- Top of page
Summary: Purpose: Although the ketogenic diet has been in use for >80 years, little agreement exists as to which patients are most likely to have dramatic, sudden, and complete seizure control.
Methods: A case–control study was performed of children with intractable epilepsy started on the ketogenic diet at our institution since June 2001. Patients with a dramatic response were defined as those becoming seizure free within 2 weeks of diet onset. These children were compared with all other patients treated with the diet over the same time period in terms of patient demographics, epilepsy characteristics, and diet parameters.
Results: Eighteen early, dramatic responders over a 3-year period were identified and compared with 89 patients who were not similarly improved. The absence of complex partial seizures as the predominant seizure type (0 vs. 23%; p = 0.02) correlated with this dramatic success. The presence of infantile spasms (39% vs. 20%; p = 0.09) approached significance, but all other variables did not.
Conclusions: An early, dramatic response to the ketogenic diet is more likely in patients with predominant seizure types other than complex partial. It may also be more likely to occur in children who have infantile spasms. In all other patient demographics and diet parameters, an equal likelihood of similar success was found.
The ketogenic diet has been available for >80 years as a treatment for pharmacoresistant epilepsy. Traditionally, patients in the toddler to school-age years with Lennox–Gastaut or multiple seizure types (including atonic and tonic seizures) were started on the diet, often after years of persistent seizures. Who is the ideal diet candidate? We know glucose transporter-1 deficiency (GLUT-1) and pyruvate dehydrogenase deficiency (PDH) patients are significantly helped (1,2). However, what about infants or adolescents (3,4)? Experience would suggest that no specific seizure type is consistently benefited by the diet (5,6). In the largest series to date evaluating the ketogenic diet, it was stated that “many questions remain: which seizure types are best treated...?” (5). Do formula-fed patients do better (7)? What about patients who are less intractable (8)?
Many of our families chose to try the ketogenic diet hoping for complete seizure freedom followed by significant reduction in anticonvulsant medications (AEDs). Stories of children like Charlie Abrahams, who have been dramatically helped by the ketogenic diet, have attracted the attention of many desperate families (9). Based on the literature, we typically counsel families that their child is just as likely to have an early, dramatic, and sustained response as is any other child being initiated on the diet; perhaps this is incorrect. We hypothesized that children who have had a remarkable response to the ketogenic diet may have some common factor(s) that can be identified before diet initiation. If true, this might have implications in identifying children for the ketogenic diet in the future, those to whom the diet should be offered earlier in their therapy.
- Top of page
Patient records were retrospectively reviewed for the period from June 2001 to June 2004; 107 patients were placed on the diet over that time period, starting with a 48-h fast, followed by gradual diet introduction over a 3-day period (6). Initial ketogenic diet admission records were reviewed for patient demographics, seizure histories and presence of underlying etiology, medications, and diet parameters. All patients had either a computed tomography scan (CT) or magnetic resonance imaging (MRI) as part of their epilepsy workup, in addition to metabolic testing. A “severe” developmental delay was defined as a developmental quotient <50% from historical records from the treating neurologist. Follow-up clinic visits were examined for diet efficacy, medication changes, diet duration, and side effects. Efficacy of the diet was assessed through patient visits, surveys, and telephone contact with parents. The Johns Hopkins Committee on Clinical Investigation approved the protocol, and informed consent was obtained.
Early, dramatic responders were defined as patients who became seizure free within 2 weeks of their ketogenic diet admission and remained seizure free afterward for ≥6 months. Two children who were included had brief seizures associated with an illness once or twice after admission. Children that became seizure free after 2 weeks on the diet, or continued to have infrequent seizures, were excluded.
Categoric data were analyzed by using both Pearson's χ2 and Fisher's exact test for independence of rows and columns. Numeric data were analyzed by using the Wilcoxon rank-sum (Mann–Whitney) test. The significance level for all tests was p = 0.05.
- Top of page
Eighteen (17%) patients were identified as early, dramatic responders over the 3-year period; 56% were boys; the mean age was 3.4 years (range, 6 months–10 years) at diet onset. Half had Lennox–Gastaut syndrome or multiple seizure types as their primary seizure disorder as per caregiver and physician reports. Four children had a history of intraparenchymal hemorrhage associated with prematurity, one child had GLUT-1 deficiency, and one had tuberous sclerosis. The median diet duration was 1.1 years (range, 9 months–3 years). Seventeen (94%) patients are medication free, with the other patient taking one of a prior three AEDs after 1 year on the diet. Eleven are currently still on the diet, with all but one of the seven that discontinued the diet doing so because of seizure freedom (acidosis with an illness in one patient). Side effects were limited to lethargy, acidosis, and kidney stones in three individual patients.
