Clinical Nutrition xxx (2014) 1e6
Contents lists available at ScienceDirect
Clinical Nutrition journal homepage: http://www.elsevier.com/locate/clnu
Original article
Dietary glycemic load, insulin load, and weight loss in obese, insulin resistant adolescents: RESIST studyq Gesa Joslowski a, b, *, Jocelyn Halim a, Janina Goletzke b, Megan Gow a, c, Mandy Ho c, Jimmy C.-Y. Louie d, Anette E. Buyken b, Chris T. Cowell a, c, e, Sarah P. Garnett a, c, e a
Institute of Endocrinology and Diabetes, The Children’s Hospital at Westmead, Sydney, Australia IEL-Nutritional Epidemiology, University of Bonn, DONALD Study at the Research Institute of Child Nutrition, Germany The Children’s Hospital at Westmead Clinical School, University of Sydney, Sydney, Australia d Faculty of Science, Medicine and Health, The University of Wollongong, Australia e Kids Research Institute at the Children’s Hospital at Westmead, Sydney, Australia b c
a r t i c l e i n f o
s u m m a r y
Article history: Received 7 November 2013 Accepted 20 January 2014
Background & aims: The optimal dietary approach for weight loss and improving insulin sensitivity in adolescents is unknown. This study aimed to explore the association between the estimated insulin demand of the diet, as measured by glycemic and insulin load, weight loss, percentage body fat and insulin sensitivity index (ISI) in obese adolescents with clinical features of insulin resistance and/or prediabetes after a 3 month lifestyle and metformin intervention. Methods: Secondary data analysis of 91 adolescents (median age 12.7 years (range 10.1e17.4) participating in a randomized controlled trial, known as RESIST; ACTRN12608000416392. Weight change between baseline and 3 months was measured by BMI expressed as percentage of the 95th centile (BMI % 95). Body composition was measured by dual energy X-ray absorptiometry and ISI was determined by an oral glucose tolerance test. Results: Higher dietary glycemic load and insulin load were associated with less weight loss (BMI %95), adjusted for sex and pubertal stage, b ¼ 0.0466, P ¼ 0.007 and b ¼ 0.0124, P ¼ 0.040, respectively. Inclusion of total energy intake in the model explained observed associations between dietary glycemic load and insulin load and change in BMI %95. Neither dietary glycemic load nor insulin load were associated with changes in percentage body fat or ISI. Dietary glycemic index and macronutrient content (% of total energy) were not associated to changes in BMI %95, percentage body fat or ISI. Conclusion: Reduced energy diet contributes to weight loss in obese, insulin resistant adolescents. Diets with a lower insulin demand were associated with a lower energy intake and may hence assist with weight loss. Ó 2014 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved.
Keywords: Glycemic load Insulin load Weight loss Insulin resistance Adolescents
1. Introduction Adolescent obesity is a global public health concern and is associated with a range of health related problems including preand type 2 diabetes. Lifestyle interventions, including diet and
q Conference presentation: European Congress on Obesity 2013, Joslowski G, Halim J, Goletzke J, Dunkley M, Ho M, Louie JCY, Buyken AE, Baur L, Cowell CT, Garnett SP. Dietary glycemic load, insulin load, and weight loss in obese and insulin resistant adolescents: RESIST study. European Congress on Obesity 2013, Obesity Facts (Suppl. 1) 224. * Corresponding author. IEL-Nutritional Epidemiology, University of Bonn, DONALD Study at the Research Institute of Child Nutrition, Heinstueck 11, 44225 Dortmund, Germany. Tel.: þ49 231 792210 34; fax: þ49 231 71 15 81. E-mail address: [email protected] (G. Joslowski).
