Abstracts

Rationale: The ketogenic (KETO) and modified Atkins (MAD) diets have been used as effective ways to treat refractory epilepsy for years₁ with clear efficacy ₂ and safety ₃ of initiating these diets IP (inpatient) and OP (outpatient) setting. Our aim was to assess if diet initiation setting corelates with patient outcomes and laboratory values.


Methods: A retrospective analysis was performed on 157 patients who started ketogenic diet from 2007 to Jan 2022. We documented the setting of diet initiation (IP/OP), the diet type (KETO vs. MAD), bicarbonate levels, bicarbonate supplement (yes or no), diagnosis, ages of patients, ketone values, seizure reduction (as >50% seizure improvement), presence of a gastrostomy tube and ratio of ketogenic det or carbohydrate limits (MAD) at each visit. We compared differences between groups using either Fisher’s exact tests for categorical variables or Wilcoxon rank sum tests for continuous variables. Spearman correlations were assessed when both variables being compared were continuous or ordinal.


Results: Patients who initiated diet IP were younger (median age 2.9) than those in an OP setting (media age 5.3). The IP cohort had a higher rate of starting Keto (99%) compared to OP (58%). Diet ratios at visit 1(V1) varied between the IP ( >50% with 3.5 or higher) and OP groups (9% 3.5 or higher) G-tube rates were higher in the IP group (46%) than the OP group (24%). Seizure improvement ( >50% seizure reduction) at V1 occurred in 75.9% of the IP group vs. 38.9% from the OP group. When MAD patients are excluded, 75.6% of the IP group reported seizure improvement vs. 45% of the OP group at V1. At visit 3 there was no statistical difference in seizure improvement between the two groups. Ketone levels (V1) are significantly higher (p=0.003) for patients initiating IP [median IQR 4.09 (2.73, 6.15)] compared to those who started OP ( median IQR 2.46 [1.03, 4.93]). Bicarbonate levels and rates of bicarbonate supplementation were not significantly different at V1.

Conclusions: We found that patients were more likely to report seizure improvement if they initiated diets in an IP setting. This improvement dissipates by visit 3. The IP groups was more likely to be younger, on a ketogenic diet, have higher initial ketone levels, a higher ketogenic ratio and more likely to have a gastrostomy tube. Bicarbonate values were not different between the two groups

Funding: Neurology Research Institute at Childrens Hospital Colorado