We compared 284 sting-allergic patients treated with either a 4 day (group 1), 6 hr (group 2) or 2 1 0 min (group 3) rapid venom immunotherapy (RVIT) protocol using honey bee (HB) or yellow jacket (YJ) venom at cumulative doses of 527.6 mug, 226.6 mug and 101.1 mug respectively. The 4 day protocol involved four times as many injections as the 2 1 0 min protocol and twice as many as the 6 hr protocol. Desensitization was conducted in a hospital providing full emergency resuscitation facilities. In group 1, 1 x 100 mug boosters were given on days 7, 10, 15 and 45 and, in groups 2 and 3, 2 x 50 mug boosters were given on day 15 and 1 x 100 mug on day 45. The patients in the three groups were comparable with regard to clinical characteristics and immunological reactivity determined by skin tests. All patients had large local reactions. Systemic reactions (SR) occurred in 28.2% of patients in group 1, 28.6% in group 2 and 6-9% in group 3. The mean total cumulative venom dose (s.e.m.) for occurrence of SR was 123.75 (+/-24.2) in group 1, 183.27 (+/- 28.5) in group 2, and 36-43 (+/-9.3) in group 3. HB led to more systemic reactions than YJ venom. The rate of SR decreased when the cumulative venom dose was reduced during RVIT. The median dose was 137-6 mug in group 1, 226.6 mug in group 2, and 21.1 mug in group 3. No systemic reactions were observed after the booster injections. The results of this study suggest that short RVIT protocols with low cumulative doses carry a lesser risk of SR.