A retrospective study on local tumor control, survival, and complications of conventional irradiation vs. accelerated hyperfractionated irradiation in patients with selected head and neck cancer sites was undertaken. A total of 1,007 consecutive patients treated with radiation alone for cure from 1974-1997 were analyzed. Excluded were female patients, patients with T1 stage of the vocal cord, and patients also treated with brachytherapy implants. There were 637 patients treated with conventional fractionation once daily (QD) in 1974-1997, at a median 2.1 Gy/fraction, to a total median dose of 71.4 Gy in a median overall time of 54 days. As was common before the mid-1980s, 39% and 22% of patients had overall times exceeding 8 and 9 weeks, respectively; 370 patients were treated with accelerated hyperfractionation twice daily (BID) from 1987-1997, at a median of 1.6 Gy/fraction, with an interfraction interval of 4-6 h, to a total median dose of 68 Gy in 40 days. Both schedules were well-balanced with respect to their pretreatment characteristics. Patients were not randomized into QD or BID. The 10-year actuarial probability of local control was 37% vs. 56% for QD and BID, respectively (P < 0.001), which reflects an increase of 19% or a 51% reduction in the local failure rate. Multivariate analysis revealed that T-stage, QD or BID schedule, and overall treatment time were significant independent factors for achieving local tumor control. The 10-year actuarial probability of cause-specific disease-free survival was 25% and 30% for QD and BID, respectively (P = 0.012). Acute morbidity was slightly higher with the BID schedule: patients requiring tube or parenteral feeding were 2.4% for BID and 0.5% for QD (P = 0.01). The 10-year actuarial probability of RTOG/EORTC Grades 3-5 late effects was 13% for both QD and BID. The lack of increase in late complications was most probably due to the lower total dose and dose per fraction in the BID schedule. This study has shown that accelerated hyperfractionated irradiation using two doses of 1.6 Gy each treatment day for less than 6 weeks in advanced head and neck cancer in male patients provides significantly better local tumor control and cause-specific disease-free survival, without increased late morbidity, than conventional fractionation delivered at the previously relaxed overall times of 7 weeks, but sometimes exceeding 8 or 9 weeks. Int.