Abstract
The outcome for patients attempting dietary therapy for epilepsy a second time is unknown. Twenty-six subjects treated with the ketogenic diet as children who then began either the ketogenic diet or a Modified Atkins Diet (MAD) at least 6 months later were evaluated. The mean age at the first diet trial was 5.6 years and at the second diet trial was 11.5 years. Most restarted dietary therapy because of persistent seizures (65%) or recurrence after seizure freedom (19%). Overall, 77% had a ≥50% seizure reduction with the first diet, and 50% with the second diet, P = .04. Individual subject responses were largely similar, with 14 (54%) having identical seizure reduction both times, 9 worse (35%) with the second attempt, and 3 (16%) improved. The second diet trial was more likely to lead to >50% seizure reduction if the first trial was started at a later age (7.4 vs 3.9 years, P = .04).
The ketogenic diet is a well-established therapy in continuous use since 1921 for the treatment of intractable epilepsy. Predominantly used for children, recognized indications for the ketogenic diet and Modified Atkins Diet (MAD) include many epileptic encephalopathies. 1 In these conditions, including infantile spasms, Lennox-Gastaut syndrome, Ohtahara syndrome, and severe myoclonic epilepsy of infancy (Dravet syndrome), seizures may become medically resistant at a very young age. Although some infants and children with these early-onset epileptic encephalopathies may have seizure-free periods with dietary therapies, in most patients the seizures are improved yet persist over time. Typically, due to concerns about metabolic side effects and the fact that a significant residual effect of seizure reduction occurs after the ketogenic diet is discontinued, the ketogenic diet is discontinued in children after approximately 2 years and other options (eg, further antiepileptic drug trials or vagus nerve stimulation) are then implemented. 1
There has been a remarkable increase in awareness and use of ketogenic diet therapy for adolescents and adults with epilepsy in recent years. 2 -4 The creation of the Modified Atkins Diet in 2003 has provided a less restrictive diet alternative in adult epilepsy. 5 Although most adolescents and adults considering diet therapy have not tried them previously, there are occasional patients who have had prior exposure to the ketogenic diet at a younger age. For some, the ketogenic diet was short-lived secondary to side effects, problems with restrictiveness, and/or rapid recognition of inefficacy. However, there are other patients who had positive experiences with the ketogenic diet, including periods of seizure freedom, but for various reasons the ketogenic diet was eventually discontinued. Both types of patients may reconsider the use of dietary therapy years later should seizures recur or remain uncontrolled.
The expected outcome for a reattempt of dietary therapy months to years after the ketogenic diet was previously tried is therefore an important clinical question, especially for those treating adults with dietary therapy. Would results with a second trial of dietary therapy be similar to the first attempt? If restrictive or fraught with side effects the first time, would a second trial (perhaps with the Modified Atkins Diet) be better tolerated? If patients tried the ketogenic diet during the first attempt, would the Modified Atkins Diet be equally beneficial with the second attempt? To our knowledge, these questions have not been previously studied.
Methods
All patients started prospectively on either the ketogenic diet or Modified Atkins Diet at Johns Hopkins Hospital in Baltimore, Maryland, or Johns Hopkins All Children’s Hospital in St. Petersburg, Florida, were evaluated. Charts and database information was examined for those children or adults who had been treated with the ketogenic diet or Modified Atkins Diet for a documented second time. Patients included were those who had their first dietary treatment with the ketogenic diet at an age <18 years. All subjects included had ≥6 months between the 2 diets (with a return to a regular diet). Patients were excluded if they continued restricting carbohydrates during the time period in between diets (and hence the second attempt mostly involved a stricter management by a ketogenic diet team). The second diet attempt was always at our 2 centers, but the first (ketogenic diet) may have been at another institution (and records were reviewed regarding that experience as well as patient or parent recall). No patient in this cohort tried the diet 3 or more times.
There was no minimum diet duration for either diet. As our Adult Epilepsy Diet Center at Johns Hopkins Hospital has been in existence since August 2010, there were several adults with more recent initiations of their second diet (typically the Modified Atkins Diet), and therefore shorter follow-up periods. Outcomes were based on clinic visit reports of seizure frequencies as well as phone and e-mail contact. Seizure reduction was classified as 0%, 1% to 50%, 51% to 89%, 90% to 99%, and seizure-free. These outcomes were based on a review of each patient’s overall improvement throughout their exposure to the ketogenic diet or Modified Atkins Diet, typically stabilizing at 6 to 12 months. The outcome from the second diet was compared to the baseline seizure frequency just before the second diet (not the first attempt) by reviewing clinic notes and seizure calendars. In addition, an overall classification was made based on neurologist, dietitian, parent, and/or patient reports to classify whether the second diet was “similar,” “worse,” or “better” than the first diet in regard to overall efficacy and tolerability.
