| | Breath acetone predicts plasma ketone bodies in children with epilepsy on a ketogenic dietReceived 18 November 2004; accepted 19 April 2005. published online 26 September 2005. Abstract ObjectiveThe high-fat ketogenic diet has long been used to treat refractory childhood seizures, but whether there is a relation between the degree of ketosis and effectiveness of seizure control remains unclear. Frequent measurements of plasma ketones are difficult in children so the goal was to determine the utility of breath acetone as a marker of systemic ketosis and seizure control in children given the ketogenic diet because of seizures refractory to medication. MethodsIn experiment I, breath acetone and plasma ketones were assessed every 2 h during an 8-h test day in seven children. In experiment II, a preliminary assessment of the possible relation between breath acetone and seizure frequency was made over 14 d in five children and one adolescent on the ketogenic diet. ResultsBreath acetone was positively and curvilinearly related to plasma acetone (r2 = 0.99, P < 0.0001), plasma acetoacetate (r2 = 0.89, P < 0.0001), and plasma β-hydroxybutyrate (r2 = 0.94, P < 0.0001). No significant relation was found between breath acetone and seizure frequency or change in seizure frequency. ConclusionsBreath acetone is indicative of systemic ketosis while on the ketogenic diet. However, owing to the wide range of seizure types and plasma acetone, more subjects will be needed to determine whether there is a clear link between breath acetone and seizure frequency or decreased seizure frequency while on the high-fat ketogenic diet.
Introduction  The high-fat ketogenic diet has been used to treat refractory childhood seizures for over 80 y but its mechanism of action remains unknown. Whether there is a threshold of ketosis for effective seizure control on this diet is unclear. Some clinical studies have demonstrated a positive relation between ketosis and seizure control [1], [2], [3], but others have found no significant association [4], [5], [6]. Therefore, the situation remains similar to that expressed by Bridge and Iob [7] more than 70 y ago: “Frequently, no improvement results in spite of severe ketosis, and at times, good results are obtained without the formation of ketone bodies … it has been impossible to establish any constant correlation.” Despite the unclear role of ketosis in seizure control, frequent monitoring of ketones (acetoacetate [AcAc], β-hydroxybutyrate [β-HBA], or acetone) is an integral aspect of the ketogenic diet regimen in the home and the clinical setting [8]. The most widely used ketone test in the home is the nitroprusside-based urinary dipstick. This non-invasive test provides a semiquantitative measurement of urinary AcAc concentration but does not reliably represent blood ketone values [3], [9], [10], [11], [12]. In contrast, plasma ketone analysis is invasive, particularly because frequent blood sampling would be required to assess the relation between ketosis and seizure control. Therefore, a less invasive measurement of systemic ketosis would be useful to determine whether seizure control on the high-fat ketogenic diet depends on achieving a certain degree of ketosis. We recently demonstrated that breath acetone is a reliable measurement of plasma ketone levels in chronically ketotic rats [13] and in mildly ketotic adult humans [14]. We also showed that children with refractory epilepsy on the high-fat ketogenic diet have breath acetone levels that are about 125 times higher than that in healthy children or children with epilepsy that is controlled by medication [15]. Breath acetone is most commonly measured by gas chromatography, which is a laboratory-based method. In collaboration with Alcohol Countermeasure Systems Co. (Mississauga, ON, Canada), we have developed and tested a prototype hand-held breath acetone analyzer (BAA) intended for rapid, home-based measurement of breath acetone. The model used in this study provided acetone values in less than 1 min and was sensitive to breath