PEDIATRICS Vol. 122 No. 2 August 2008, pp. e334-e340 (doi:10.1542/peds.2007-2410)
ARTICLE |
Growth of Children on Classical and Medium-Chain Triglyceride Ketogenic Diets
a University College London-Institute of Child Health, London, England
b Great Ormond Street Hospital for Children National Health Service Trust, London, England
c National Centre for Young People with Epilepsy, Lingfield, England
d North West London Hospitals National Health Service Trust, London, England
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
OBJECTIVE. The objectives of this study were to examine growth in children on classical and medium-chain triglyceride ketogenic diets and to investigate any association between growth and calorie or protein intake.
METHODS. Weight, height, and BMI z scores were recorded for children who were initiated on 1 of 2 ketogenic diets at baseline and after 3, 6, and 12 months, if continued. Mean calorie and protein intakes during treatment were calculated for children who completed 12 months on the diet. Changes in growth were compared between the 2 diets, and the association between growth and dietary intake was examined.
RESULTS. Seventy-five children provided growth data. Weight z scores decreased significantly between baseline and 3, 6, and 12 months; height z scores showed no change at 3 months but decreased significantly by 6 and 12 months. This was more significant in the younger and ambulatory children. Subdivision according to diet type showed weight z scores to decrease significantly in the medium-chain triglyceride group only at 3 and 6 months and in both groups at 12 months. Height z scores decreased significantly in both groups by 6 and 12 months. Forty children completed 12 months of treatment; in this group, the slopes of best-fit regression lines of serial z-score measures were used to represent growth trend. There were no significant differences in mean slope between classical and medium-chain triglyceride diet groups for weight, height, or BMI. There was no significant difference in mean calorie intake during the 12 months between the 2 diets, but the medium-chain triglyceride group had significantly higher protein intake. There was no significant correlation between calorie or protein intakes and the slope of the best-fit line for weight, height, or BMI.
CONCLUSIONS. Both weight and height z scores decreased during diet treatment. By 12 months, there was no difference in outcome between classical and medium-chain triglyceride protocols despite the increased protein in the latter diet.
Key Words: ketogenic diet • epilepsy • growth • central nervous system diseases
Abbreviations: MCT—medium-chain triglyceride
After early reports of the beneficial action of the ketogenic diet in seizure control,1,2 it has been widely used as a successful treatment for intractable childhood epilepsy.3–7 There have been concerns, however, about whether this treatment could adversely affect growth, possibly as a result of the restricted calorie and limited protein content of the dietary regimen. Although Couch et al8 found that children were able to maintain linear growth and weight percentiles after 6 months on the ketogenic diet, other data have been less encouraging, with both weight9 and height10–12 compromised. This seemed more of a problem in the younger age group and after a longer time on the diet. The majority of children included in previous studies were following the classical ketogenic diet protocol: the alternative medium-chain triglyceride (MCT) diet is usually more generous in protein. Growth data for children who are on the MCT diet are very limited, and no studies have examined the relation between calorie or protein intake and growth in children who are on a ketogenic diet.
The aim of this study was to examine growth in children after classical and MCT diets, by using weight and height measures, and to investigate any association between growth and calorie or protein intake. The study formed part of a larger randomized, controlled trial on efficacy and tolerability of the ketogenic diet reported separately.13
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Children included were between 2 and 16 years of age, had tried at least 2 anticonvulsant medications, and had at least 7 seizures weekly. They had no enzyme deficiencies of organic acid metabolism or history of hyperlipidemia or renal stones. All children had been randomly assigned as part of the larger trial to receive either the classical or MCT ketogenic diet, by using a computer program that used the minimization method to ensure a close balance between the treatment groups for 3 defined age groups (2–6, 7–11, and 12–16 years).