All patients (or parents) prospectively consented to be included in a central database at Johns Hopkins Hospital or at Johns Hopkins All Children’s Hospital. Information was retrospectively reviewed from electronic medical records and this database. This study was approved by the Johns Hopkins Hospital Institutional Review Board. Fisher exact test was used to compare categorical variables, and a 2-tailed t test was used to compare means. An alpha level of P <.05 was used to determine significance.
Results
Demographics
Twenty-six patients fulfilled the inclusion criteria for this study. Nineteen patients received their second diet at Johns Hopkins Hospital as children, 2 at Johns Hopkins All Children’s Hospital as children, and 5 as adults at Johns Hopkins Hospital. Etiologies included Lennox-Gastaut syndrome (n = 11), idiopathic focal epilepsy (n = 3), myoclonic astatic epilepsy (n = 2), tuberous sclerosis complex (n = 2), Williams syndrome (n = 1), Ohtahara syndrome due to ARX (aristaless related homeobox) mutation (n = 1), Angelman syndrome (n = 1), juvenile myoclonic epilepsy (n = 1), alternating hemiplegia of childhood (n = 1), Group B streptococcus meningitis (n = 1), atypical absence epilepsy (n = 1), and 15q deletion (n = 1). All were having frequent, at times daily, seizures at each diet trial. Seventeen (65%) had moderate or profound intellectual disability and developmental delay.
First Diet Attempt Experience
The first diet trial with the ketogenic diet was started from January 1990 to March 2012 in these 26 subjects. Demographic and outcome data with first diet treatment on these patients is provided in Table 1. The mean age at ketogenic diet onset was 5.6 years (range: 6 months to 15 years); 13 (50%) were female. Seven (27%) had the ketogenic diet started at a different center. Antiepileptic drug burden was mean 1.1 (range: 0-3) at time of ketogenic diet onset), with 3 subjects having insufficient documented information from their prior center regarding antiepileptic drugs used. In those receiving anticonvulsants, valproate was the most commonly used (n = 13), followed by topiramate (n = 4), levetiracetam (n = 3), and lamotrigine (n = 2). Three infants had infantile spasms at the first diet attempt because of etiologies of tuberous sclerosis complex (n = 2) and Williams syndrome (n = 1).
Demographics of Patients Starting Dietary Therapies Twice (n = 26).
Abbreviation: SD, standard deviation.
aEight patients remain on the diet at this time; therefore, this number will increase as they continue the diet.
bFor those who have discontinued the diet attempt at this time.
Influence of Factors on Likelihood of ≥50% Seizure Reduction (Success) With Second Diet Attempt.
Abbreviation: SD, standard deviation.
The outcomes with the first diet were generally positive, with 20 (77%) having a >50% seizure reduction over time. Twelve (46%) had >90% response, including 5 (19%) who became seizure-free for at least 6 months. All 3 of the patients with infantile spasms were responders (90%-100% spasm reductions). The mean diet duration was 19.8 months (range: 0.5-120 months). Reasons for discontinuation included inefficacy (or gradual loss of efficacy over time) in 11 (42%), restrictiveness in 6 (23%), planned discontinuation (sometimes because of years of seizure freedom) in 6 (23%), and 3 (12%) due to adverse effects.
These side effects included kidney stones, stomach pain (despite reflux medications), and vomiting (despite reflux medications) in these 3 subjects. All 3 used the Modified Atkins Diet with their second diet attempt primarily as an attempt to prevent adverse effects. This approach was successful in the first 2 patients, but not the child with vomiting (it recurred even on the Modified Atkins Diet).
Time Period Between Diet Attempts
The time period between the discontinuation of the first diet and the initiation of the second diet attempt was a mean of 64.8 months (median: 48 months, range: 6-264 months). Nine patients had gastrostomy tubes for only the second attempt, perhaps influencing their decision to retry the diet (assured compliance). Five patients were adults when the second diet was started. Twenty-three (88%) were intellectually disabled, with minimal compliance issues reported.
Reasons for retrying a ketogenic diet were persistent seizures after years of trying other medical options in 17 (65%) of the patients. In 5 patients, a resumption of seizures after seizure freedom on the diet led to the decision to retry this therapy (after medications failed to control seizures again). Three (12%) had status epilepticus, and a decision was made to implement the ketogenic diet via nasogastric tube to control their seizures. Lastly, 1 patient was retried on the ketogenic diet with some modifications to the ratio 6 months after discontinuing the ketogenic diet for side effects, with the hope that it would be better tolerated following these adjustments. Her seizures had been reduced by 50% when the ketogenic diet was initially stopped.
