Insomnia treatment response as a function of objectively measured sleep duration

To examine the potential moderating effect of objectively measured sleep duration at baseline on the response to cognitive behavioral therapy for insomnia (CBT-I), administered singly or combined with medication (CBT-I + Med).

Based on the average PSG-derived sleep duration across two baseline nights and the type of treatment received, 159 adults with insomnia (50.3 ± 10.1 years; 61.0% women) were classified into one of four groups: participants with short sleep duration (ie, ≤ 6 h) treated with CBT-I (n = 26) or CBT-I+Med (n = 25), and participants with normal sleep duration (ie, > 6 h) treated with CBT-I (n = 54) or CBT-I+Med (n = 54). Primary outcome measures were sleep/wake parameters derived from a sleep diary and insomnia severity and secondary outcomes were beliefs about sleep, fatigue, depression and anxiety.

Patients with both short and normal sleep durations at baseline improved significantly on most sleep continuity parameters with CBT-I administered singly or combined with medication. Irrespective of treatment received, participants with short sleep duration also showed significantly greater improvements in subjective sleep (ie, reduced wake after sleep onset, increased sleep efficiency) relative to those with normal sleep duration. Conversely, participants with normal sleep duration showed greater improvements on some measures of daytime functioning and sleep satisfaction.

There was no moderating effect of baseline sleep duration on treatment response to cognitive behavioral therapy. Despite some marginal differential treatment response on selected daytime functioning outcomes, the benefits from CBT-I were not significantly different as a function of short or normal sleep duration at baseline. Further prospective investigation of insomnia phenotypes taking into account other variables than sleep duration is warranted in order to develop more targeted insomnia therapies.

www.clinicaltrials.gov (#NCT00042146).