Abstract
Background : Ketamine is efficient for short-term reduction of suicidal ideas. Predictive factors and outcome trajectories are poorly characterized. Methods : Secondary analyses were conducted on the KETIS study (Abbar et al. BMJ 2022): 156 suicidal patients were randomized to two intravenous infusions of racemic ketamine (0.5 mg/kg) or placebo. Response or remission was assessed over six weeks based on the Beck Scale for Suicidal Ideation (SSI). We calculated i) predictive values of 12 baseline variables on remission ii) outcome courses, and iii) positive (PPV) and negative predictive values. Results : In multivariate analyses, bipolar disorder, lower patient-rated suicidal ideas, and higher physical pain were predictive of suicidal remission at day 3. No clinical factor predicted remission at week 6. Twenty and 24 different clinical courses were identified in early (day 3) and later (week 6) follow-up, respectively, including around 40 % sustained remission, 50 % fluctuating course and 10 % no response. Suicidal remissions at day 1 and day 3 were highly predictive of remissions at day 3 and week 6 (PPV = 96.8 and 92.6 %). Limitations : SSI may not be adapted for rapid variations and repeated measures. Conclusions : Clinical factors were poorly predictive of remission. Fluctuations in suicidal ideas were frequent, even after ketamine (although less than placebo), necessitating vigilance and multimodal care. Remission at day 1 after one infusion was highly predictive of future remission. The benefits of a second infusion will have to be tested.