Second Diet Attempt Experience
The second attempt at dietary therapy occurred from November 1996 to February 2015. The mean age at the second diet was 11.5 years (range: 0.8-28 years). Patients were receiving a mean of 2.4 (range: 0-4) antiepileptic drugs at onset of the second diet, which was more than the first diet (P < .001). Eight patients had vagus nerve stimulation insertion in the interim; none had epilepsy surgery. Sixteen (62%) were started on the classic ketogenic diet and 10 (38%) on the Modified Atkins Diet.
As a group, seizure reduction was somewhat less than compared with the first diet attempt. Thirteen (50%) had ≥50% seizure reduction over time (compared to 77%, P = .04); 5 (19%) had ≥90% seizure reduction (compared to 46%, P = .04). However, there was no difference in likelihood of seizure freedom (15% vs 19%, P = .5) with the second diet. When examined for individual differences in seizure reduction, 14 (50%) had similar seizure reduction between diets, 9 (35%) worse, and 3 (15%) improved.
Four patients were seizure-free with the second diet attempt. One child with myoclonic-astatic epilepsy had 90% to 99% seizure reduction with the ketogenic diet but discontinued due to restrictiveness. She is currently seizure-free with the Modified Atkins Diet and has had better tolerability as well. The other 3 children were also seizure-free with the first diet attempt. The first patient was 11 years old when he started the second diet (Modified Atkins Diet Modified Atkins Diet) for recurrence of seizures due to atypical absence epilepsy. He discontinued the diet again after 8 years of seizure and medication freedom, and remains seizure-free a second time. A second patient aged 27 years with presumed myoclonic-astatic epilepsy at his second trial of dietary therapy discontinued the Modified Atkins Diet after 11 months. He was seizure-free with the Modified Atkins Diet, but because his primary purpose was valproate discontinuation (not seizure freedom), and this was not successful when attempted, the Modified Atkins Diet was stopped. He has since been lost to follow-up. The third seizure-free patient started the ketogenic diet at age 18 years for alternating hemiplegia of childhood after seizures had returned following her first diet therapy discontinuation.
Six children had <50% seizure reduction with the ketogenic diet yet retried the diet a second time. Only 1 patient had ≥50% seizure reduction with the second diet trial (a 15-year-old started on the ketogenic diet to stop status epilepticus). The diet was successful, and he remains on the ketogenic diet to date with 50% seizure reduction compared to the period prior to status epilepticus.
Three patients had infantile spasms as the cause for trying the ketogenic diet the first time, with all having 90% seizure reduction or greater. The second attempt (all 3 with the ketogenic diet) was less successful. Two with tuberous sclerosis complex had no improvement and discontinued the ketogenic diet after 1 and 2 months. Another child with Williams syndrome who had 90% to 99% seizure reduction with the first attempt has had 50% to 90% reduction with the second ketogenic diet trial and remains on the diet after 4 years to date.
The ketogenic diet and Modified Atkins Diet as the second diet were equal in efficacy: 7/16 (44%) vs 6/10 (60%) with >50% seizure reduction respectively, P = .23. Of other factors also examined for likelihood of a >50% seizure reduction with the second diet, only age at first diet was significant (7.4 vs 3.9 years, P = .04). Of 5 patients who were adults at the time of the second diet, more than half (60%) were responders, with 2 (40%) seizure-free and 1 (20%) reporting 90% to 99% improvement. The ketogenic diet was successful in stopping status epilepticus in 1 of 3 patients who tried it for this purpose. Antiepileptic drugs were reduced successfully in 3 children; however, they were added in 5 over time.
Diet duration to date is 21.4 months (range: 1-100 months), which is not significantly different than the first diet attempt, P = .8. Eight patients (31%) remain on dietary therapy at this time, with those discontinuing doing so for inefficacy (16/18), restrictiveness and a gallstone (1/18), and due to a planned discontinuation after 8 years of seizure freedom (1/18). Fourteen (54%) have been on the second diet longer than the first diet (to date). Presence of a gastrostomy tube for the second diet trial did not influence this occurrence (5/14 who were on the second diet longer vs 4/12 with shorter diet durations, P = .6). Potential factors predicting a successful second diet trial are listed in Table 2.