Prestudy diet calorie intake was calculated from a 4-day food record by using a computer program (Compeat Pro 5.8 [Compeat Pro, Nutrition Systems, Banbury, UK]). Initial calorie prescription for ketogenic diets was based on an average between this prediet intake and recommendations from Johns Hopkins for energy requirements on the ketogenic diet14 but taking into account weight and height (both current and recent trends), United Kingdom requirements,15 physical activity levels, seizure activity, and medications. A ketogenic ratio of 4:1 was aimed for in the classical ketogenic diets but was modified between 3:1 and 5:1 as necessary. The MCT in the MCT ketogenic diets was usually started at 40% to 45% of energy and was increased up to 60% when necessary; carbohydrate was usually started at 15% of energy and was reduced to a lowest value of 12% when necessary. Reduction of carbohydrate to improve ketosis was done only when an increase in MCT was not possible as a result of poor tolerance. Alteration in calorie prescriptions were made as needed during the course of follow-up; these were generally applied in increases or decreases of 100-kcal (420-kJ) increments, and waiting at least 2 weeks before any additional changes were applied; 50-kcal (210-kJ) increments were used for children with a very low daily energy prescription, for example, the very young or nonambulatory. Protein intake was increased as needed to meet requirements.
Weight and height measures were performed at the diet initiation appointment (baseline) and at each subsequent follow-up clinic visit (after 3, 6, and 12 months). Hospital clinic nurses, under the supervision of the ketogenic diet study nurse, performed all growth measures at the clinic visits. The same set of scales was used for most of the children at the main hospital center. Weight was measured to the nearest 0.1 kg by using standard calibrated digital scales that weighed the child while seated (Marsden Weighing Co, Henley-on-Thames, UK). Children who were unable to sit unaided were weighed while held by a parent or caregiver; this person was then reweighed on the same scales without holding the child, and the difference between the 2 weights was calculated to give the child's weight. Height was measured, without shoes, to the nearest 1 mm for children who were able to stand by using a wall-mounted stadiometer. Children who were unable to stand independently had their length measured. Three people were usually needed for this type of measurement. The child was positioned on a flat surface with the head held lightly. The legs and back were held straight with toes pointing vertically, and length was measured with a flexible tape measure. This method could be used only when the child was able to lie flat; it was not possible to monitor linear growth accurately for a few children with severe scoliosis or spastic quadriplegia.
Weights and heights were compared with reference values by using SD scores (z scores), which express the difference between the measurement of an individual and the median value of the reference population as a proportion of the SD of the reference population. An Excel add-on computer program, lmsGrowth, was used to convert the measures into z scores16; this used British 1990 reference data.17 BMI (weight in kilograms divided by the square of height in meters) was calculated for all individuals. The expression of this index as a z score, by using the same computer program, based on BMI reference curves for children and young adults,18 allowed the use of this index to assess body shape. For children who completed 12 months of the study, mean calorie and protein intakes per kilogram body weight were calculated by using the average of the values for calorie and protein that were prescribed for the child at each of 4 time points: diet initiation, 3 months, 6 months, and 12 months.
Ethics
Ethical permission for the study was obtained from the ethics committees for each of the 3 centers involved. Parents or caregivers of all children who were enrolled in the study were asked to give written consent before the child was randomly assigned to receive either the classical or the MCT ketogenic diet.
Statistical Analysis
Student's paired t tests were used to compare z scores for weight, height, and BMI at baseline with 3 months, 6 months, and 12 months. For children who continued the diet for 12 months, linear regression was used for each child separately to determine the gradient of the line of best fit of their serial weight z scores. The resulting gradient value was used to represent the overall change in weight in that child during the 12-month period. This process was repeated for both height and BMI z scores.
The unpaired t test was used to compare mean z-score values at baseline of girls versus boys and ambulatory versus nonambulatory groups (ambulatory defined as being able to walk aided or unaided). Three age groups were defined, as for the initial randomization (2–6, 7–11, and 12–16 years); analysis of variance was used to do the same comparison between the 3 age groups. The paired t tests to examine differences in z scores for weight, height, and BMI between baseline and 3, 6, and 12 months were performed separately for girls and boys, for whether the child was ambulatory or nonambulatory, and for each of the 3 age groups. For children who completed 12 months of treatment, the unpaired t test was used to compare the mean gradient of the line of best fit of serial z scores for weight, height, and BMI between girls and boys and the ambulatory and nonambulatory groups. Analysis of variance was used to perform the same comparison between the 3 age groups.
