Purpose: The purpose of this investigation was to determine the irradiation tolerance level and complication rates of the proximal vagina to combined external irradiation and low dose rate (LDR) brachytherapy. Also, the mucosal tolerance for fractionated high dose rate (HDR) brachytherapy is further projected based on the biological equivalent dose (BED) of LDR for an acceptable complication rate. Materials and methods: Two hundred seventy-four patients with stages I-IV cervical carcinoma treated with irradiation therapy alone from 1987 to 1997 were retrospectively reviewed for radiation-associated late sequelae of the proximal vagina. All patients received LDR brachytherapy and 95% also received external pelvic irradiation. Follow-up ranged from 15 to 126 months (median, 43 months). The proximal vagina mucosa dose from a single ovoid (single source) or from both ovoids plus the tandem (all sources), together with the external irradiation dose, were used to derive the probability of a complication using the maximum likelihood logistic regression technique. The BED based on the linear-quadratic model was used to compute the corresponding tolerance levels for LDR or HDR brachytherapy. Results: Grades 1 and 2 complications occurred in 10.6% of patients and Grade 3 complications occurred in 3.6%. There were no Grade 4 complications. Complications occurred from 3 to 71 months (median, 7 months) after completion of irradiation, with over 60% occurring in the first year. By logistic regression analysis, both the mucosal dose from a single ovoid or that from all sources, combined with the external irradiation dose, demonstrate a statistically significant fit to the dose response complication curves (both with P=0.016). The single source dose was highly correlated with the all source dose with a cross-correlation coefficient 0.93. The all source dose was approximately 1.4 times the single source dose. Over the LDR brachytherapy dose rate range, the complication rate was relatively stable to small variations of the underlying tumor biological characteristics and the dose rate. The complication rates change approximately an absolute 1% over the range of the alpha-beta ratio (α/β) from 2 to 4 Gy and repair constant (μ) of 0.46/h to 0.60/h. The complication rates increased an absolute 2% over the mucosa dose rate from 1.75 to 3.50 Gy/h. They markedly increased as the dose rate increased above 3.00 Gy/h as in HDR brachytherapy. The projected HDR Grade 3 tolerance varied from 25 Gy for one fraction to 57 Gy for six fractions in addition to 20 Gy external irradiation for nominal 3-5% complication rates. The traditional LDR tolerance dose of 150 Gy was shown to yield nominal 11% and 4% Grades 1 and 2 and Grade 3 sequelae, respectively. Conclusions: The traditional 150 Gy LDR tolerance dose (single source plus external irradiation) can be relaxed to 175 Gy or equivalently a full mucosal dose of 238 Gy (all sources plus external irradiation) for a nominal 5% Grade 3 complication rate. Higher fractionation is necessary with four to six fractions in HDR therapy for similar rates of sequelae. The mucosal surface dose from a single ovoid, which can be readily computed, remains a convenient tolerance check for treatment planning purposes.