exercise, with metformin can lead to improvements in weight and insulin sensitivity in adolescents.1 Yet, little is known about the optimal dietary approach for weight loss in obese adolescents, including those at risk of developing type 2 diabetes. The conventional therapeutic approach focuses on restricting energy by reducing fat and increasing carbohydrates which may not be the preferred option to treat obese adolescents with insulin resistance. It is speculated that this diet might induce higher levels of postprandial glycemia and/or insulinemia and increase insulin resistance potentially leading to type 2 diabetes.2,3 Intervention studies in overweight and obese adults indicate an efficacy of low glycemic index and/or glycemic load diets on weight loss4 especially for individuals with a compensatory increased insulin secretion.5,6 But there is a paucity of data relating dietary glycemic index or glycemic load and weight loss in obese adolescents and
0261-5614/$ e see front matter Ó 2014 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved. http://dx.doi.org/10.1016/j.clnu.2014.01.015
Please cite this article in press as: Joslowski G, et al., Dietary glycemic load, insulin load, and weight loss in obese, insulin resistant adolescents: RESIST study, Clinical Nutrition (2014), http://dx.doi.org/10.1016/j.clnu.2014.01.015
2
G. Joslowski et al. / Clinical Nutrition xxx (2014) 1e6
results from intervention studies are conflicting. One 6 month intervention trial indicated a beneficial effect of dietary glycemic load on weight loss,7 however, after 3 and 24 month interventions Kirk et al. and Mirza et al., respectively, found that ad-libitum diets with reduced glycemic load were as effective as portion size controlled or low fat diets.8,9 Dietary glycemic load is considered to be an indirect estimate of insulin demand resulting from carbohydrate containing foods.10 Nevertheless, insulin secretion is also stimulated by dietary protein. Dietary protein and fat may both act synergistically with carbohydrates to increase insulin levels and reduce glycemia.11 The new concept of the food insulin index was developed to quantify postprandial insulin responses to all foods including foods with low or no carbohydrate amounts.10 The average insulin load of the diet can be calculated by summing the product of food insulin index, energy content and consumption frequency over all recorded food items, hence estimating the insulin demand of the overall diet. Novel results from the healthy free living DONALD population suggest that a habitually higher insulin demand during puberty may predispose to higher percentage body fat in adulthood.12 However, it is unknown whether glycemic or insulin load affects weight loss in obese adolescents with clinical features of insulin resistance. The aim of this study was to explore the association between the estimated insulin demand of the diet, as measured by glycemic and insulin load and weight loss, percentage body fat and insulin sensitivity index (ISI) in obese adolescents with clinical features of insulin resistance and/or prediabetes after a 3 month lifestyle and metformin intervention. 2. Materials and methods 2.1. Participants This study is secondary data analysis of a randomized control trial, known as RESIST (Australian New Zealand Clinical Trial Registration Number 12608000416392). The primary aim of RESIST was to determine the efficacy and effectiveness of two structured lifestyle interventions differing in diet composition on insulin sensitivity, in adolescents with clinical features of insulin resistance and/or prediabetes treated with metformin. Prediabetes was defined according to the American Diabetes Association, as impaired fasting glucose 5.6e6.9 mmol/L and/or impaired glucose tolerance 2 h post load 7.8e11.0 mmol/L.13 Clinical features of insulin resistance were defined as a fasting insulin (pmol/L) to glucose (mmol/L) ratio greater than 20 with one or more of the following: acanthosis nigricans, polycystic ovarian syndrome, hypertension, fasting highdensity lipoprotein cholesterol less than 1.03 mmol/L or fasting triglycerides 1.7 mmol/L or greater, as previously described.14 The design and methods of the study have been previously published14 as well as the 6 months intention-to-treat analysis.15 This secondary data analysis was conducted after the adolescents had completed 3 months of intensive dietary intervention. At baseline 111 participants aged 10e17 years were recruited and randomized to either a high carbohydrate, low fat diet (55%e60% of total energy as carbohydrate (moderate glycemic load), 30% fat, and 15% protein) or a moderate carbohydrate, increased protein diet (40%e45% of total energy as carbohydrate (moderate glycemic load), 30% fat, and 25%e30% protein). Both groups were educated and instructed to consume low-moderate glycemic index foods. Both
diets were prescriptive and two different energy levels were prescribed depending upon age: 6000e7000 kJ (10e14 year olds) or 7000e8000 kJ (15e17 year olds). All participants were commenced on metformin and received the same overall lifestyle intervention. The only difference between the two groups was the macronutrient content of the diets. This study focuses on those 91 participants who completed the initial 3 months of the trial, had at least one assessment of dietary intake and had anthropometry, body composition and insulin sensitivity measured at baseline and 3 months. There was no significant difference in baseline age, anthropometry, body composition or insulin sensitivity between those RESIST participants who included or excluded (n ¼ 20) from this study, data not shown. However, there was a higher proportion of females who were not followed up or excluded from this study compared to those participants who were included (85% vs. 54%; P ¼ 0.010). The study was approved by The Children’s Hospital at Westmead Human Research Ethics Committee (07/CHW/12), Sydney South West Area Health, Western Zone (08/LPOOL/195) and Sydney South West Area Health Service, Royal Prince Alfred Hospital (08/RPAH/455). Written informed consent from parents and assent from the young people was sought prior to their enrolment in the study.
