Second meal effects of dietary calcium and vitamin D - Nature

European Journal of Clinical Nutrition (2008) 62, 872–878

& 2008 Nature Publishing Group All rights reserved 0954-3007/08 $30.00 www.nature.com/ejcn

ORIGINAL ARTICLE

Second meal effects of dietary calcium and vitamin D MJ Soares and W Chan She Ping-Delfos Program of Nutrition, School of Public Health, Curtin University of Technology, Perth, Western Australia, Australia

Objective: To determine the second meal effects of calcium and vitamin D on postprandial glucose, insulin, non-esterified fatty acids (NEFA) and glycerol. Methods: Eight volunteers aged (mean7s.e.m.) 55.571.2 years and body mass index 29.071.6 kg/m2, completed a randomized within-subject design that compared a low calcium–low vitamin D breakfast and an isocaloric high calcium–high vitamin D breakfast (HCB). Four hours following each breakfast, a very low calcium standard lunch was ingested. Serial blood collections were made on the hour over a duration of 8 h. Postprandial responses were calculated as the percentage change (D) from the fasting value for breakfast meals, and the 4th hour breakfast value for each lunch, respectively. Non-parametric tests of significance were employed. Results: The change in glucose, insulin, serum ionized calcium (iCa2 þ ) and intact parathyroid hormone was not different between the two breakfasts, or the two lunches. However, HCB resulted in a lesser suppression of NEFA that significantly carried over to lunch (P ¼ 0.036, Wilcoxon test). A similar pattern of change in glycerol did not attain overall statistical significance. DNEFA and Dglycerol were related at lunch (Spearman’s r ¼ 0.52, P ¼ 0.04). Relative to breakfast, both lunches resulted in significantly higher glucose and insulin responses (P ¼ 0.011, Wilcoxon test). Conclusions: The data are suggestive of second meal effects of calcium and vitamin D. Our observations of higher glucose and insulin after lunch may include the involvement of second meal factors as well.

European Journal of Clinical Nutrition (2008) 62, 872–878; doi:10.1038/sj.ejcn.1602803; published online 23 May 2007 Keywords: postprandial; second meal; calcium bioavailability; NEFA; glycerol; sequential

Introduction Individuals in Westernized societies exist in a postprandial state for most of the day, due to the ingestion of several sequential meals. The amount of glucose in a meal and its bioavailability determines its glucose tolerance and that of a subsequent meal (Jenkins et al., 1982; Wolever et al., 1988; Brighenti et al., 2006). This ‘second meal phenomenon’ has been shown to occur even when the type of fat in the preceding meal has been manipulated. Accordingly, Robertson et al. (2002) demonstrated that the rank order of insulin sensitivity established at breakfast (saturated fatty acids (SFA)on6 polyunsaturated fatty acids (PUFA)on3 PUFAomonounsaturated fatty acids (MUFA)) was carried Correspondence: Dr MJ Soares, Department of Nutrition, Dietetics & Food Science, School of Public Health, Curtin University of Technology, GPO Box U 1987, Perth, Western Australia 6845, Australia. E-mail: [email protected] Contributors: MJS generated the idea, planned the study, analysed the data and wrote the paper. WCSP-D planned and prepared the test meals, executed the study, assisted with data analysis and writing of the paper. Received 16 October 2006; revised 21 February 2007; accepted 18 April 2007; published online 23 May 2007

over to lunch. Such data would indicate that quantitative, as well as qualitative, changes to the composition of the first meal of the day, could have implications for postprandial metabolism over the day. Zemel et al. (2000) were the first to suggest that intracellular calcium reciprocally influenced the activities of enzymes involved in fat synthesis and lipolysis. Based on these and other studies, Zemel and co-workers hypothesized that dietary calcium could regulate adiposity (Zemel, 2002; Zemel et al., 2004). The model predicted that increasing calcium intake would suppress parathyroid hormone (PTH) and hence decrease intracellular calcium. Either directly or via insulin, the lower intracellular calcium would attenuate lipogenesis but stimulate lipolysis. The reciprocal regulation of fat metabolism would, over time, result in a change in body composition that favoured lean tissue mass. While there has been a flurry of cellular, animal and epidemiological work testing the relationship between dietary calcium and body weight control, clinical studies are only now beginning to unravel the potential mechanistic basis of calcium’s effects on metabolism (Gunther et al., 2005; Zemel et al., 2005a, b).