Mean baseline z-score values were compared between the 2 diet groups by using the unpaired t test. Paired t tests were performed separately for the 2 diet groups to examine differences in z scores for weight, height, and BMI between baseline and 3, 6, and 12 months. For the children who completed 12 months of treatment, the unpaired t test was used to compare the mean gradient of the line of best fit of serial z scores for weight, height, and BMI between the 2 diet groups.
The association between growth and energy or protein intake was examined for the children who completed 12 months of treatment only. Mean energy and protein per kilogram of intake during the 12 months was compared between the 2 diets by using the unpaired t test. Pearson correlation coefficient was used to examine the correlation between the mean energy and protein intake per kilogram of a child and the gradient of the line of best fit of their serial z scores for weight, height, and BMI. SPSS 13 (SPSS Inc, Chicago, IL) was used for all statistical analysis.
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
A total of 75 children were included in this analysis: 42 boys and 33 girls. Fifty of these children were defined as ambulatory and 25 as nonambulatory. Forty were aged between 2 and 6 years, 25 between 7 and 11 years, and 10 between 12 and 16 years. Thirty-seven were following the classical ketogenic diet, and 38 were following the MCT diet. Because not all children continued on the diet beyond 3 months, the number of children who provided weight and height data at each time point decreased as the study progressed. Accurate height/length measurements were not possible for all children, and calculation of BMI was possible only when weight and height data both were present. Table 1 shows the number of children who provided data at each time point and summarizes z scores, with mean, SD, and maximum and minimum values.
|
There was a general trend of falling weight and height z score during the course of ketogenic diet treatment. For examination of the significance of any changes, paired differences in z score between baseline and 3, 6, and 12 months were computed (Table 2). There was a significant fall in weight z score between baseline and 3, 6, and 12 months. Height z score showed no change between baseline and 3 months, but by 6 months, there was a highly significant fall, which continued at 12 months. The significant fall in BMI z score between baseline and 3 months would reflect the fall in weight; by 6 and 12 months, the height had also decreased, and BMI showed no significant change.
|
The gradient of the lines of best fit of serial z scores for weight, height, and BMI was computed for all children who continued the diet for 12 months and was used to represent the change that occurred during that time in an individual child. The mean gradient for weight (n = 40) was –0.0280 (range: –0.1700 to 0.0600), for height (n = 32) was –0.0434 (range: –0.1000 to 0.0200), and for BMI (n = 32) was –0.0044 (range: –0.2200 to 0.1600), again indicating an overall downward trend in all growth z scores.
Gender
Mean baseline weight, height, and BMI z-score values showed boys as a group to be slightly heavier and shorter and to have a higher BMI than girls (Table 3), but the differences were not significant (P = .748, P = .748, and P = .861 for weight, height, and BMI z scores, respectively). Analysis of the paired differences in z scores between baseline and 3, 6, and 12 months separately for girls and boys showed the fall in weight z score between initiation and 3 months to be significant for boys (P = .010) but not for girls (P = .206); this was also the case at 12 months (P = .007 for boys and .068 for girls). Neither group had a significant change in height z score at 3 months. The fall in height z score between baseline and 6 months, although significant in both groups, seemed more marked in boys (P = .003) than in girls (P = .037). This was also the case at 12 months (P = .000 for boys and P = .002 for girls).
|
The mean gradients of the lines of best fit for serial z scores for weight, height, and BMI were plotted for each gender separately for children who completed the 12-month study period (Table 4). There was no significant difference in mean gradient between girls and boys for weight (P = .920), height (P = .732), or BMI (P = .923).