2.2. Anthropometry Weight and height were measured according to standard procedures. Weight was measured to the nearest 100 g using electronic scales. Height was measured twice, to the nearest 0.1 cm using a wall mounted stadiometer, and the average value was used for data analysis. Body mass index (BMI, kg/m2) was calculated. Z-scores for weight, height, and BMI were calculated from age and sex specific reference values.16 BMI was expressed as a percentage of the 95th centile (BMI %95 centile).17 Overweight and obesity were defined according to the International Obesity Task Force criteria.18
2.3. Body composition Dual energy X-ray absorptiometry (DEXA; Prodigy, Lunar-GE, Madison, WI USA) equipped with propriety software version 13.6 was used to measure body composition. The manufacturer recommended scan mode was used for total body mass measurements. When possible, standard positioning techniques were used. When the participant width exceeded the maximum scan width, they were “mummy wrapped”, with arms placed in a lateral position. Scans were analyzed using manufacturer recommended techniques. Repeated measurements in children are often considered unethical, but precision of repeated measurement in adults expressed as the percent coefficient of variation has been shown to be 2.2% for percentage body fat.19 Fat free mass index (kg of fat free mass/m2) was calculated.
2.4. Insulin sensitivity Insulin sensitivity was measured by the ISI determined from an oral glucose tolerance test performed after an overnight fast. The dose of glucose was 1.75 g/kg of body weight to a maximum of 75 g. Plasma glucose and insulin was sampled every 30 min for 2 h as previously described.14 The ISI was calculated using the following formula20:
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
10 000=
insulinfasting glucosefasting ðmean 2h glucose mean 2h insulinÞ:
Please cite this article in press as: Joslowski G, et al., Dietary glycemic load, insulin load, and weight loss in obese, insulin resistant adolescents: RESIST study, Clinical Nutrition (2014), http://dx.doi.org/10.1016/j.clnu.2014.01.015
G. Joslowski et al. / Clinical Nutrition xxx (2014) 1e6
3
2.5. Nutritional assessment
2.8. Statistical analysis
Dietary intake was assessed by 24 h dietary recalls using a standardized three-pass methodology14 which has been previously used in Australian adolescents (2007 Australian National Children’s Nutrition and Physical Activity Survey).21 To assist with estimating the amounts of foods a food model booklet was used.21 Recalls were collected by trained dieticians, face-to-face in the hospital at weeks 6 and 12 and a telephone interview at week 9. Overall, 213 24 h dietary recalls were included in the analyses and on average participants provided two 24 h dietary recalls (1e3 per participant). The foods consumed were entered into FoodWorks version 6.0.2539 (Xyris Software Inc., Brisbane, QLD 4101, Australia) by research dieticians for nutritional analysis. Macronutrient and energy intake were calculated as means of the 24 h dietary recalls of each participant using the Australian Food and Nutrient Database (AusNut) compiled in 2007 by Food Standards Australia and New Zealand and amended by product and brand specific information using AusNut (AllFoods) and AusNut (Brands) compiled in 1999 by Food Standards Australia and New Zealand.
