Ketamine is associated with improvement of depression and neuropathic pain when sub-anesthetic doses are used. The NMDA receptor is involved in learning and memory, and antagonists such as ketamine may be associated with cognitive impairment. The objective of this study is to investigate whether repeat outpatient ketamine infusion is associated with cognitive dysfunction.


With IRB approval, patients undergoing repeat outpatient ketamine infusions were asked to complete a Manos 10 point clock test before and after each ketamine infusion and on each follow up clinic visit. One visit includes 3 infusions, one per day. The clock is scored using a 10 point clock drawing criteria, based on how well a patient can draw a clock positioned at a specific time. Intact cognition is represented by a score of 10, probable cognitive deficit, less than 8, and severely impaired at less than 4. Pain scores were also obtained pre and post infusion and on follow up visits.

Fixed effects mixed model regression was used to test the effect of visit number in patients that had at least 2 visits. Visits 2 through 7+ were compared with visit 1 for each patient. The results were adjusted for age, race, gender, BMI, pain, ketamine dose (mg/kg), opioid use, and antidepressant use. Pre and post ketamine clock scores were averaged for each visit. Analysis was also performed using averages of only pre scores and averages of only post scores.


148 patients received ketamine infusion, however most patients (n=95, 64%) had just 1 visit (i.e., 3 days of infusion). Mean clock scores and pre scores showed decrease cognition with repeat visits, however not significant. Post scores showed a significant decline in cognitive function with repeat infusion visits (p=.03 for visit effect) after adjusting for gender, race, age, bmi, pain, ketamine dose, opioids, and antidepressants. No visit effect was seen in follow-up clock scores.

There were large differences in the numbers of patients who had 1 versus 2 versus more than 2 visits. Subject differences at each visit, rather than the visit number itself, may explain cognition differences. Therefore, we also made this comparison using a fixed effects mixed model regression to test the effect of visit number, using each patient as their own control, using patients with 2 or more visits (n=53, 36%).

In this model, we found that after adjusting for covariates, visit had a strong significant effect (p=.003). When patients came for repeat infusions, their average cognitive function level (mean of pre and post scores) declined from their initial infusion visit. Using pre or post clock scores, or follow-up scores, results were similar, with declines at all visits subsequent to visit 1.

In the fixed effects mixed model, visits where antidepressants were present had lower mean scores (p<.0001), compared to visits where antidepressants were absent, for the same patient. The effect of opioids was not significant (p=.31). Across patients, the antidepressant effect was significant at a trend level (p=.055). Visits where patients had used antidepressants had lower post clock scores after adjusting for covariates. Using scores measured at pre-infusion each day, we found a significant effect of antidepressants (p=.0498). The mean drop in pre clock scores for visits with antidepressants present was 1.5 points.

Level of pain and opioid use was not associated with clock scores in any of the analyses.


Ketamine use maybe associated with decreased cognition. In a prior pilot study, repeat ketamine infusions showed decreased clock scores only after the fourth visit, whereas this study shows a significant decrease after the second visit. Antidepressant use may have a negative effect on cognition, which may indicate that depression may also influence cognition. Pain level and opioid use do not appear to be associated with changes in cognition.