|
Ambulatory Status
Ambulatory children as a group were heavier and taller and had a higher BMI than their nonambulatory counterparts at baseline (Table 3), but the differences in mean weight, height, or BMI baseline z score were not significant (P = .126, P = .574, and P = .316, respectively). Analysis of the paired differences in z scores between baseline and 3, 6, and 12 months separately for ambulatory and nonambulatory groups showed the fall in weight z score between baseline and 3 months to be highly significant for the ambulatory group (P = .000) but not the nonambulatory group (P = .334); this was also the case at 6 months (P = .000 for ambulatory and P = .505 for nonambulatory) and at 12 months (P = .001 for ambulatory and P = .406 for nonambulatory). A similar pattern of difference was seen with height z score: neither group had a significant change at 3 months, but, by 6 months, this decreased significantly in the ambulatory group (P = .000) but not in the nonambulatory group (P = .950). By 12 months, both groups had a significant decrease in height z score, but this continued to be more marked in the ambulatory children (P = .000 for ambulatory and P = .032 for nonambulatory).
The mean gradients of the lines of best fit for serial z scores for weight, height, and BMI were plotted for the ambulatory and nonambulatory groups separately for children who completed the 12-month study period (Table 4). There was no significant difference in mean gradient between the 2 groups for weight (P = .596), height (P = .386), or BMI (P = .777).
Age
The children at baseline seemed as a group to get lighter, shorter, and thinner with increasing age, with a progressive fall in mean z score for weight, height, and BMI in the older groups (Table 3). This difference was significant for both weight and height z score (P = .018 for both) but not BMI z score (P = .233). Analysis of the paired differences in z scores between baseline and 3, 6, and 12 months separately for the 3 age groups showed the fall in weight z score between baseline and 3 months to be significant in the 2- to 6-year-old children (P = .015) but not in the older groups (P = .593 for 7- to 11-year-old group, P = .065 for 12- to 16-year-old group); this was also the case at 6 months (P = .026 for 2- to 6-year-old group, P = .506 for 7- to 11-year-old group, and P = .336 for 12- to 16-year-old group), and 12 months (P = .006 for 2- to 6-year-old group, P = .535 for 7- to 11-year-old group, and P = .074 for 12- to 16-year-old group). No age group had a significant change in height z score at 3 months, but, by 6 months, this decreased significantly in the 2- to 6-year-old group (P = .016) and the 7- to 11-year-old group (P = .024) but not in the 12- to 16-year-old group (P = .099). Changes in height z score were similar at 12 months: this decreased significantly in the 2- to 6-year-old group (P = .000) and the 7- to 11-year-old group (P = .025) but not in the 12- to 16-year-old group (P = .126).
The mean gradients of the lines of best fit for serial z scores for weight, height, and BMI were plotted for each age group separately for children who completed the 12-month study period (Table 4). There was no significant difference in mean gradient among the 3 age groups for either weight (P = .305) or BMI (P = .495), but this was significant for height (P = .044).
Type of Diet
At baseline, the children who were randomly assigned to start the MCT diet seemed as a group to be heavier and taller and to have a higher BMI than the classical diet group (Table 3); however, differences in mean baseline z-score values were not significant for weight (P = .165), height (P = .085), or BMI (P = .635). Analysis of the paired differences in z scores between baseline and 3, 6, and 12 months separately for the 2 diet groups showed the fall in weight z score between baseline and 3 months to be significant for the MCT group (P = .014), but there was no significant change in the classical group (P = .146); this was also the case at 6 months (P = .014 for MCT and P = .332 for classical). By 12 months, there was a significant fall in weight z score in both diet groups (P = .014 for MCT and P = .041 for classical). Neither diet group had a significant change in height z score at 3 months, but, by 6 months, this decreased significantly in both MCT (P = .003) and classical (P = .029) groups. By 12 months, the fall in height z score was highly significant for both diet groups (P = .000 for MCT and classical). There were no significant changes in BMI z score on either diet.
The mean gradients of the lines of best fit for serial z scores for weight, height, and BMI were plotted for each diet separately for children who completed the 12-month study period (Table 4). There was no significant difference in mean gradient between the 2 diet groups for weight (P = .611), height (P = .912), or BMI (P = .748).