acetone in the range of 500 to 9000 nmol/L (unpublished data). The objective of experiment I therefore was to determine whether breath acetone could be used to predict plasma ketone level (AcAc, β-HBA, or acetone) in children whose refractory seizures were treated using the ketogenic diet. In experiment II, the objective was to evaluate whether there is a significant relation between breath acetone and seizure frequency or change in seizure frequency in children with refractory seizures on a ketogenic diet. Materials and methods Both experiments were approved by the ethical review committees of the University of Toronto, the Hospital for Sick Children, Bloorview MacMillan Children’s Hospital, and Hôpital Ste-Justine. All guardians and subjects were fully informed of the experimental procedures before giving written consent and assent. Experiment I: breath acetone and plasma ketones Design The characteristics of the seven subjects, their ketogenic diet formulations, and length of time on the ketogenic diet are presented in Table 1. Each subject arrived at the laboratory between 6:30 and 9:00 am. Blood and breath samples were collected at fasting and every 2 h for 8 consecutive hours (five samples in total) during a single day. Subjects consumed their home-prepared ketogenic meals and drinks and took their medications at the usual times. Blood sampling and analysis Five minutes before blood collection, an electric heating blanket was placed on the hand to stimulate vasodilatation. Using disposable lancet pens, 375 μL of blood was collected from fingertips into 75-μL heparinized capillary tubes. To separate the plasma, blood samples were centrifuged in 2.5-mL heparinized Microfuge tubes at 2500g for 10 min at 4°C. An enzymatic assay kit (Sigma, St. Louis, MO, USA) was used to analyze plasma β-HBA. The AcAc assay developed by Harano et al. [16] was used, but instead of using 50 μL of plasma for the AcAc analysis, 25 μL of plasma was diluted in 25 μL of water to avoid reaching the maximum detection limit of the assay [14]. Plasma acetone was analyzed as the dinitrophenylhydrazine derivative by high-performance liquid chromatography using protein free plasma extracts. To remove plasma proteins, 50 μL of plasma was mixed with 100 μL of acetonitrile and centrifuged at 2500g for 10 min at 4°C. To prepare sufficient reagent for the derivation of 50 samples, 2.5 mg of 2,4-dinitrophenylhydrazine were dissolved in a 2.5-mL mixture of concentrated HCl, distilled water (4:6 vol/vol), and incubated in a water bath for 20 min at 60°C. To derivatize the acetone, 50 μL of plasma supernatant was combined with 50 μL of 2,4-dinitrophenylhydrazine (Sigma). Acetone standards were prepared by using control plasma spiked with known amounts of acetone. Derivatized samples were loaded onto an autosampler and automatically injected into a C18 Symmetry column (5-μm particles, 25 cm × 4.2 mm inner diameter; Waters, Milford, MA, USA) of a high-performance liquid chromatographic system (Agilent 1100 Series, Agilent, Palo Alto, CA, USA). The ultraviolet absorbance detector was set to 365 nm. The mobile phase was 63:37 (vol/vol) acetonitrile:water. The flow rate was 1.0 mL/min. Breath sampling and analysis Breath acetone was analyzed in all subjects by gas chromatography [14], [15]. The coefficients of variation across triplicate breath and calibrator samples were 6.7 ± 0.7% and 2.2 ± 0.5%, respectively. In four subjects, breath acetone was also analyzed using the BAA prototype. The remaining three subjects were incapable of using the BAA or their breath acetone values were below its sensitivity limit (∼500 nmol/L). The coefficient of variation across triplicate BAA measures averaged 7.2 ± 0.7%. Experiment II: breath acetone and seizure frequency The physical characteristics of the six subjects, including ketogenic diet type, length of time on the ketogenic diet, and anticonvulsant medications, are listed in Table 2. Some subjects in experiment I were also in experiment II. Breath acetone was the only ketone measured in experiment II. Breath