Calcium intake and second meal events MJ Soares and W Chan She Ping-Delfos

873 We have recently documented that an acute increase in dietary calcium (both dairy and elemental) resulted in a significantly lesser suppression of circulating non-esterified fatty acids (NEFA) (Cummings et al., 2006). This was accompanied by greater whole body fat oxidation following calcium. In the present study, we sought to understand whether increases in calcium and vitamin D at breakfast, could modify the response to a standard lunch (SL). The study was designed to fulfil the following three aims: (i) to validate our previous findings of differences in NEFA/ glycerol following high calcium single meals, (ii) to uncover carry-over effects of a high calcium breakfast on a subsequent lunch (second meal phenomenon), and (iii) to examine the potential inter-relationships between calcium bioavailability and end points of postprandial metabolism.

Subjects and methods Subjects were recruited by advertisement in the local media. They were screened by a telephone questionnaire and inclusion criteria included: (i) history of weight stability (72 kg for preceding 12 months); (ii) not on medication affecting metabolic rate or body composition; (iii) absence of symptoms of communicable disease; (iv) no history of cardiovascular symptoms or hypertension; (v) not on hormone replacement therapy, (vi) Australian, of European descent. Eleven subjects completed the protocol. All subjects gave written informed consent to participate in the study that had been approved by the Human Research Ethics Committee of Curtin University of Technology (HR 42/ 2003). All measurements were made in the clinical rooms of the School of Public Health, on the Bentley Campus. Owing to incomplete blood sampling, the data on only eight subjects (four men and four women) formed the basis of this study, and their characteristics are shown in Table 1.

Table 1 Physical and metabolic characteristics of the subjects

Age (years) Weight (kg) Body mass index (kg/m2) Waist circumference (cm) Fat mass (kg)b Fat free mass (kg)b Fasting TAG (mmol/l) Fasting LDL (mmol/l) Fasting HDL (mmol/l) Fasting vitamin D (mmol/l) Fasting PTH (pmol/l) Fasting glucose (mmol/l) Fasting insulin (mIU/ml)

Meana

s.e.m.

Range

55.50 82.16 29.00 94.62 28.58 52.21 1.56 2.83 1.31 67.88 2.92 5.39 6.98

1.21 5.93 1.64 5.57 3.52 4.65 0.14 0.19 0.12 7.01 0.28 0.11 0.75

52–62 51.90–102.50 21.86–35.19 69.50–119.80 16.39–43.77 33.90–72.38 0.71–2.42 1.77–3.57 0.81–1.78 34.00–101.00 1.40–4.60 4.45–6.10 2.87–11.70

Abbreviations: HDL, high-density lipoprotein; LDL, low-density lipoprotein; PTH, parathyroid hormone; TAG, triacylglycerol. a Mean values of the two postprandial visits. b Fat mass and fat free mass measured using dual-energy X-ray absorptiometry (DPX, Lunar Corporation, USA).

Study design The study was designed as a single blind, within-subject, randomized comparison of the acute responses to two breakfast–lunch combinations, separated by a 2-week interval. Subjects were instructed to maintain their habitual intake and activity patterns during this period.

Test meals The nutrient composition of the test meals is given in Table 2. The macronutrient content was determined from Australian Food Composition Tables (English and Lewis, 1991) and manufacturer’s product information. All subjects completed a palatability questionnaire that enquired about amount, taste, oiliness and overall acceptability of each meal. Each answer was scored on a 10 cm visual analogue scale (VAS), anchored by the most negative to the most positive response for each question.

Anthropometry and body composition measurements Standing height was measured using a wall-mounted scale (PE 27, Mentone Educational Centre, Moorabbin, VIC, Australia) and recorded to the nearest 0.1 cm. Body weight was measured after an overnight fast on each occasion, immediately after voiding, with subjects wearing light indoor clothing and no shoes and recorded to the nearest 100 g (Tanita System 502, Tanita Corporation, Tokyo, Japan). Mid arm and waist circumferences were measured as described by Norton and Olds (2000). Body composition was measured using dual-energy X-ray absorptiometry (DPX, Lunar Corporation, Madison, WI, USA).

Measurement protocol Subjects were requested to abstain from any strenuous exercise for 36 h before the measurement. All subjects were provided with 2 l of de-ionized water for ad libitum consumption the day before the study. They were also provided a low fat, low calcium meal on each night before the trial day. Subjects arrived at the laboratory after a 12 h overnight fast, and emptied their bladder after which they were weighed. They rested for an hour, after which they provided a fasting collection of urine and blood. They were then given the breakfast meal (Table 2), which they consumed within 15 min. The meal palatability questionnaire was completed in this period. In between hourly blood collections, subjects remained sedentary and spent their time in bed either lying down, or sitting up and listening to music or reading. Some subjects elected to sit at a table close to the bed, where they did craft work, needlework or solved crosswords as in our previous studies. In all instances, subjects returned to a supine position, 30 min before blood collection (Soares et al., 2004; Cummings et al., 2006). Four hours following breakfast, subjects were given lunch, which European Journal of Clinical Nutrition


874 Table 2 Nutrient composition of test mealsa Low calcium breakfast

Energy content (kJ) Protein (g) (% of total energy) Total fat (g) (% of total energy) Carbohydrate (g) (% of total energy) Fibre (g) Calcium (mg) Vitamin D (IU)b GIc Volume (ml) Weight (g)

High calcium breakfast

Standard lunch

Mean

s.e.m.