These 18 responders were compared with the 89 patients who were not as successful over the 3-year period (Table 1). Twenty-three (26%) of the 89 control patients had >90% seizure reductions. The only factor that reached significance in this comparison was the presence of complex partial seizures as the primary seizure type; none of the early, dramatic responders had this as their primary seizure type versus 20 (23%) of the control group; p = 0.02. Of the 23 patients with a >90% response to the diet in the control group, three (13%) had complex partial seizures. Of the 20 patients with complex partial seizures started on the diet in the past 3 years, only seven (35%) had a ≥50% response at 3 months. EEG evidence of focal spike waves or slowing was not seen in the dramatic responder group but was seen in 15% of the control group; p = 0.07. Presence of infantile spasms neared significance, with 39% of the dramatic responders having infantile spasms as their seizure type, compared with 20% of the control group; p = 0.09. No impact of any one particular AED was noted on the likelihood of having a dramatic response.
Table 1. Comparison of early, dramatic responders with all other ketogenic diet patients, 2001–2004
| ||Early, dramatic responder (n = 18)||Control ketogenic group (n = 89)||p Value|
|Male gender||10 (56%)||50 (56%)||0.95|
|Age at first seizure (yr)||1.5||1.4||0.56|
|Age at diet onset (yr)||3.4||4.4||0.54|
|Structural brain abnormality||5 (29%)||24 (27%)||0.93|
|Severe developmental delay||7 (39%)||37 (42%)||0.83|
|Large ketosis during first 2 wk||17 (94%) ||79 (89%)||0.47|
|BMI (body mass index)||16.4||16.6||0.84|
|BMI/weight percentile for age||48.9%||48.7%||0.87|
| Lennox–Gastaut/multiple||9 (50%)||35 (39%)||0.30|
| Infantile spasms||7 (39%)||18 (20%)||0.09|
| Complex partial||0 (0)||20 (23%)||0.02|
| Other||2 (11%)||15 (17%)||0.54|
| Focal spikes or slowing||0 (0)||13 (15%)||0.07|
| High-voltage spike–wave or multifocal discharges||12 (67%)||57 (64%)||0.82|
|Medications attempted before diet||4.2||5.4||0.25|
|Average number of seizures per month before diet||1,553||2,476||0.44|
| 3:1 ratio (%)||6 (33%)||22 (25%)||0.83|
| All-liquid (formula)||6 (33%)||32 (36%)||0.42|
- Top of page
Despite the ketogenic diet's long-standing use as a treatment for epilepsy, many questions persist regarding predictability of which patients are most likely to do well. The aim of this study was to provide further insight into these questions by examining these “super” responders and comparing them with a control group, realizing that many of these control patients had success that most neurologists would consider excellent.
In this comparison of exceptionally successful cases from the past 3 years with all other patients, only one variable reached significance. Having complex partial seizures as the primary type appeared to be a negative prognostic factor for a dramatic response on the diet. None of these responders from the past 3 years had complex partial seizures as their main seizure type, compared with nearly one fourth of the control population of ketogenic diet patients. Patients with complex partial seizures also were overall less improved at 3 months than was the overall patient historically in several studies (33% vs. ∼50% with a >50% seizure reduction) (5,6). One other study also showed a negative impact of partial compared with generalized seizures on outcomes (10). We do not believe patients with complex partial seizures should necessarily be discouraged from attempting the diet; however, based on this study, it seems much less likely that they will have a complete and sustained response within days.
The average age of the early, dramatic responders at the start of the diet was a full year younger than that of the general ketogenic diet population, which, although not statistically significant, also may indicate that the ketogenic diet might be most effective if introduced at an earlier age. Nearly double the percentage of the dramatic responders had with infantile spasms versus the general ketogenic diet group. Infantile spasms were the second largest subgroup among the early, dramatic responders, after Lennox–Gastaut and multiple seizures. A previous study by our group also suggested that patients with recalcitrant infantile spasms might improve with the ketogenic diet (11).
Several drawbacks occurred with this study. Because of our strict definition of an early, dramatic response, many patients with not quite as exceptional outcomes were excluded from analysis, including those 99% improved, seizure free after >2 weeks, and even one patient seizure free initially, only to have a later recurrence of occasional seizures. Thus many patients who responded well to the diet were included in the control group. Another drawback deals with the actions of the patients and parents themselves. Patient compliance to the ketogenic diet is nearly impossible to measure; ketosis can fluctuate even in compliant children, and reports are subjective by their nature. It is possible that some of the 23 control patients with >90% seizure reduction may have actually been superresponders initially, had compliance been better, although most families tend to follow the diet instructions well for the first few weeks at least. None of these children had routine EEGs or video monitoring at 2 weeks on the diet to document seizure reduction definitively. In addition, more patients over a period >3 years could possibly indicate a trend not seen.
Can you predict which patients will have an early and dramatic success? No, although patients with primarily complex partial seizures are unlikely to have this response. Possibly the expansion of our knowledge of the human genome will help prospectively to identify children likely to respond to the ketogenic diet with a laboratory test. Initial diet parameters (ratio, calories, formula base) may be important nutritionally and lead to better ketosis, but, at least in the first 2 weeks, did not correlate with dramatic success. So although a quick, dramatic response to the diet cannot be predicted with certainty, our data suggest that even patients with structural brain abnormalities, thousands of seizures monthly, and/or severe developmental delay have a substantial chance of being miracle responders and therefore should not be excluded from starting the diet.