samples were obtained in triplicate, one to three times daily from each subject for 2 wk. Caregivers maintained a log of daily seizure frequency throughout the study period. Subject numbers 2, 3, and 6 were physically unable to use the BAA so they were visited at home once or twice daily, and breath acetone was collected for analysis by gas chromatography. Subject numbers 1, 4, and 5 were able to use the BAA so their breath acetone was measured by BAA only. Weekly home visits were conducted to calibrate and ensure optimal functioning of the BAA. During the home visits, breath samples were also collected for subsequent analysis by gas chromatography to determine whether any subjects were exhaling isopropanol, a metabolite of acetone that can yield a false breath acetone value on the BAA (Musa-Veloso, unpublished data). | | |  | Subject number | Sex | Age (y) | Time on KD (mo) | KD ratio | MCT oil (g/d) | Weight (kg) (%) | Height (cm) (%) | AEDs before KD | AEDs during KD | |
|---|
| 1 | M | 9 | 10 | 4:1 | 0 | 26.5 (21) | 125.0(<5) | None | None | |
| 2 | M | 19 | 22 | 4:1 | 18 | 40.5(<5) | 160.0(<5) | CLB, CBZ | CLB, CBZ | |
| 3 | M | 13 | 31 | 3:1 | 15 | 34.0(<5) | 149.0(13) | CLB | None | |
| 4 | M | 10 | 8 | 40% MCT | 93 | 33.6(50) | 138.4(37) | LTG, TPM | LTG, TPM | |
| 5 | F | 11 | 64 | 3.5:1 | 0 | 35.1(29) | 145.0(36) | VPA | VPA, CLB, PB | |
| 6 | M | 8 | 35 | 3.25:1 | 0 | 20.7(<5) | 115.0(<5) | VPA, CLB | CLB | | | | |
Throughout the 2-wk study, subject number 4, who was without seizure, had breath acetone levels that were below the sensitivity limitations of the BAA. To investigate the reason for these unexpectedly low values, this subject agreed to participate in three additional 8-h tests on separate days, in each of which triplicate breath samples were collected hourly and analyzed by gas chromatography. During these 3 d of the study, medications and ketogenic meals were consumed as usual. Statistical analyses All values are expressed as mean ± standard error of the mean. A repeated measures one-way analysis of variance was performed using SAS statistical software (SAS Institute, Cary, NC, USA) to determine whether any of the measurements were significantly different from fasting values (P < 0.05). Tukey’s test was then conducted to determine where significant differences existed. To determine the relation between breath acetone and each of the plasma ketones (experiment I), non-linear regression analyses were conducted using GraphPad Prism (GraphPad, San Diego, CA, USA). To determine the relations between plasma ketones, linear regression analyses were conducted using GraphPad Prism. To evaluate the relation between breath acetone measured by gas chromatography and that measured by BAA, linear regression analysis was conducted. The statistical method described by Bland and Altman [17] was employed to determine the agreement between breath acetone as determined by gas chromatography versus that measured by the BAA. To evaluate possible within-subject and between-subject relations between breath acetone and seizure control (experiment II), a repeated measures linear regression analysis was conducted between mean daily breath acetone and daily seizure frequency using SAS statistical software. Results Experiment I: breath acetone and plasma ketones Compared with morning fasting values (0 h), all three plasma ketones were modestly but significantly increased 8 h later in the midafternoon. Plasma AcAc increased from 1.7 ± 0.4 to 2.1 ± 0.3 mmol/L, β-HBA increased from 3.5 ± 0.7 to 4.4 ± 0.6 mmol/L, and acetone increased from 3.5 ± 1.0 to 4.5 ± 1.0 mmol/L (all P < 0.05; Fig. 1A). Mean values over the 8 h were −1.9 ± 0.2 mmol/L (AcAc), 3.6 ± 0.3 mmol/L (β-HBA), and 4.0 ± 0.8 mmol/L (acetone), respectively. Hence, plasma acetone was about 12% higher than plasma β-HBA and about twice as high as plasma AcAc. Averaged over the day, the ratio of plasma β-HBA to AcAc did not change significantly from 2.0 ± 0.1 mmol/L. Mean breath acetone remained stable throughout the day, averaging 5063 ± 933 nmol/L (Fig. 1B). Although isopropanol is occasionally found on the breath in subjects on the ketogenic diet, none was detected in these subjects. Breath acetone was positively and curvilinearly related to all three plasma ketones (P < 0.0001; Fig. 2). A linear relation was also observed between plasma acetone and plasma AcAc (r2 = 0.98, P = 0.011; Fig. 3A), between plasma acetone and plasma β-HBA (r2 = 0.99, P = 0.0073; Fig. 3B), and between plasma AcAc and plasma β-HBA (r2 = 0.92, P = 0.0006; Fig. 3C). The equations for these curves are presented in the legend to Fig. 3. Breath acetone measured by gas chromatography was positively and linearly related to measurements made using the BAA (r2 = 0.99, P = 0.0002; Fig. 4A). All BAA measurements were found to be within 10% of the values obtained by gas chromatography (Fig. 4B). During the additional 3 d of testing of subject number 4, breath acetone increased throughout the day with day-long averages of 124 ± 20, 182 ± 49, and 215 ± 43 nmol/L, respectively (Fig. 5). Experiment II: breath acetone and seizure frequency Mean breath acetone and number of seizures before and during treatment with the ketogenic diet are presented in Table 3. Breath acetone values varied widely (means of 174 to 14342 nmol/L), representing an 82-fold range over the six subjects. No significant relation between breath acetone and seizure frequency was detected. The possibility that breath acetone might be related to change in seizure frequency was also evaluated, but no statistically significant relation was found. Discussion Experiment I of this study shows that, in children on a high fat ketogenic diet, breath acetone accurately predicts plasma acetone, AcAc, and β-HBA (Fig. 2). This was previously shown in rats on the high-fat ketogenic diet [13] and in adults in short-term mild ketosis [14] but had not previously been established in children maintaining higher ketosis induced by the high-fat ketogenic diet. To our knowledge, significant positive relations between AcAc, β-HBA, and acetone in plasma have not been described before. Other studies have shown a positive linear relation between breath acetone up to 25000 nmol/L and plasma acetone up to 15 mmol/L [18], [19], [20], [21]. In the present study, plasma acetone up to 9 mmol/L was positively but curvilinearly related to breath acetone up to 18000 nmol/L. A positive, linear relation between breath acetone and blood β-HBA has been reported in five patients during 10 to 36 d of fasting [22]. At breath acetone level higher than 7500 nmol/L, the relation became curvilinear as plasma β-HBA began to plateau above 3 mmol/L. In the present study, plasma β-HBA began to plateau at 4 to 5 mmol/L. Compared with plasma β-HBA and plasma AcAc, the wider range in plasma and breath acetone suggests a potentially important physiologic role of plasma acetone. β-HBA and AcAc dissociate in blood to become anions. Consequently, these ions affect blood pH, bicarbonate concentrations, and arterial gases [18]. Although AcAc is an acid, acetone is neutral and carbonic acid is a weak buffering acid, so the formation of acetone and carbon dioxide from AcAc counteracts acidosis. Unlike β-HBA or AcAc, acetone is strongly lipophilic and hydrophilic. Thus, although the distribution of β-HBA and AcAc is limited to the water compartment [23], acetone is distributed throughout the body including fat and membranes [24]. Acetone is also freely diffusible into the brain but β-HBA and AcAc require the monocarboxylic acid transporter to cross the blood-brain barrier. Recent reports have suggested that acetone has anticonvulsant properties [25], [26]. Some studies examining the relation between ketosis and seizure control have reported a positive association [1], [2], [3], but others have found no significant association [4], [5], [6]. In the present study, no significant within-subject or between-subject associations were found between breath acetone and seizure frequency. Although we only had six subjects in experiment II, the multiple breath acetone sampling per day over 