survival rates were comparable with other series employing either LDR or HDR procedures. Of all parameters analysed, stage of disease was the only parameter significantly correlated with complications in univariate and multivariate analysis.