Non-convulsive status epilepticus, often abbreviated NCSE, is one of the most elusive emergencies in neurology. Unlike the generalized convulsive form of status epilepticus, in which a patient is visibly wracked by continuous seizures, NCSE produces only subtle signs: confusion, unresponsiveness, a vacant stare, or an unexplained failure to wake up. The seizure activity is hidden inside the brain and can only be detected with electroencephalography, or EEG, which records the brain’s electrical rhythms through electrodes on the scalp. The growing availability of continuous EEG monitoring in intensive care units has led to far more diagnoses of NCSE in recent years, particularly among critically ill patients whose only symptom might be that they simply do not regain consciousness.

For generalized convulsive status epilepticus, escalating to therapeutic coma — the continuous administration of anesthetic drugs such as propofol, midazolam, barbiturates or ketamine, sometimes combined with inhalational agents like sevoflurane — is an established and often life-saving strategy when conventional antiseizure medications fail. The goal is to suppress electrical activity so deeply that the brain enters a pattern called burst suppression, visible on the EEG, in the hope of halting the neuronal storm before it inflicts lasting damage. But whether the same aggressive approach benefits patients with the non-convulsive form of the condition has long been contested. Current guidelines explicitly emphasize that treatment in NCSE should be individualized and driven by the underlying cause, noting that anesthetic therapy may not always be required. In practice, however, therapeutic coma is still frequently deployed in refractory cases.

To test the risk-benefit balance, the research team led by Eyad Altarsha and Kristian Barlinn at Technische Universität Dresden analyzed 283 adults treated for NCSE at their center between January 2010 and March 2020. Of these, 111 patients — 39.2 percent — received therapeutic coma, defined as continuous intravenous or inhalational anesthetic administration aimed at seizure control. The remaining 172 were managed without it. The researchers tracked a predefined set of clinically relevant in-hospital complications, including pneumonia, sepsis, cardiac arrhythmias requiring intervention, acute kidney injury, the need for renal replacement therapy, venous thromboembolism, and cardiopulmonary resuscitation. Secondary outcomes included in-hospital mortality, length of intensive care stay, and whether the status epilepticus was successfully terminated before discharge.

Because doctors tend to reserve therapeutic coma for the sickest patients, a simple comparison would be deeply misleading — a problem statisticians call confounding by indication. To address this, the team used a technique known as inverse probability of treatment weighting based on propensity scores. In essence, they built a statistical model estimating each patient’s likelihood of receiving therapeutic coma based on age, sex, seizure etiology, whether the patient was comatose at presentation, a history of epilepsy, and pre-admission antiseizure medication use. They then reweighted the two groups so that, on paper, patients who received the coma and those who did not were statistically comparable. After weighting, all baseline characteristics were well balanced, with standardized mean differences below 0.1, allowing a fairer estimate of the treatment’s association with outcomes.

The results were stark. In the weighted analyses, therapeutic coma was associated with roughly twentyfold higher odds of pneumonia, sixfold higher odds of sepsis, fourteenfold higher odds of clinically significant cardiac arrhythmias, and more than fourfold higher odds of acute renal failure. The need for renal replacement therapy was nearly five times more likely, and the odds of requiring cardiopulmonary resuscitation were more than eleven times higher. The association with venous thromboembolism showed a similar direction but narrowly missed statistical significance. These patterns are biologically plausible: deep sedation, mechanical ventilation, and prolonged immobilization are all well-recognized drivers of intensive care complications, from ventilator-associated pneumonia to cardiovascular instability.

Crucially, the excess harm did not appear to be explained by ventilation alone. When the researchers restricted their analysis to the subgroup of patients who were intubated in both groups, therapeutic coma remained independently associated with higher risks of pneumonia, sepsis, and cardiac arrhythmias, along with a longer intensive care stay. A further sensitivity analysis limited to patients treated with three or more antiseizure medications — a pragmatic proxy for genuinely refractory disease — produced even more dramatic estimates, with mortality remaining significantly elevated and some complications, such as cardiac arrhythmias and cardiac arrest, occurring exclusively in patients exposed to therapeutic coma. Excluding patients with hypoxic NCSE, a particularly severe variant, did not change the picture.

The study also uncovered a troubling dose-response relationship. Within the therapeutic coma group, every additional hour under anesthesia increased the odds of pneumonia by about 2 percent, which translates to an approximately 89 percent increase in pneumonia odds for each additional 24-hour cycle of coma. Sepsis risk rose by roughly 46 percent per additional 24-hour cycle. Meanwhile, mortality told its own story: crude in-hospital death rates were 43.2 percent in the coma group versus 22.7 percent without coma, and after weighting, therapeutic coma carried a relative risk of death of 1.86 — an absolute mortality increase of roughly 24 percentage points. Intensive care stays were approximately four times longer. Notably, an analysis of the circumstances of death suggested the excess was not driven by earlier withdrawal of care: in the coma group, deaths were split almost evenly between withdrawal of life-sustaining treatment and death under ongoing intensive care from complications, whereas among patients managed without coma, the vast majority of deaths followed explicit treatment-limitation decisions.

Against all this harm, the expected benefit failed to materialize. Successful termination of NCSE before hospital discharge occurred in only 70.3 percent of patients treated with therapeutic coma, compared with 84.3 percent of those managed without it — and after adjustment, therapeutic coma was actually associated with a lower likelihood of seizure resolution. The authors are careful to stress the limits of their design. This was a retrospective, single-center study; the decision to induce coma was made at physician discretion and may have been influenced by unmeasured factors such as perceived refractoriness or systemic instability; NCSE diagnoses relied on contemporaneous routine EEG interpretation rather than centralized re-adjudication; and long-term functional and cognitive outcomes were not available. Residual confounding by disease severity cannot be excluded, and the findings should not be read as proof that treatment escalation is always futile.

Nevertheless, the consistency of the associations across multiple statistical approaches, sensitivity analyses, and outcome measures lends the results considerable weight, and they sharpen an increasingly important distinction in neurocritical care: what saves the brain in convulsive status epilepticus may harm it in the non-convulsive form. The definitive answer will have to come from randomized evidence, and the ongoing FAST randomized controlled trial — comparing acute sedation in intensive care against high-dose intravenous antiseizure medication for non-convulsive status epilepticus — is expected to deliver exactly that. Until then, the study suggests clinicians should think twice before reaching for the anesthetic dial in patients whose seizures announce themselves not through convulsions, but only through the quiet electrical whispers of an EEG trace.

Subject of Research: Association of therapeutic coma with complications and mortality in non-convulsive status epilepticus

Article Title: Therapeutic coma in non-convulsive status epilepticus is associated with increased complications and mortality: a retrospective cohort study

Article References: Altarsha, E., Diel, N. J., Braun, M., Schoene, D., Kaadan, A., Moustafa, H., Winzer, S., Sedghi, A., Siepmann, T., Lesser, M., Huttner, H. B., & Barlinn, K. (2026). Therapeutic coma in non-convulsive status epilepticus is associated with increased complications and mortality: a retrospective cohort study. Journal of Neurology, 273(10), Article 563. https://doi.org/10.1007/s00415-026-14090-8

Image Credits: AI Generated

DOI: 10.1007/s00415-026-14090-8

Keywords: non-convulsive status epilepticus, therapeutic coma, status epilepticus, neurocritical care, electroencephalography, pneumonia, sepsis, in-hospital mortality, propensity score weighting, intensive care, anesthetic agents, retrospective cohort study