Mean

s.e.m.

Mean

s.e.m.

1349.38 11.41 14.37 11.02 30.21 43.09 51.09 4.35 248.17 12.30 50.28 248 254

1.34 0.02

1343.14 11.49 14.55 11.20 30.86 42.34 50.44 3.85 543.23 348.6 50.35 252 262

2.22 0.02

1375.39 11.86 14.66 11.41 30.69 44.26 51.48 3.02 47.51 24.66 61.39

1.30 0.01

0.02 0.03 0.00 0.56 0.02 0.02

0.04 0.04 0.00 1.74 1.15 0.02

0.00 0.06 0.00 0.07 0.00 0.02

Abbreviation: GI, glycemic index. a Obtained from English and Lewis (1991) and manufacturer’s nutrient composition. b Calculated from Weihrauch and Tamaki (1999) and manufacturer’s nutrient composition. c Calculated from Foster-Powell et al. (2002).

they consumed within 15 min. All subjects were offered a buffet meal before they returned home. De-ionized water was allowed ad libitum over the first postprandial visit, but the time and amount consumed were noted. These amounts were kept constant for subsequent visits. Two separate postprandial urine collections, at hours 4 and 8 were also made. The volume and duration of all collections were noted, urines were kept on ice and an aliquot was frozen at 801C. Total urinary nitrogen excretion was estimated by the Kjeldahl technique. Urinary calcium excretion (uCa) was measured using the GBC Flame Atomic Absorption Spectrophotometer (AAS) (GBC-HG3000, Avanta S) and results expressed in mg/4 h.

Blood assays Venous blood samples were drawn at baseline and at the end of each hour following the breakfast or lunch meal. Blood for determination of insulin, total cholesterol (TC), low-density lipoprotein, high-density lipoprotein (HDL) cholesterol, NEFA and glycerol was left to stand for 30 min at room temperature and then centrifuged at 2500 r.p.m. for 10 min. Serum was extracted and the samples stored at 801C until later analysis. Blood glucose was measured on whole blood using Accu-Chek Active glucose strips (Roche Diagnostic, Castle Hill, NSW, Australia). Serum insulin was measured by enzyme-linked immunosorbent assay based on two monoclonal antibodies (Dako Diagnostic, Cambridgeshire, UK), according to manufacturer’s instructions. Serum TC and HDL was measured as described earlier (Cummings et al., 2006). NEFA and glycerol were determined by an enzymatic colorimetric method (WAKO NEFA C Kit, Novachem, Collingwood Vic, Australia and Randox Ltd, Crumlin, UK, respectively). In our hands, individual CVintra% were o5% European Journal of Clinical Nutrition

for glycerol and o2% for glucose, insulin, NEFA, TC, HDL and TAG.

Statistical analysis All data are presented as mean7s.e.m. Change between fasting and fed states for the breakfast meals was calculated as the percentage change from fasting value over 4 h. As the intervals of measurement in the postprandial period were equal, this summary statistic was analogous to determining the incremental area under the curve (Mathews et al., 1990). All metabolic end points following breakfast had returned to baseline values by 4th hours, except serum ionized calcium (iCa2 þ ) and intact parathyroid hormone (iPTH). The postprandial change following lunch hence utilized the breakfast 4th hour value as baseline. Non-parametric tests of significance were employed at the 5% level. Friedman test for related samples was used to determine the overall statistical significance of meal effects. Post hoc comparisons used the Wilcoxon signed-rank test, and were limited to that between breakfasts, between lunches and between breakfast and lunch. Spearman’s rank correlation coefficients between changes in insulin, NEFA, glycerol, iCa2 þ and iPTH were separately calculated for breakfast and lunch. Data were analysed using the SPSS for Windows (version 14, SPSS Inc., USA) statistical software package.

Results Subjects were lean to obese and had glucose, insulin, lipid and parathyroid hormone concentrations within the normal fasting range (Table 1). There were no differences in body weight (82.0975.91 vs 82.0475.98 kg, P ¼ 0.75), fasting


875 Table 3 The influence of dairy calcium and vitamin D at breakfast on postprandial metabolism following a standard lunch Meal-induced change

Low calcium trial LCB

D D D D D D D

Glucose (%) Insulin (%) NEFA (%) Glycerol (%) Ionized Ca2 þ (%) Intact PTH (%) Urinary calcium excretion (%)

Friedman testa

High calcium trial SL

HCB

SL

(P-values)

Mean

s.e.m.