All statistical analyses were carried out using SAS (version 9.2, SAS Institute, Cary, NC, USA). A P-value
Contents lists available at ScienceDirect
Clinical Nutrition journal homepage: http://www.elsevier.com/locate/clnu
Original article
Dietary glycemic load, insulin load, and weight loss in obese, insulin resistant adolescents: RESIST studyq Gesa Joslowski a, b, *, Jocelyn Halim a, Janina Goletzke b, Megan Gow a, c, Mandy Ho c, Jimmy C.-Y. Louie d, Anette E. Buyken b, Chris T. Cowell a, c, e, Sarah P. Garnett a, c, e a
Institute of Endocrinology and Diabetes, The Children’s Hospital at Westmead, Sydney, Australia IEL-Nutritional Epidemiology, University of Bonn, DONALD Study at the Research Institute of Child Nutrition, Germany The Children’s Hospital at Westmead Clinical School, University of Sydney, Sydney, Australia d Faculty of Science, Medicine and Health, The University of Wollongong, Australia e Kids Research Institute at the Children’s Hospital at Westmead, Sydney, Australia b c
a r t i c l e i n f o
s u m m a r y
Article history: Received 7 November 2013 Accepted 20 January 2014
Background & aims: The optimal dietary approach for weight loss and improving insulin sensitivity in adolescents is unknown. This study aimed to explore the association between the estimated insulin demand of the diet, as measured by glycemic and insulin load, weight loss, percentage body fat and insulin sensitivity index (ISI) in obese adolescents with clinical features of insulin resistance and/or prediabetes after a 3 month lifestyle and metformin intervention. Methods: Secondary data analysis of 91 adolescents (median age 12.7 years (range 10.1e17.4) participating in a randomized controlled trial, known as RESIST; ACTRN12608000416392. Weight change between baseline and 3 months was measured by BMI expressed as percentage of the 95th centile (BMI % 95). Body composition was measured by dual energy X-ray absorptiometry and ISI was determined by an oral glucose tolerance test. Results: Higher dietary glycemic load and insulin load were associated with less weight loss (BMI %95), adjusted for sex and pubertal stage, b ¼ 0.0466, P ¼ 0.007 and b ¼ 0.0124, P ¼ 0.040, respectively. Inclusion of total energy intake in the model explained observed associations between dietary glycemic load and insulin load and change in BMI %95. Neither dietary glycemic load nor insulin load were associated with changes in percentage body fat or ISI. Dietary glycemic index and macronutrient content (% of total energy) were not associated to changes in BMI %95, percentage body fat or ISI. Conclusion: Reduced energy diet contributes to weight loss in obese, insulin resistant adolescents. Diets with a lower insulin demand were associated with a lower energy intake and may hence assist with weight loss. Ó 2014 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved.
Keywords: Glycemic load Insulin load Weight loss Insulin resistance Adolescents
1. Introduction Adolescent obesity is a global public health concern and is associated with a range of health related problems including preand type 2 diabetes. Lifestyle interventions, including diet and
q Conference presentation: European Congress on Obesity 2013, Joslowski G, Halim J, Goletzke J, Dunkley M, Ho M, Louie JCY, Buyken AE, Baur L, Cowell CT, Garnett SP. Dietary glycemic load, insulin load, and weight loss in obese and insulin resistant adolescents: RESIST study. European Congress on Obesity 2013, Obesity Facts (Suppl. 1) 224. * Corresponding author. IEL-Nutritional Epidemiology, University of Bonn, DONALD Study at the Research Institute of Child Nutrition, Heinstueck 11, 44225 Dortmund, Germany. Tel.: þ49 231 792210 34; fax: þ49 231 71 15 81. E-mail address: [email protected] (G. Joslowski).
