Using Classification and Regression Tree Modeling to Investigate the Effects of Subanesthetic Ketamine Infusion on Remission of Suicidal Symptoms

Background: As accumulating evidence suggests the antisuicidal properties of ketamine, elucidating its underlying mechanisms and predictors of treatment response has become crucial. Whether a combination of clinical markers can enhance the prediction of the antisuicidal response to ketamine remains unclear. Methods: Using data from our previous randomized placebo-controlled and open-label trials, this post hoc analysis evaluated 67 patients with treatment-resistant depression and prominent suicidal ideation who received a single ketamine infusion of 0.5 mg/kg. A classification and regression tree model was used to identify the combinations of clinical and demographic characteristics at baseline that can predict the antisuicidal response to ketamine infusion. Results: Combinations of clinical predictors, including mild or moderate depression severity, a shorter duration of the current episode, no more than four trials of antidepressant failures, low or moderate current suicide risks, and a history of suicide attempts, offered superior predictive power for the rapid and sustained antisuicidal effects of sub-anesthetic ketamine infusion, outperforming the predictive power of any individual predictor. Conclusions: Clinicians can use our findings to identify individuals who are most likely to benefit from the antisuicidal effects of ketamine. Further research is required to corroborate these findings. Social Media Abstract: CART found combined baseline markers, including mild/moderate depression, shorter episode, ≤4 failed antidepressants, low/moderate suicide risk, and prior attempts, predict rapid/sustained antisuicidal response of sub-anesthetic ketamine in TRD.