Brain stimulation did not prevent postoperative delirium in older adults

What happened

A randomized clinical trial published August 18 found that weak electrical stimulation delivered through the scalp after surgery did not reduce postoperative delirium overall in adults aged 65 and older. The result matters because delirium can disrupt recovery and threaten cognitive and functional independence after an operation, yet preventive options remain limited.

The MODEST trial tested two forms of transcranial electrical stimulation at University Medicine Greifswald in Germany. Investigators randomly assigned 225 patients undergoing elective surgery to alpha-frequency transcranial alternating-current stimulation, or tACS; transcranial direct-current stimulation, or tDCS; or sham stimulation. The mean age was 73, and the trial excluded people with known neurological or psychiatric conditions and those using certain central nervous system–active medicines.

The intervention was one 20-minute session in the post-anesthesia care unit. The tACS protocol targeted posterior parietal regions with a 9.5-Hz alternating current, while tDCS targeted the left dorsolateral prefrontal cortex. Both active protocols used a 2-mA current. Sham stimulation produced an initial scalp sensation before the current was stopped.

The primary result was negative

The primary endpoint was at least one positive delirium assessment on the surgical ward during as many as five postoperative days, measured twice daily with the 3-minute Confusion Assessment Method. Although 225 patients were randomized, 180 were included in the main analysis.

Ward delirium occurred in 11.5% of the analyzed tACS group, compared with 9.2% of the sham group. The adjusted odds ratio was 1.60, with a 95% confidence interval from 0.47 to 5.46. That wide interval included both possible benefit and possible harm, and the difference was not statistically significant.

The secondary comparison was also negative: delirium occurred in 14.3% of the tDCS group and 9.2% of the sham group. Neither active protocol produced a significant reduction in delirium in the post-anesthesia care unit. Delirium severity, motor presentation and postoperative pain also did not differ substantially among the groups.

Reported adverse-event ratings were not significantly different, and the investigators reported no serious adverse events. Those observations are reassuring within this trial but do not establish safety across larger, more medically diverse surgical populations.

A subgroup signal needs another trial

A prespecified analysis suggested that tACS might reduce delirium in the recovery unit after longer operations. The interaction between surgical duration and treatment reached nominal statistical significance, and model-derived estimates favored tACS for procedures lasting three to five hours.

That finding should not be treated as a positive efficacy result. The primary endpoint was negative, relatively few participants developed delirium, and the researchers did not adjust the subgroup analyses for multiple testing. The interaction’s confidence interval was also close to the null boundary. At most, it identifies a narrower hypothesis for a properly powered confirmatory trial.

Why the evidence remains limited

The study was double-blind and sham-controlled, design features that strengthen causal inference. However, it was conducted at one German center across a heterogeneous mix of operations. The sham group’s 9.2% ward-delirium rate was substantially below the 22.2% rate used in planning the sample size, leaving the trial underpowered to detect the expected effect. The authors estimated only 46% power to detect a reduction comparable to an earlier study.

Clinical assessments may also miss subtle neurological changes, and a single stimulation session does not resolve whether another timing, repeated sessions or individualized targeting would perform differently. The practical conclusion is narrower: this protocol should not be described as an established way to prevent postoperative delirium. The longer-surgery signal warrants testing, but it does not overturn the overall null result.