exercise, with metformin can lead to improvements in weight and insulin sensitivity in adolescents.1 Yet, little is known about the optimal dietary approach for weight loss in obese adolescents, including those at risk of developing type 2 diabetes. The conventional therapeutic approach focuses on restricting energy by reducing fat and increasing carbohydrates which may not be the preferred option to treat obese adolescents with insulin resistance. It is speculated that this diet might induce higher levels of postprandial glycemia and/or insulinemia and increase insulin resistance potentially leading to type 2 diabetes.2,3 Intervention studies in overweight and obese adults indicate an efficacy of low glycemic index and/or glycemic load diets on weight loss4 especially for individuals with a compensatory increased insulin secretion.5,6 But there is a paucity of data relating dietary glycemic index or glycemic load and weight loss in obese adolescents and
0261-5614/$ e see front matter Ó 2014 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved. http://dx.doi.org/10.1016/j.clnu.2014.01.015
Please cite this article in press as: Joslowski G, et al., Dietary glycemic load, insulin load, and weight loss in obese, insulin resistant adolescents: RESIST study, Clinical Nutrition (2014), http://dx.doi.org/10.1016/j.clnu.2014.01.015
2
G. Joslowski et al. / Clinical Nutrition xxx (2014) 1e6
results from intervention studies are conflicting. One 6 month intervention trial indicated a beneficial effect of dietary glycemic load on weight loss,7 however, after 3 and 24 month interventions Kirk et al. and Mirza et al., respectively, found that ad-libitum diets with reduced glycemic load were as effective as portion size controlled or low fat diets.8,9 Dietary glycemic load is considered to be an indirect estimate of insulin demand resulting from carbohydrate containing foods.10 Nevertheless, insulin secretion is also stimulated by dietary protein. Dietary protein and fat may both act synergistically with carbohydrates to increase insulin levels and reduce glycemia.11 The new concept of the food insulin index was developed to quantify postprandial insulin responses to all foods including foods with low or no carbohydrate amounts.10 The average insulin load of the diet can be calculated by summing the product of food insulin index, energy content and consumption frequency over all recorded food items, hence estimating the insulin demand of the overall diet. Novel results from the healthy free living DONALD population suggest that a habitually higher insulin demand during puberty may predispose to higher percentage body fat in adulthood.12 However, it is unknown whether glycemic or insulin load affects weight loss in obese adolescents with clinical features of insulin resistance. The aim of this study was to explore the association between the estimated insulin demand of the diet, as measured by glycemic and insulin load and weight loss, percentage body fat and insulin sensitivity index (ISI) in obese adolescents with clinical features of insulin resistance and/or prediabetes after a 3 month lifestyle and metformin intervention. 2. Materials and methods 2.1. Participants This study is secondary data analysis of a randomized control trial, known as RESIST (Australian New Zealand Clinical Trial Registration Number 12608000416392). The primary aim of RESIST was to determine the efficacy and effectiveness of two structured lifestyle interventions differing in diet composition on insulin sensitivity, in adolescents with clinical features of insulin resistance and/or prediabetes treated with metformin. Prediabetes was defined according to the American Diabetes Association, as impaired fasting glucose 5.6e6.9 mmol/L and/or impaired glucose tolerance 2 h post load 7.8e11.0 mmol/L.13 Clinical features of insulin resistance were defined as a fasting insulin (pmol/L) to glucose (mmol/L) ratio greater than 20 with one or more of the following: acanthosis nigricans, polycystic ovarian syndrome, hypertension, fasting highdensity lipoprotein cholesterol less than 1.03 mmol/L or fasting triglycerides 1.7 mmol/L or greater, as previously described.14 The design and methods of the study have been previously published14 as well as the 6 months intention-to-treat analysis.15 This secondary data analysis was conducted after the adolescents had completed 3 months of intensive dietary intervention. At baseline 111 participants aged 10e17 years were recruited and randomized to either a high carbohydrate, low fat diet (55%e60% of total energy as carbohydrate (moderate glycemic load), 30% fat, and 15% protein) or a moderate carbohydrate, increased protein diet (40%e45% of total energy as carbohydrate (moderate glycemic load), 30% fat, and 25%e30% protein). Both groups were educated and instructed to consume low-moderate glycemic index foods. Both
diets were prescriptive and two different energy levels were prescribed depending upon age: 6000e7000 kJ (10e14 year olds) or 7000e8000 kJ (15e17 year olds). All participants were commenced on metformin and received the same overall lifestyle intervention. The only difference between the two groups was the macronutrient content of the diets. This study focuses on those 91 participants who completed the initial 3 months of the trial, had at least one assessment of dietary intake and had anthropometry, body composition and insulin sensitivity measured at baseline and 3 months. There was no significant difference in baseline age, anthropometry, body composition or insulin sensitivity between those RESIST participants who included or excluded (n ¼ 20) from this study, data not shown. However, there was a higher proportion of females who were not followed up or excluded from this study compared to those participants who were included (85% vs. 54%; P ¼ 0.010). The study was approved by The Children’s Hospital at Westmead Human Research Ethics Committee (07/CHW/12), Sydney South West Area Health, Western Zone (08/LPOOL/195) and Sydney South West Area Health Service, Royal Prince Alfred Hospital (08/RPAH/455). Written informed consent from parents and assent from the young people was sought prior to their enrolment in the study.