14 d led to more than 250 breath acetone measurements in this experiment. These results suggest that good seizure control was not statistically dependent on achieving a particular breath acetone level or ketone threshold in plasma. The lack of association between breath acetone and seizure control therefore seems to be due in large part to the wide cross-subject range in breath acetone (120 to 14000 nmol/L). Another possibility is that breath acetone may be more clearly related to a change in seizure frequency rather than to actual seizure frequency per se. We found no such a relation but may not have been able to adequately assess this possibility because four of the six subjects were without seizure during the 2-wk period when breath acetone was collected. High breath acetone concentration in some subjects could indicate slower acetone catabolism. Acetone can be metabolized to lactate, pyruvate, and acetyl coenzyme A via the propanediol or methyl-glyoxal pathways and subsequently recycled into amino acids, fatty acids, cholesterol, or glucose [27], [28], [29]. Carbon from acetone is incorporated into cholesterol and other lipids, glycogen, and several amino acids [19], [24], [30], [31], [32]. In one tracer study, less than 30% of a dose of [14C] acetone given to fasted humans was eliminated as acetone in urine, breath, and sweat, whereas 60% was recycled into other products such as CO2, glucose, and amino acids [24]. Up to11% of plasma glucose production was calculated to potentially be derived from acetone during prolonged fasting. Incorporation of radioactivity from acetone into glucose has been observed in rats [31], ketotic cows [33], [34], fasted humans [24], and diabetic humans [19]. Thus, subjects with unusually high breath acetone may actually be converting more into glucose, which would impair seizure control. Conversely, our subject number 4 who was without seizure despite very low fasting breath acetone may reflect more efficient excretion of acetone. Summary Breath acetone is a good, non-invasive marker of systemic ketosis but shows that ketosis varies widely in children given similar formulations of the high-fat ketogenic diet. Owing to this variability in breath acetone and markedly differing seizure frequencies in different forms of epilepsy, establishing a relation between ketosis and seizure control will require frequent measurements of both parameters over an extended period, preferably from the point at which the diet is introduced. Acknowledgments Mary Ann Ryan is thanked for technical assistance. References [1].
[1]
Huttenlocher PR
.
Ketonemia and seizures
(metabolic and anticonvulsant effects of two ketogenic diets in childhood epilepsy)
.
Pediat Res
. 1976;10:536–540
.
MEDLINE |
CrossRef
[2].
[2]
Whedon B
, Sirven JI
, O’Dwyer J
, Sperling MR
.
Therapeutic serum beta-hydroxybutyrate levels in adult ketogenic diet patients
.
Epilepsia
. 1999;40(suppl 7):221
.
[3].
[3]
Gilbert DL
, Pyzik PL
, Freeman JM
.
The ketogenic diet
(seizure control correlates better with serum β-hydroxybutyrate than with urine ketones)
.
J Child Neurol
. 2000;15:787–790
.
MEDLINE |
CrossRef
[4].
[4]
Dekaban AS
.
Plasma lipids in epileptic children treated with the high fat diet
.
Arch Neurol
. 1966;15:177–184
.
MEDLINE [5].
[5]
Schwartz RM
, Boyes S
, Aynsley-Green A
.
Metabolic effects of three ketogenic diets in the treatment of severe epilepsy
.
Dev Med Child Neurol
. 1989;31:152–160
.
MEDLINE [6].
[6]
Fraser DD
, Whiting S
, Andrew RD
, Macdonald EA
, Musa-Veloso K
, Cunnane SC
.
Elevated polyunsaturated fatty acids in blood serum obtained from children on the ketogenic diet
.
Neurology
. 2003;60:1026–1029
.
[7].
[7]
Bridge EM
, Iob LV
.
The mechanism of the ketogenic diet in epilepsy
.
Johns Hopkins Med J
. 1931;48:373–389
.
[8].
[8]
Freeman JM
, Freeman JB
, Kelly MT
.
The ketogenic diet
(a treatment for epilepsy)
. 3rd ed. New York: Demos Publications; 2000;
.
[9].
[9]
Freund G
.
The calorie deficiency hypothesis of ketogenesis tested in man
.
Metabolism
. 1965;14:985–990
.
MEDLINE |
CrossRef
[10].