Dietary Intake
There was no significant difference in mean energy intake during the 12 months between the 2 diets, but the MCT group had a significantly higher protein intake (Table 5). There were no significant correlations between the mean energy or protein intake per kilogram of a child and the gradients of the lines of best fit of their serial z scores for weight, height, or BMI (Table 6).
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
The z-score results for weight, height, and BMI for these children on 2 ketogenic diets range widely, from below –3 to above 3; they are clearly a group of varying size and shape, who span the whole centile range. Although the use of mean z scores at each time point may be limited by differing numbers included at each point, they do indicate a downward trend in both weight and height during the course of the study. The paired differences show the fall in weight z score to be significant at all time points and height highly significant by 6 and 12 months.
These results add weight to the growing body of evidence that the ketogenic diet does seem to have an adverse impact on growth. Although some of the previous study results seem conflicting, this discrepancy may be attributable to differing follow-up measurement periods. Couch et al8 found no problems with either weight or height in their group of children after 6 months on the diet, and Liu et al9 found no height change after 4 months but weight to decrease by 10%. These are short time frames within which to assess growth trends; any negative impact on longitudinal height gain will not usually be noticed until after a few months. In contrast, Williams et al11 reported height centiles to fall from baseline in 86% of their children after 24 months on the diet, and Peterson et al12 found height z score to show a significant decrease from baseline by 12 months, with the greatest fall between 6 and 12 months. No significant change was found in weight z score. Vining et al10 also reported height z score to show no problems up to 6 months, then to decline. It is clear that the more long-term the measurement period, the more accurately a conclusion can be reached about the impact of ketogenic diet treatment on growth.
Changes in BMI z score are more difficult to interpret; this is influenced by both weight and height, changes in either resulting in a change in BMI. An overall fall in an individual's weight and height would be likely to manifest as no change in a BMI z score; therefore, there seemed to be no significant changes in BMI during the study period, similar results to those of Peterson et al,12 who also reported using BMI centiles and found no change after 12 months.
The use of a gradient of line of best fit of serial measurements for children who completed 12 months of dietary treatment is recommended as the appropriate statistical method for this type of datum.19 This enables a mean gradient to be obtained for subgroups of children on the diet, which allows a statistical comparison between these groups. Initially, the fall in weight and height seemed more marked in boys, in the ambulatory group, and in the younger children. When a mean gradient was obtained for each of these subgroups, no significant differences were found between the genders or between nonambulatory and ambulatory groups. The 3 age groups had a significantly different height gradient, which decreased with increasing age, indicating more of a fall in z score; this significance was not seen for weight. The increased problems in the younger age group are likely to be attributable to their faster growth rate at this age; therefore, any negative impact would be more adverse. Vining et al10 also examined age and ambulatory status as variables and reported age <1 year as a negative predictor but found a greater decline in weight z scores when nonambulatory. This discrepancy in results may have been attributable to our ambulatory group's being heavier than their nonambulatory counterparts at baseline, thereby having a lower calorie prescription at the outset. Numbers in our study were also considerably smaller: Vining et al10 had 133 children remaining on the diet at 12 months of an initial 237. A similar dropout rate occurred in our sample, with only 40 children providing weight data and 32 height data by 12 months. This is a clear limitation of our study, because subgroups for statistical comparisons were small.
The use of a randomization process should equal out the 2 diet groups, so it was surprising to find the MCT group considerably heavier and taller with a higher mean BMI z score than the classical group at initiation, although this difference was not statistically significant. This may be responsible in part for the differing pattern of weight z score between children on the 2 diets. Weight z score seemed initially to increase in the classical diet group, the subsequent decline was not significant until 12 months on the diet; however, it showed a significant fall at all time points in the MCT diet group. This was the opposite of results reported by Liu et al,9 who found more problems with the classical diet group after 4 months. Early weight changes between the 2 diets would be expected, as a result of the differing nature of the diet initiation process. MCT fat was built up over a 1- to 2-week period, resulting in reduced initial calories, whereas the classical diet was commenced on a full-energy prescription when tolerated. Details of dietary prescriptions used by Liu et al9 are not available; it may be that they were more generous in calories on the MCT diet, with a quicker increase to full prescription. A statistical comparison of weight z score between the 2 diet groups for children who completed our study, by using the mean gradients of line of best fit to represent overall pattern of change during the 12 months, showed no significant difference between the children on the classical and MCT diets.