2.2. Anthropometry Weight and height were measured according to standard procedures. Weight was measured to the nearest 100 g using electronic scales. Height was measured twice, to the nearest 0.1 cm using a wall mounted stadiometer, and the average value was used for data analysis. Body mass index (BMI, kg/m2) was calculated. Z-scores for weight, height, and BMI were calculated from age and sex specific reference values.16 BMI was expressed as a percentage of the 95th centile (BMI %95 centile).17 Overweight and obesity were defined according to the International Obesity Task Force criteria.18
2.3. Body composition Dual energy X-ray absorptiometry (DEXA; Prodigy, Lunar-GE, Madison, WI USA) equipped with propriety software version 13.6 was used to measure body composition. The manufacturer recommended scan mode was used for total body mass measurements. When possible, standard positioning techniques were used. When the participant width exceeded the maximum scan width, they were “mummy wrapped”, with arms placed in a lateral position. Scans were analyzed using manufacturer recommended techniques. Repeated measurements in children are often considered unethical, but precision of repeated measurement in adults expressed as the percent coefficient of variation has been shown to be 2.2% for percentage body fat.19 Fat free mass index (kg of fat free mass/m2) was calculated.
2.4. Insulin sensitivity Insulin sensitivity was measured by the ISI determined from an oral glucose tolerance test performed after an overnight fast. The dose of glucose was 1.75 g/kg of body weight to a maximum of 75 g. Plasma glucose and insulin was sampled every 30 min for 2 h as previously described.14 The ISI was calculated using the following formula20:
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
10 000=
insulinfasting glucosefasting ðmean 2h glucose mean 2h insulinÞ:
Please cite this article in press as: Joslowski G, et al., Dietary glycemic load, insulin load, and weight loss in obese, insulin resistant adolescents: RESIST study, Clinical Nutrition (2014), http://dx.doi.org/10.1016/j.clnu.2014.01.015
G. Joslowski et al. / Clinical Nutrition xxx (2014) 1e6
3
2.5. Nutritional assessment
2.8. Statistical analysis
Dietary intake was assessed by 24 h dietary recalls using a standardized three-pass methodology14 which has been previously used in Australian adolescents (2007 Australian National Children’s Nutrition and Physical Activity Survey).21 To assist with estimating the amounts of foods a food model booklet was used.21 Recalls were collected by trained dieticians, face-to-face in the hospital at weeks 6 and 12 and a telephone interview at week 9. Overall, 213 24 h dietary recalls were included in the analyses and on average participants provided two 24 h dietary recalls (1e3 per participant). The foods consumed were entered into FoodWorks version 6.0.2539 (Xyris Software Inc., Brisbane, QLD 4101, Australia) by research dieticians for nutritional analysis. Macronutrient and energy intake were calculated as means of the 24 h dietary recalls of each participant using the Australian Food and Nutrient Database (AusNut) compiled in 2007 by Food Standards Australia and New Zealand and amended by product and brand specific information using AusNut (AllFoods) and AusNut (Brands) compiled in 1999 by Food Standards Australia and New Zealand.
All statistical analyses were carried out using SAS (version 9.2, SAS Institute, Cary, NC, USA). A P-value
