Objective: This longitudinal study examined whether high frequency heart-rate variability (HF-HRV) and HF-HRV reactivity to stress moderates response to cognitive behavioural therapy for insomnia (CBT-I) within a stepped-care framework in cancer patients with comorbid insomnia.
Methods: 177 participants (86.3% female; Mage=55.3, SD=10.4) were randomized to receive either stepped-care or standard CBT-I and were followed for 12 months following treatment. HRV measures were assessed at pre-treatment during a rest and worry period. Insomnia symptoms were assessed using the Insomnia Severity Index (ISI) and daily sleep diary across five timepoints. 
Results: Resting HF-HRV significantly predicted pre-treatment sleep efficiency but not ISI score. No significant time x HF-HRV or CBT-I group x time x HF-HRV interactions were found, indicating that HF-HRV does not predict differential responses to the different CBT-I group. HRV reactivity was not cross-sectionally or longitudinally related to any outcome variables. In exploratory analyses, significant insomnia severity x time x HF-HRV interactions were observed, suggesting that HF-HRV may predict treatment responses differently based on initial insomnia severity. 
Conclusion: Although resting HF-HRV was related to initial sleep efficiency, HF-HRV measures did not significantly predict response to either form of CBT-I. Resting HF-HRV may predict certain treatment outcomes when initial insomnia severity is considered, however these results are exploratory and of unclear clinical significance.