Status epilepticus (SE), defined as either continuous seizure activity or two or more sequential seizures without recovery of consciousness between seizures, is a medical emergency associated with a significant morbidity and mortality. In some patients with SE, seizures do not respond to antiseizure medications (ASMs) and require anesthetic therapy. If seizures persist 24-h post initiation of anesthesia or reoccur after anesthesia is weaned, it is considered super-refractory status epilepticus (SRSE) (1), which may result in a delayed, prolonged increase in intracranial pressure, cardiorespiratory and metabolic crisis, and a substantial mortality rate of 23–57% (2, 3).

Current available evidence for the treatment of SRSE is lacking. In the case of drug treatment failure, immunotherapy such as hormones, gamma globulin, and plasma exchange, may be given when the etiology of SRSE is presumed to be autoimmune. Alternative therapies such as hypothermia, surgical resection, ketogenic diet therapy and neuromodulation (including vagus nerve stimulation, deep brain stimulation, electroconvulsive therapy and transcranial magnetic stimulation) may also be tried, all with varying levels of success (1).

Among them, the ketogenic diet (KD) is high-fat, low-carbohydrate diets that mimic a fasting state and induce ketone body production through fat metabolism. The exact mechanism of KD is still controversial, which is the focus of current research. The antiepileptic effect of KD may be related to a variety of mechanisms, including neurotransmitters, ion channels, g protein-coupled receptors, gut microbiota, mitochondria, DNA methylation, inflammation, and so on (4–10). Studies have shown that, ketone bodies may act by activating various synergistic pathways, which are still being evaluated and chronic ketosis is thought to stabilize and reduce synaptic hyperexcitability in order to conserve energy (11). Moreover, The KD has been shown to exhibit antiepileptogenic properties (12, 13) and has become a well-established treatment for pediatric drug-resistant epilepsy. Recently, emerging evidences have indicated that the KD may also be beneficial for SRSE (14–21), not only for children but also for adults. However, most evidence has come from case reports, case series, or retrospective studies (14, 16, 18, 19). Currently, there are limited prospectively collected data on practical considerations and long-term outcomes of KD in patients with SRSE in the intensive care units (ICUs), such as a study of ketogenic diet treatment of SRSE in the pediatric intensive care unit confined to children (21). Moreover, some studies relied on only urinary ketosis (17–19), which does not correlate well with seizure control (22–24). In ICUs, coexistent medical problems specific to critically ill patients may act as obstacles during KD therapy. Such challenges include decreased gastrointestinal (GI) function due to a low level of consciousness and immobility; prolonged coma therapy; increased susceptibility to infection; malnutrition and metabolic disorders; and complex interactions between the KD and concomitant treatments. Therefore, many issues remain to be addressed for the management of these SRSE patients in the critical care setting, and more data are required to assess the practice of the KD therapy in adults with SRSE patients in ICUs. Here, we carried out a single-center prospective study of the administration of the KD to patients who had SRSE in the ICU, and then presented our findings about the efficacy and adverse effects of the KD. Simultaneously, short-term (during hospitalization) tolerability and long-term (3 months after KD) tolerability are also evaluated.

From July 2020 to April 2024, we prospectively recruited 12 patients (9 females and 3 males) with SRSE and implemented ketogenic treatment. The demographic and baseline characteristics are described in Table 1. All patients were diagnosed with SRSE before KD initiation. Eight patients were diagnosed with cryptogenic new-onset refractory status epilepticus, two with virus encephalitis, two with encephalopathy. The median SE onset age was 34 (range 16–69, interquartile range [IQR] 18–52) years, with a median admission GCS score of 5 and a median FOUR score of 5. Prior to KD therapy, all patients were in a comatose state and 11 of them were assisted with mechanical ventilation. The median duration of SE prior to the KD was 21 days (range 4–46).

Our experience supports that KD is applicable in the management of SRSE patients in the ICU. In the vast majority of critically ill SRSE patients, ketogenesis can be achieved and has a positive effect on seizure control, including resolution of SE, and weaning off of anesthetic agents, intravenous ASMs and mechanical ventilators. The KD may also cause numerous adverse effects in SRSE patients in the ICU, including gastrointestinal intolerances, malnutrition and metabolic abnormalities. However, most KD-related adverse reactions can be reversed after being carefully monitored and actively managed.

To assess whether those patients in the ICU can achieve ketosis and the level of ketosis at which SE is resolved, we utilized both urine ketones and serum beta-hydroxybutyrate as monitoring measures. Similar to previous findings, ketosis was achieved at median of 2 days (20). A study on a modified low-proportion KD in adults with SRSE reported an 83% resolution rate after a median of 9 days post-KD initiation (range 2–21) (25). However, suboptimal ketosis levels (<1.5 mmol/L) are frequently encountered (26). In our cohort, patients 6 and 10 failed to achieve adequate ketosis and consequently did not experience SE resolution following KD therapy. Intriguingly, despite initial insufficient ketosis in patients 3, 4, 5, 11, and 12, they all resolved SRSE and were successfully weaned off anesthetics, suggesting that an optimal ketosis level may not be a prerequisite for KD efficacy in these cases (25). Lowe et al. (27) demonstrated that children on a MCT-KD achieved similar seizure control despite lower urinary ketone levels. Some studies have indicated that the KD can more efficiently eliminate glutamate, the predominant excitatory neurotransmitter, and convert glutamine to γ-aminobutyric acid (GABA), the primary inhibitory neurotransmitter, thereby exerting an anti-seizure effect (28). The mechanisms by which seizures are controlled despite insufficient ketosis are not fully elucidated and may involve direct effects of ketone bodies on neuronal excitability, alterations in energy metabolism, and other potential biological mechanisms. Therefore, if ketones are low but the patient shows clinical benefit such as freedom from seizures, further action may not be required.

