The research, published in Frontiers in Anesthesiology, was led by a team from the Keck School of Medicine of the University of Southern California and Los Angeles General Medical Center. The investigators mined the TriNetX Research Network, a federated platform that aggregates deidentified records from over 70 healthcare organizations, to compare psychiatric outcomes among adults undergoing vascular surgery between July 2006 and July 2026. Vascular surgery patients were chosen deliberately: they carry a heavy opioid burden — one cited retrospective analysis found opioids were prescribed for 78 percent of common vascular procedures — making them prime candidates for opioid-sparing ketamine regimens, yet their underlying vascular disease also predisposes them to postoperative delirium and cognitive dysfunction.

The study’s design hinged on three cohorts. Patients who received ketamine alone on the day of surgery formed one group; those who received both ketamine and midazolam formed a second; and patients receiving neither drug served as controls. Crucially, the researchers excluded anyone with a preexisting psychiatric diagnosis, as well as patients with chronic pain syndromes, ischemic heart disease, heart failure, or cerebrovascular disease — conditions known to raise the risk of postoperative delirium and cognitive impairment. Anyone who suffered a major adverse cardiac event or died within 31 days of surgery was also removed. The primary outcome was any new psychiatric diagnosis recorded within 30 days after the operation, spanning four domains: psychotic, mood-related, anxiety-related, and behavior-related disorders.

To level the playing field, the team used propensity score matching, a statistical technique that pairs patients across cohorts based on age, sex, race, ethnicity, and socioeconomic indicators. The matching worked remarkably well: after pairing, all measured covariates showed standardized mean differences below 0.01 in the ketamine-versus-control comparisons, meaning the groups were nearly indistinguishable on every variable the database could capture. The ketamine-only cohort, initially a small fraction of the total — roughly 1,400 to 1,800 patients depending on the time window — was matched against equally sized control groups drawn from tens of thousands of vascular surgery patients.

The headline result was stark. Compared with controls, patients who received ketamine alone had roughly doubled odds of receiving a new psychiatric diagnosis within 30 days of surgery — an odds ratio of 2.25 for the composite outcome, with confidence intervals running from 1.81 to 2.79. The behavioral domain was hit hardest, with an odds ratio of 2.79, followed by anxiety at 2.21 and mood disorders at 1.69. These elevations were not fleeting: the same pattern appeared at one day, 14 days, and 30 days after surgery, with absolute event rates in the ketamine group hovering between roughly 17 and 20 percent, compared with 8 to 10 percent in matched controls. Psychotic outcomes could not be reliably analyzed because too few events occurred in either group.

Patients who received ketamine plus midazolam fared somewhat better on the behavioral front. Their odds of a new behavioral diagnosis within 30 days were 1.85 times those of controls — elevated, but noticeably lower than the 2.79-fold risk seen with ketamine alone. When the two ketamine groups were compared head-to-head, the ketamine-only cohort showed 44 percent higher odds of a new behavioral diagnosis (odds ratio 1.44) and 22 percent higher odds of a composite psychiatric diagnosis (odds ratio 1.22) than the combination group. The same behavioral signal appeared at one day and 14 days, suggesting that whatever midazolam is doing, its effect on behavior-related outcomes emerges early and persists.

Yet the combination was far from benign. Compared with controls, patients receiving both drugs still faced roughly doubled odds of new anxiety diagnoses (odds ratio 2.63) and elevated odds of mood and behavioral diagnoses across all time points. In other words, midazolam appeared to blunt — not eliminate — the psychiatric risk associated with ketamine, and the combination group’s overall psychiatric burden remained substantially higher than that of patients who received neither drug. The authors are careful on this point: the data show an association, not proof that midazolam actively protects the brain, and the possibility that midazolam itself carries neuropsychiatric liabilities cannot be excluded.

The pharmacological logic behind the combination is straightforward. Ketamine works primarily by blocking NMDA receptors, producing anesthesia and potent analgesia but also triggering the hallucinations, agitation, and vivid dreams known as emergence phenomena. Midazolam, by contrast, enhances GABA-A receptor signaling — the brain’s main inhibitory pathway — and has long been used clinically to dampen ketamine’s psychotomimetic edge. Prior small trials support this: a randomized emergency department study found reduced recovery agitation when midazolam was coadministered with ketamine, and earlier work showed midazolam attenuates ketamine-induced abnormal perceptions and thought processes. What has been missing is evidence on longer-term psychiatric outcomes, which is precisely the gap this database study attempts to fill.

The study’s limitations are substantial and the authors acknowledge them candidly. TriNetX does not record drug doses, timing, or anesthetic technique, so the researchers could not distinguish between a low sedative dose of midazolam given before induction and a large dose given at emergence. Frailty — a powerful predictor of postoperative delirium and psychiatric morbidity — could not be calculated from the available data, though the cohorts’ mean ages of roughly 50 to 65 suggest a relatively younger population than the groups in which frailty prevalence climbs above 20 percent. Surgical duration, concurrent agents such as propofol or dexmedetomidine, and differential follow-up rates all remain unmeasured confounders. And because the analysis was restricted to vascular surgery, the findings may not generalize to other surgical populations.

Still, the study is the first of its kind to examine midazolam-ketamine coadministration against 30-day psychiatric outcomes in adult surgical patients, and its scale — hundreds of thousands of records across a national network — gives it a statistical power that no single-center trial could match. The authors stop short of recommending the combination as a protective strategy, emphasizing that prospective randomized controlled trials are needed before any change in clinical practice. But for anesthesiologists weighing opioid-sparing regimens against neuropsychiatric risk in a vulnerable vascular surgery population, the message is clear: the drugs patients receive on the day of surgery may echo in their mental health for weeks afterward, and the choice of what to pair with ketamine could matter more than previously appreciated.

Subject of Research: Neuropsychiatric outcomes of ketamine and midazolam coadministration after vascular surgery

Article Title: Neuropsychiatric outcomes of ketamine and midazolam coadministration following vascular surgery: a multicenter database study

Article References: Lee, E., Yang, A., Chwa, J., & Chou, J. (2026). Neuropsychiatric outcomes of ketamine and midazolam coadministration following vascular surgery: a multicenter database study. Frontiers in Anesthesiology, 5, Article 1888859. https://doi.org/10.3389/fanes.2026.1888859

Image Credits: AI Generated

DOI: 10.3389/fanes.2026.1888859

Keywords: ketamine, midazolam, vascular surgery, anesthesia, psychiatric outcomes, postoperative delirium, opioid-sparing, TriNetX, propensity score matching, emergence phenomena, NMDA receptors, benzodiazepines