[10]
Livingston S
.
Dietary management of epilepsy
.
In:
Thomas CC
editors.
Comprehensive management of epilepsy in infancy, childhood, and adolescence
. IL: Springfield; 1972;p. 378–405
.
[11].
[11]
Schwartz RM
, Eaton J
, Bower BD
, Aynsley-Green A
.
Ketogenic diets in the treatment of epilepsy
(short-term clinical effects)
.
Dev Med Child Neurol
. 1989;31:145–151
.
MEDLINE [12].
[12]
Laffel L
.
Ketone bodies
(a review of physiology, pathophysiology and application of monitoring to diabetes)
.
Diabetes Metab Res Rev
. 1999;15:412–426
.
MEDLINE |
CrossRef
[13].
[13]
Likhodii SS
, Musa K
, Cunnane SC
.
Breath acetone as a measure of systemic ketosis assessed in a rat model of the ketogenic diet
.
Clin Chem
. 2000;48:115–120
.
MEDLINE [14].
[14]
Musa-Veloso K
, Likhodii SS
, Cunnane SC
.
Breath acetone is a reliable indicator of ketosis in adult volunteers consuming ketogenic meals
.
Am J Clin Nutr
. 2002;76:65–70
.
MEDLINE [15].
[15]
Musa-Veloso K
, Rarama E
, Comeau F
, Curtis R
, Cunnane SC
.
Epilepsy and the ketogenic diet
(Assessment of ketosis in children using breath acetone)
.
Pediatr Res
. 2002;52:443–448
.
MEDLINE |
CrossRef
[16].
[16]
Harano Y
, Ohtsuki M
, Ida M
.
Direct automated assay method for serum or urine levels of ketone bodies
.
Clin Chim Acta
. 1985;151:177–183
.
MEDLINE |
CrossRef
[17].
[17]
Bland JM
, Artman DO
.
Statistical methods for assessing agreement between two methods of clinical measurement
.
Lancet
. 1986;i:307–310
.
[18].
[18]
Sulway MJ
, Malins JM
.
Acetone in diabetic ketoacidosis
.
Lancet
. 1970;2:736–740
.
MEDLINE [19].
[19]
Owen OE, Trapp VE, Skutches CL, Mozzoli MA, Hoeldtke RD, Boden G, et al.
Acetone metabolism during diabetic ketoacidosis
.
Diabetes
. 1982;31:242–248
.
MEDLINE [20].
[20]
Levey S
, Balchum OJ
, Medrano V
, Jung R
.
Studies of metabolic products in expired air
(II. Acetone)
.
J Lab Clin Med
. 1964;63:574–584
.
MEDLINE [21].
[21]
Brechner VL
, Bethune RWM
.
Determination of acetone concentration in arterial blood by vapor phase chromatography of alveolar gas
.
Diabetes
. 1965;14:663–665
.
MEDLINE [22].
[22]
Tassopoulos CN
, Barnett D
, Fraser TR
.
Breath acetone and blood sugar measurements in diabetes
.
Lancet
. 1969;i:1282–1286
.
[23].
[23]
Owen OE
, Caprio S
, Reichard GA
, Mozzoli MA
, Boden G
, Owen RS
.
Ketosis of starvation
(a revisit and new perspectives)
.
Clin Endocrinol Metab
. 1983;12:359–379
.
CrossRef
[24].
[24]
Reichard GA
, Haff AC
, Skutches CL
, Paul P
, Holroyde CP
, Owen OE
.
Plasma acetone metabolism in the fasting human
.
J Clin Invest
. 1979;63:619–626
.
MEDLINE |
CrossRef
[25].
[25]
Likhodii SS
, Burnham WM
.
Ketogenic diet
(does acetone stop seizures?)
.
Med Sci Monit
. 2002;8:HY19–HY24
.
MEDLINE [26].
[26]
Rho JM
, Anderson GD
, Donevan SD
, White HS
.