Changes in height z score were very similar for the 2 diets, both showing significant decline by 6 and 12 months; again, there was no significant difference in overall pattern of change during the 12 months between the classical and MCT diet groups, represented by the mean gradients for height z score of the 2 groups. There were no significant changes in BMI on either diet.
The mean energy intake during 12 months in the group of 40 children who completed the study was not significantly different between the 2 diet groups; however, as would be expected from the differing nature of the dietary prescriptions, the MCT group had a significantly higher mean protein intake per kilogram of body weight. Although low protein intake could be a factor in poor growth on a ketogenic diet, the more generous protein allowance of the MCT diet in this study clearly did not improve outcome; growth was not significantly different between children who were on the classical and MCT diets. There also seemed to be no correlation between either the energy or protein intake of a child and the gradient of their lines of best fit for weight, height, or BMI. This would throw into question the importance of a role that either of these 2 nutrients would have in determining the long-term growth outcome of a child on this diet and raises the question of what else might be causing the problems. Could it be high ketones, as suggested by the results of Peterson et al,12 or some other, as-yet-unknown process having an influence? The following question also must be asked: what would the growth of this group be like had they not been on a ketogenic diet? Frequent seizures and antiepileptic medications are known to affect growth adversely, but for many of the children in this study, numbers of both were reduced over time because of the successful diet treatment; however, this study was not designed to investigate the effect of different medications on weight; additional studies are needed to assess whether different treatment regimens affect outcome.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
The risk of the ketogenic diet's adversely affecting growth of the children who use it as a treatment must be weighed against the benefits that it has in terms of seizure reduction. It is not clear what the long-term effect would be on growth trajectories; neither is it easy to give advice on how to manage the problem by dietary changes. Increasing calories and protein or reducing ketosis level in an attempt to limit any negative impact on growth will also run the risk for reducing dietary efficacy. As discussed, the evidence on the effect of the ketogenic diet on growth rate is controversial, and although it clearly does play a part, without a much larger trial that is designed to answer this question specifically, the extent will be unknown.
|
ACKNOWLEDGMENTS |
|---|
Funding was received for this project from HSA, Smiths Charity, Scientific Hospital Supplies, and the Milk Development Council. The UCL Institute of Child Health received funding as a National Institute for Health and Research Specialist Biomedical Research Centre.
Annie Jones, MSC, Ltd, provided some editorial input in the preparation of the manuscript.
|
FOOTNOTES |
|---|
Accepted Mar 17, 2008.
Address correspondence to Elizabeth G. Neal, SRD, MSc, BSc, Institute of Child Health–Neurosciences Unit, 30 Guilford St, London WC1N 1EH, United Kingdom. E-mail: l.neal{at}ich.ucl.ac.uk
The authors have indicated they have no financial relationships relevant to this article to disclose.
| What's Known on This Subject There have been 5 previous studies on growth in children after a ketogenic diet, suggesting that both weight gain and linear growth may be compromised. Children were mainly using the classical diet protocol.
|
| What This Study Adds To our knowledge, this is the first study to compare growth in children on both classical and MCT ketogenic diets and to examine the relationship between calorie or protein intake and growth in children on this treatment.
|
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
1. Wilder RM. The effects of ketonemia on the course of epilepsy. Mayo Clin Proc. 1921;2 :307 –308
2. Peterman MG. The ketogenic diet in the treatment of epilepsy: a preliminary report.
Am J Dis Child. 1924;28
:28
–33
3. Huttenlocher PR. Ketonaemia and seizures: metabolic and anticonvulsant effects of two ketogenic diets in childhood epilepsy. Pediatr Res. 1976;10 (5):536 –540[Web of Science][Medline]
4. Kinsman SL, Vining EP, Quaskey SA, Mellits D, Freeman JM. Efficacy of the ketogenic diet for intractable seizure disorders: review of 58 cases. Epilepsia. 1992;33 (6):1132 –1136[CrossRef][Web of Science][Medline]
5. Vining EP, Freeman JM, Ballaban-Gil K, et al. A multi-center study of the efficacy of the ketogenic diet.
Arch Neurol. 1998;55
(11):1433
–1437
6. Freeman JM, Vining EP, Pillas D, Pyzik PL, Casey JC, Kelly LM. The efficacy of the ketogenic diet: 1998—a prospective evaluation of intervention in 150 children.