Previous observational studies have documented the efficacy of the KD in treating SE, with a success rate of 82% in achieving SE cessation across 31 patients (29). However, the management of SRSE in the ICU setting diverges from that of other epilepsy patients, necessitating further prospective data collection. A recent study reported a 50% response rate to KD within 7 days in eight pediatric patients with SRSE in the pediatric intensive care unit (PICU) (21). Despite this, prospective data on adult patients remain scarce. In our prospective study, significant clinical improvement was observed in nine out of 12 patients (adults and adolescents) who resolved SE following KD initiation, despite prior treatment failures. In recent studies, the authors found that resolution of SE usually occurred within 7 days (15, 30, 31). Similarly, we resolved of SE within a median of 3 days (range 1–18) after initiating KD. Notably, all nine KD-responsive patients were successfully weaned off anesthetic agents and ASMs following KD initiation. This underscores an additional potential advantage of KD in the treatment strategy for RSE: a reduction in the number of medications required for seizure management.

The safety of the KD diet in critically ill patients with RSE warrants careful consideration. KD therapy is predominantly administered via the enteral route. A meta-analysis encompassing 13 KD studies with 932 participants with drug-resistant epilepsy revealed that the most frequently reported adverse effects were vomiting, constipation, and diarrhea (32). In our cohort, gastrointestinal intolerances were observed in 58.3% of patients, manifesting as mild constipation, gastric retention and diarrhea. Given that critical RSE patients are often comatose and subjected to prolonged use of ASMs, anesthetics, and mechanical ventilation, they are predisposed to heightened gastrointestinal complications. In such scenarios, the introduction of a high-fat diet to critically ill SRSE patients may exacerbate their condition, potentially undermining the efficacy of the diet and ketosis development, as exemplified by patient 6, thus rendering the KD potentially unsafe and inefficient. For patients with impaired gastric motility, preemptive placement of a nasojejunal tube was employed to mitigate gastric retention. Moreover, administering the KD at a continuous rate minimized the risk of gastrointestinal issues. KD has been shown to affect glucose and lipid metabolism in murine models, as confirmed by Li et al. (33). Although eight patients in our study developed abnormal lipid levels, no instances of hypoglycemia were recorded, possibly attributed to the continuous feeding strategy and maintenance of adequate pancreatic and liver function during KD implementation. To optimize ketosis, we initiated treatment with the classic 4:1 KD ratio, with a temporary protein restriction of 0.5 g/kg/day in the initial phase. This approach led to hypoproteinemia in seven patients and anemia in six. Subsequent adjustments to the KD ratio with additional protein and sucrose-Fe supplementation resolved these issues. Therefore, for critical RSE patients in ICU, the initial 4:1 KD may need to be adapted later to ensure the positive nitrogen balance and meet daily nutritional requirements. Additionally, metabolic acidosis is a significant complication in ICU patients and a notable adverse effect of KD, identified as a major barrier in similar studies. Ketogenic diets increase the acid load, potentially leading to an initial phase of uncompensated metabolic acidosis and chronic low-grade metabolic acidosis (34). Untreated, this can result in dehydration and excessive weight loss, with chronic metabolic acidosis associated with complications such as nephrolithiasis and bone mineral loss (35). Consequently, metabolic acidosis must be monitored in the ICU, and the initial 4:1 KD may require adjustment in the presence of acid–base disorders to correct the imbalance caused by the diet. Interestingly, a study suggested that potassium citrate supplementation during KD therapy can prevent compensatory metabolic acidosis (36). In our study, potassium citrate was administered, and metabolic acidosis was observed in only three patients.

In our analysis of the three-month post-treatment follow-up data from all 11 KD responders, we found that although the outcomes of the patients treated with the KD were variable, the neurofunctions, including functional independent for activities of daily life evaluated with the mRS and cognitive function evaluated with the MMSE and MoCA, demonstrated continued improvement under the KD regimen, persisting even after hospital discharge. These findings align with previous reports suggesting that KD treatment may enhance both seizure control and cognitive outcomes in patients with prolonged SRSE (37).

We conclude that the KD for SRSE may be feasible, effective, and relatively safe in the ICU. In the vast majority of critically ill SRSE patients, ketogenesis can be achieved and has a positive effect on seizure control, including resolution of SE, and weaning off of anesthetic agents and intravenous antiseizure medications (ASMs). In addition, although side effects are varied, most of them could be corrected.