Acetoacetate, acetone, and dibenzylamine (a contaminant in L-(+)-β-hydroxybutyrate) exhibit direct anticonvulsant actions in vivo
.
Epilepsia
. 2002;43:358–361
.
MEDLINE |
CrossRef
[27].
[27]
Argiles JM
.
Has acetone a role in the conversion of fat to carbohydrate in mammals?
.
Trends Biochem Sci
. 1986;11:61–63
.
CrossRef
[28].
[28]
Kalapos MP
.
Possible physiological roles of acetone metabolism in humans
.
Med Hypotheses
. 1999;53:236–242
.
Abstract |
Full-Text PDF (183 KB)
|
CrossRef
[29].
[29]
Kalapos MP
.
On the mammalian acetone metabolism
(from chemistry to chemical implications)
.
Biochem Biophys Acta
. 2003;1621:122–139
.
[30].
[30]
Price TD
, Rittenberg D
.
The metabolism of acetone
(I. Gross aspects of catabolism and excretion)
.
J Biol Chem
. 1950;185:449–459
.
MEDLINE [31].
[31]
Sakami W
, Rudney H
.
The metabolism of acetone and acetoacetate in the mammalian organism
.
Brookhaven Symp Biol
. 1952;5:176–191
.
MEDLINE [32].
[32]
Mourkides GA
, Hobbs DC
, Koeppe RE
.
The metabolism of acetone-2-C14 by intact rats
.
J Biol Chem
. 1959;234:27–30
.
MEDLINE [33].
[33]
Luick JR
, Black AL
, Simesen MG
, Kametaka M
, Kronfeld DS
.
Acetone metabolism in normal and ketotic cows
.
J Dairy Sci
. 1967;50:544–549
.
MEDLINE [34].
[34]
Black AL
, Luick JR
, Lee SL
, Knox K
.
Glucogenic pathway for acetone metabolism in the lactating cow
.
Am J Physiol
. 1972;222:1575–1580
.
MEDLINE a Department of Nutritional Sciences, University of Toronto, Toronto, Ontario, Canada b Department of Pharmacology, University of Toronto, Toronto, Ontario, Canada c Alcohol Countermeasure Systems, Mississauga, Ontario, Canada d Department of Neurology, Hôpital Ste-Justine, Montréal, Quebec, Canada e Bloorview MacMillan Children’s Centre, Toronto, Ontario, Canada  Corresponding author. Tel.: 819-821-1170, ext. 2670; fax: 819-829-7141
The Bloorview MacMillan Children’s Hospital Foundation, Dairy Farmers of Canada, NSERC, Stanley Thomas Johnson Foundation, and the University of Toronto Awards Division are thanked for their financial support. PII: S0899-9007(05)00226-1 doi:10.1016/j.nut.2005.04.008 © 2006 Elsevier Inc. All rights reserved. | |
|
FULL TEXT00226-1/journalimage?img=PIIS0899900705002261.gr1.lrg.gif&fig=fig1&kwhquery=&issn=0899-9007&ishighres=false&allhighres=false&free=yes','journalimage');)
00226-1/journalimage?img=PIIS0899900705002261.gr2.lrg.gif&fig=fig2&kwhquery=&issn=0899-9007&ishighres=false&allhighres=false&free=yes','journalimage');)
00226-1/journalimage?img=PIIS0899900705002261.gr3.lrg.gif&fig=fig3&kwhquery=&issn=0899-9007&ishighres=false&allhighres=false&free=yes','journalimage');)
00226-1/journalimage?img=PIIS0899900705002261.gr4.lrg.gif&fig=fig4&kwhquery=&issn=0899-9007&ishighres=false&allhighres=false&free=yes','journalimage');)
00226-1/journalimage?img=PIIS0899900705002261.gr5.lrg.gif&fig=fig5&kwhquery=&issn=0899-9007&ishighres=false&allhighres=false&free=yes','journalimage');)