Of the 7 children who were on the first diet for fewer than 6 months, all were on the second diet slightly longer (14.1 vs 2.4 months, P = .3). However, there was no clear improvement concurrently in seizure control, with only 1 patient having 50% to 90% seizure reduction. Six of these 7 children with brief first diet durations discontinued the second diet because of inefficacy again.
Tolerability was largely similar between diets, with 23 (88%) having a similar, 1 worse (ketogenic diet for both diets), and 2 better experience (ketogenic diet for both diets). Nine children discontinued the first diet (ketogenic diet) because of side effects and restrictiveness. Of these 9 children, 5 followed the Modified Atkins Diet the second time and 4 the ketogenic diet. Only 1 (on Modified Atkins Diet) had a clear improvement in restrictiveness documented with the second diet. Seven of these patients stopped diet therapy the second time, but none did so for restrictiveness or side effects.
Discussion
Our study suggests that a variety of patients start dietary therapy a second time, often years later. The majority did so for continued seizures after an initial positive response to the ketogenic diet and failed response to further antiepileptic drug trials and/or vagus nerve stimulation. Many of these patients had Lennox-Gastaut syndrome or other childhood-onset epilepsies that would not be expected to resolve, and two-thirds were intellectually disabled. Having a gastrostomy tube for the ketogenic diet or utilizing the Modified Atkins Diet for adults likely helped influence the decision to retry dietary therapy because of perceived improved tolerability and limited negative effect on lifestyle. 1,5,6
Although as a group, the chances of improvement were less overall with the second diet attempt, when examined individually, approximately half of patients had identical responses, including 12% who had a greater reduction in seizures during the second trial. If a child was seizure-free during the first diet attempt, the majority became seizure-free again. If a child had poor seizure response with the first diet attempt, nearly all had a poor response again. Based on these findings, a neurologist can counsel a family or patient that the first response is likely to predict the outcome with repeat treatment. Older children at the time of the first diet attempt were more likely to respond the second time (or perhaps younger children are less likely to improve). This may reflect the severe nature of some infantile-onset epilepsies (eg, Ohtahara syndrome, infantile spasms) and that their epilepsy may be just as intractable later in life and not respond well to dietary therapy.
Additionally, tolerability was similar between diet attempts, as well as between the ketogenic diet and Modified Atkins Diet. This is an encouraging finding as some patients may be unwilling to retry dietary therapy because of a perception of more difficulty as an adolescent or adult, but this was largely not the case. The Modified Atkins Diet had advantages for adults considering dietary therapy, but if a gastrostomy tube is present, the ketogenic diet was also well tolerated. The increased availability of computer programs for creating daily meal plans, parents support groups, and Internet-based recipe collections has possibly also led to an improvement in the tolerability of dietary therapies in recent years. 7 However, comparatively few of these patients were intellectually normal; therefore, tolerability and compliance in normal adolescents and adults may still be problematic. It is therefore important that patients and families understand fully the restrictiveness of the ketogenic diet and Modified Atkins Diet, even if it is a reattempt.
Future studies would be improved by prospectively surveying patients and families regarding their personal opinions about the different diet experiences, including satisfaction and quality of life scales. Although only 3 patients presented with status epilepticus at the time of the second diet initiation, this subpopulation is of high interest in the epilepsy community and may be a unique indication for retrying the ketogenic diet emergently, even if it was unsuccessful previously. 8,9 Additionally, as antiepileptic drug reduction is often a secondary goal of many families and patients, a future study could examine whether a planned medication wean in all patients could be successful. 10 Although only 2 patients in this cohort were able to reduce medications, many did not attempt to do so. Lastly, a prospective study could examine the first diet attempt to determine if any diet modifications (eg, ratio adjustment, calorie reduction, fasting) were of particular benefit. 11 These modifications could be implemented specifically to improve efficacy with the second diet attempt.
Conclusion
Using dietary therapy a second time was feasible and well tolerated in this group of children and adults, typically with continued or recurrent seizures. Outcomes for the second diet attempt were largely similar to the first attempt, both positive and negative. As more adults are being considered for dietary therapy for refractory epilepsy, these findings may help neurologists predict and counsel patients regarding the likelihood of response to treatment.
Footnotes
Author Contributions
EK, SD, MC, and ZT contributed to the study concept and design. AS entered and provided data. EK drafted the manuscript. EK, SD, MC, ZT, BH, SW, AS, and SB take responsibility for the integrity of this paper.
Declaration of Conflicting Interests
The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article. EK is on the Scientific Advisory Board for Atkins Nutritionals. The other authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
Ethical Approval
Institutional review board ethical approval was provided and consent obtained.