Oops!…I Did it Again,I Gave You Some Carbs: Prescribing Errors in Inpatient Ketogenic Diet Management

Abstract

Prescribing Errors in Hospitalized Patients on the Ketogenic Diet

Yen CY, Kronisch L, Whitley K, Carew B, Zaldana A, Diaz-Medina G, Katyayan A, Cokley JA. American Journal of Health-System Pharmacy. 2025.

Purpose: A ketogenic diet (KD) is recommended as a nonpharmacological treatment option in pediatric patients with epilepsy. Prescribing errors for these patients can result in inadvertent carbohydrate exposure, increasing the risk of loss of ketosis and breakthrough seizures. The objective of this study was to evaluate the incidence of inadvertent carbohydrate exposure in hospitalized children on the traditional KD. Methods: This was a retrospective cohort study of patients with epilepsy receiving KD therapy while admitted to the hospital. Patients 18 years of age or younger diagnosed with epilepsy and/or intractable epilepsy, receiving antiseizure medications on the traditional KD or KD total parenteral nutrition, or with GLUT-1 genetic disorder were included. The primary endpoint was the incidence of patient admissions with unintended orders for carbohydrate-containing medications during hospitalization. Results: A total of 42 patients accounting for 66 inpatient admissions were included in this study. The total incidence of admissions with an inadvertent carbohydrate-containing medication order placed was 52% for intravenous (IV) medications and 64% for oral medications. Patients averaged 2 carbohydrate-containing orders per admission for both IV and oral medications. The most commonly prescribed carbohydrate-containing medications were given at least once before being discontinued. Of the IV medications documented in this study, 6 were premix products diluted in carbohydrate-containing solutions and did not have an alternative file built to facilitate dilution in normal saline. Conclusion: Because of their restricted carbohydrate allowance and the possibility of carbohydrate-containing product excipients, patients on the KD are at increased risk for receiving inappropriate carbohydrate containing medications during hospital admissions.

A ketogenic diet therapy (KDT) is an adjunctive treatment for patients of all ages with antiseizure medication-resistant epilepsy ¹ and standard of care for managing glucose transporter type 1 deficiency syndrome (Glut1DS). ² The macronutrient composition is high in fat and low in carbohydrate (often 20 g or fewer per day) with adequate protein intake. The goal is to produce nutritional ketosis through fatty acid metabolism, which can be reduced or prevented by high carbohydrate ingestion. Individuals on KDT may be admitted to a hospital to treat medical emergencies or may be electively admitted to an epilepsy monitoring unit (EMU) for assessment and treatment. KDT may also be initiated in the hospital during an elective admission ³ or in the setting of neurologic emergencies such as new-onset infantile spasms ⁴ and refractory status epilepticus. ⁵ Maintenance of nutritional ketosis and carbohydrate restriction have both been shown to be important mechanisms of action in seizure reduction, ⁶ status epilepticus elimination, and prevention of neurologic symptoms in individuals with Glut1DS.

Yen and colleagues ⁷ conducted a 6-month retrospective chart review of 66 admissions of 42 patients (18 years and younger, median age 6.6 years) with epilepsy or Glut1DS on a “traditional” ketogenic diet (prescribed as a ratio of fat to carbohydrate and protein combined in grams and excluding modified versions) at Texas Children's Hospital in Houston to identify the frequency of inadvertent prescription of carbohydrate-containing (≥300 mg) medications. They discovered inadvertent prescription of carbohydrate-containing intravenous (IV) medications in 32 (52%) admissions and oral medications in 41 (64%) admissions, with an average of two per admission. However, doses of these medications received by patients were lower than prescribed when errors were identified by the ketogenic diet-trained multidisciplinary treatment team prior to administration.

The authors described important study limitations including the inability to determine the clinical impact of carbohydrate-containing medication doses received with regard to reduction in ketone body production or exacerbation of epileptiform activity or underlying neurologic symptoms, given the retrospective nature of the study and absence of video electroencephalography (EEG) for review in all patients. They provided several practical strategies for approaching the problem of inadvertent prescribing of carbohydrate-containing medications including adding a critical alert best-practice advisory for “ketogenic diet” in the electronic medical record (EMR) and incorporating a dedicated pharmacologist trained in KDT management into the treatment team. The authors proposed implementing a standardized physician order entry program that calculates the carbohydrate content of medications, indicating that one is under development. Regular training of staff and healthcare providers on KDT management was also stressed.

In addition to the extensive strategies, tools and protocols recommended by the authors, ⁷ lectures and educational modules could be designed for all staff and heath care providers. Similar content made available on-line could be used in hospitals without experience managing these patients. Patients and families could also be provided materials to share in case of hospital admission and emergency contact information for their treating KDT team. Like a critical alert best-practice advisory, adding the term “ketogenic diet” to a patient's problem list could trigger a notification to appear on the patient's storyboard banner in their EMR. Leveraging tools and technology may help to eliminate prescribing errors.

As a general rule, crushed tablets have a lower carbohydrate content than liquids and syrups ⁸ and use of these products is ideal while acknowledging the increased time and effort required for medication preparation and administration. The pellets in capsules can clog feeding tubes and may be unsafe. Equally important to ketogenic diet training is establishing comprehensive and accurate communication during hand-offs between healthcare providers, particularly in settings with multiple trainees and temporary staff who may be less familiar with KDT. Otherwise, medication prescriptions may be changed for convenience reasons (switching crushed tablets to liquids) without recognition of the rationale for the decision to choose this form of delivery or the potential risks.

Some IV medications are only available in preparations that contain carbohydrate-rich fluids. When faced with this limitation, it is important to weigh the benefits of appropriately treating the underlying condition versus exposing the patient to the theoretical risk of losing nutritional ketosis and experiencing worsening of symptoms. For instance, many IV preparations of antibiotics contain dextrose and the risk of not adequately treating an infection may lower seizure threshold to a greater extent than impairing ketone body production, in addition to exposing the patient to other unavoidable complications. Therefore, a multidisciplinary approach is essential in making these decisions.

While the purpose of this study was to evaluate medication errors in KDT management, another challenge lies in assuring that the diet prescription itself is ordered and administered as intended. This process involves the ordering provider, the nutrition team, food services, and other staff that may be involved in food preparation and delivery. Anecdotal accounts of incorrect diet prescriptions and delivery of incorrect food items (“regular” soda rather than sugar-free) or whole meals (wrong food tray or enteral formula given to the wrong patient) have been described, even in centers with dedicated KDT programs, multidisciplinary teams and staff and healthcare provider training.

The clinical impact of inadvertent administration of carbohydrate-containing medications or incorrect diet prescriptions to patients on KDT is unknown, particularly the degree to which these errors could sabotage nutritional ketosis and the carbohydrate threshold at which this could occur in different patient populations. Studies have shown that there is not a linear correlation between ketone body concentration and clinical response to KDT and that in some circumstances, a transient increase in carbohydrate intake may not significantly impair ketosis. ⁹ Prospective collection of data from controlled settings such as patients on KDT admitted electively to an EMU or for KDT initiation during treatment for infantile spasms or RSE may be beneficial. In many of these cases, patients receive long-term EEG monitoring which could be used to identify the temporal correlation and severity of clinical response to inadvertent carbohydrate-containing medication administration as well as the carbohydrate threshold leading to ketone reduction and adverse events.

In summary, KDT may be indicated in controlling seizures in patients with refractory epilepsy or rare conditions such as Glut1DS. However, standardized protocols and education are necessary to reduce rates of inadvertent carbohydrate administration to hospitalized patients using KDT while also taking into consideration the acuity and severity of the medical conditions being treated.