Pediatrics. 1998;102
(6):1358
–1363
7. Kang HC, Kim YJ, Kim DW, Kim HD. Efficacy and safety of the ketogenic diet for intractable childhood epilepsy: Korean multicentre experience. Epilepsia. 2005;46 (2):272 –279[CrossRef][Medline]
8. Couch SC, Schwarzman F, Carroll J, et al. Growth and nutritional outcomes of children treated with the ketogenic diet. J Am Diet Assoc. 1999;99 (12):1573 –1575[CrossRef][Web of Science][Medline]
9. Liu YM, Williams S, Basualdo-Hamond C, Stephens D, Curtis R. A prospective study: growth and nutritional status of children treated with the ketogenic diet. J Am Diet Assoc. 2003;103 (6):707 –712[CrossRef][Web of Science][Medline]
10. Vining EP, Pyzik P, Mc Grogan J, et al. Growth of children on the ketogenic diet. Dev Med Child Neurol. 2002;44 (12):796 –802[CrossRef][Web of Science][Medline]
11. Williams S, Basualdo-Hammond C, Curtis R, Schuller R. Growth retardation in children with epilepsy on the ketogenic diet: a retrospective chart review. J Am Diet Assoc. 2002;102 (3):405 –407[CrossRef][Web of Science][Medline]
12. Peterson SJ, Tangey CC, Pimentel-Zablah EM, Hjelmgren B, Booth G, Berry-Kravis E. Changes in growth and seizure reduction in children on the ketogenic diet as a treatment for intractable epilepsy. J Am Diet Assoc. 2005;105 (5):718 –725[CrossRef][Web of Science][Medline]
13. Neal EG, Chaffe HM, Schwartz RH, et al. The ketogenic diet for the treatment of childhood epilepsy: a randomised controlled trial. Lancet Neurol. 2008;7 (6):500 –506
14. Freeman JM, Freeman JB, Kelly MT. The Ketogenic Diet: A Treatment for Epilepsy. 3rd ed. New York, NY: Demos Medical Publishing; 2000
15. Department of Health. Report No. 41: Dietary Reference Values for Food and Energy for the United Kingdom. London, United Kingdom: HMSO; 1991
16. Cole TJ, Pan. H. lmsGrowth: An Excel Add-in to Convert Measurements to z-Scores. Centre for Epidemiology and Biostatistics, School of Public Health, Faculty of Medicine, CUHK; 1999
17. Freeman JV, Cole TJ, Chinn S, Jones PR, White EM, Preece MA. Cross sectional stature and weight reference curves for the UK, 1990.
Arch Dis Child. 1995;73
(1):17
–24
18. Cole T, Freeman JV, Preece M. Body mass index reference curves for the UK, 1990.
Arch Dis Child. 1995;73
(1):25
–29
19. Matthews JN, Altman DG, Campbell MJ, Royston P. Analysis of serial measures in medical research.
BMJ. 1990;300
(6719):230
–235
PEDIATRICS (ISSN 1098-4275). ©2008 by the American Academy of Pediatrics
CiteULike 
Connotea 
Del.icio.us 
Digg 
Facebook 
Reddit 
Technorati 
Twitter What's this?
This article has been cited by other articles:
|
|
 |
|
  E. H. Kossoff, B. A. Zupec-Kania, and J. M. Rho Ketogenic Diets: An Update for Child Neurologists J Child Neurol, August 1, 2009; 24(8): 979 - 988. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
