We developed a new method to score sleep fragmented by respiratory events that we call the T-sleep (transitional sleep scoring) method. Five control polysomnograms from subjects without sleep-related breathing abnormalities were scored by the Rechtschaffen and Kales (R&K) method, and 10 polysomnograms from patients with severe obstructive sleep apnea were scored by both the R&K and the T-sleep method. Comparative analyses were performed on sleep variables of the control and apnea polysomnograms, and interscorer correlations were assessed for sleep and apnea variables. The interscorer correlations were high for both R&K control scoring and for apnea recordings scored by the T-sleep method. The number of sleep stage events documented for the 10 apnea recordings was significantly less for the T-sleep method than the R&K method (36, SD 17.0 vs. 332, SD 144.0; p = 0.0002). The T-sleep method was shown to be an effective, accurate and quick method for scoring sleep in patients with sleep-related breathing disorders.