Mean

s.e.m.

Mean

s.e.m.

Mean

s.e.m.

2.82 115.5 28.3 16.2 1.75 12.8 43.66

1.37 22.16 5.03 10.88 0.19 7.36 29.94

5.39** 228.9** 29.9 13.7 1.07** 5.06 55.11

4.12 51.8 6.12 13.33 0.57 8.5 35.02

0.82 90.7 19.9 9.9 1.96 15.23 33.05

1.59 21.48 5.27 14.95 0.93 6.69 37.22

7.07** 274.6** 16.5* 46.0 1.62 6.11 59.66

2.2 56.2 6.54 34.4 0.51 8.63 47.97

0.012 0.011 0.047 0.165 0.009 0.583 0.825

Abbreviations: HCB, high calcium and vitamin D breakfast; LCB, low calcium and vitamin D breakfast; NEFA, non-esterified fatty acids; PTH, parathyroid hormone; SL, standard lunch. n ¼ 8. Post hoc tests employed Wilcoxon signed-rank test: *Po0.05 vs SL following LCB, **Po0.05 vs corresponding breakfast. a Friedman test used to determine overall statistical significance.

Second meal effects of calcium and vitamin D Serum ionized calcium was higher following high calcium– high vitamin D breakfast (HCB) (Table 3) and suppression of iPTH was correspondingly greater (Table 3), but no statistical difference was detected. Glycaemia and insulinaemia was similar between breakfasts, and similar between lunches. There was a lesser suppression of NEFA following HCB that reached significance (P ¼ 0.036) at lunch (Figure 3, Table 3). A similar rank order of effects was observed with glycerol, with HCB being higher than low calcium–low vitamin D breakfast (LCB) (Wilcoxon test, P ¼ 0.025). However, as overall statistical significance was not reached, these data are not shown in Table 3 or Figure 4.

Effect of sequential meals: breakfast vs lunch Fasting glucose (5.4170.18 vs 5.3870.13 mmol/l, P ¼ 0.88), insulin (6.9071.15 vs 7.0571.05 mIU/ml, P ¼ 0.29) and NEFA (0.5770.06 vs 0.4870.06 mEq/l, P ¼ 0.16) were similar between visits. Fasting glycerol (0.08470.01 vs 0.06670.01 mmol/l, P ¼ 0.06) tended to be higher on the LCB day, despite randomization. To account for any difference in baseline values, change in these variables was therefore expressed as a percentage of basal values (Table 3). Both glucose (P ¼ 0.012) and insulin (P ¼ 0.011) were significantly higher after lunch (Table 3, Figures 1 and 2). These changes were accompanied by significantly lower DiCa2 þ (P ¼ 0.009, Table 3). uCa was not different (Table 3). There

7.50 Blood glucose concentrations (mmol/L)

iCa2 þ (1.2070.01 vs 1.2170.01 mmol/l, P ¼ 0.83) and fasting serum iPTH (2.7070.37 vs 3.1470.43 pmol/l, P ¼ 0.29) between visits. The nutrient composition of the test meals is shown in Table 2. There were no differences in VAS scores for amount or oiliness between breakfast and lunch. Lunch, however, scored higher on taste (5.870.71 vs 7.370.48, Wilcoxon test, P ¼ 0.013), and overall acceptability (6.670.69 vs 8.070.38, Wilcoxon test, P ¼ 0.075).

*

7.00 6.50 6.00 5.50 5.00 4.50 4.00 0

1

2

3 4 5 Time (hour)

6

7

8

Figure 1 Blood glucose concentrations of eight subjects after a low dairy calcium and vitamin D breakfast (LCB J) and high dairy calcium and vitamin D breakfast (HCB K), each followed by an SL. Values are means with their standard error of mean shown by vertical bars. *P ¼ 0.012, lunch vs respective breakfast (Wilcoxon test).

was a significant correlation between DNEFA and Dglycerol across both lunches (Spearman’s r ¼ 0.52, P ¼ 0.04).

Discussion The study of sequential meal ingestion in humans, usually explores two facets of postprandial metabolism. One is the effect of manipulating dietary composition of a first meal on metabolic responses to the next. This is conventionally termed ‘second meal effects’. Accordingly, investigators have examined the effects of variations in glycemic index (GI), amount of fat and type of fatty acid in the first meal (usually breakfast), on the metabolic response to a standard second meal (usually lunch) (Jenkins et al., 1982; Ercan et al., 1994; Frape et al., 1997a, b; Robertson et al., 2002; Brighenti et al., European Journal of Clinical Nutrition


876 Breakfast 0.00 60.0

Serum NEFA concentrations (mEq /L)

Serum Insulin concentrations (µIU/ml)

70.0

*

50.0 40.0 30.0 20.0 10.0

Lunch

-0.05

-0.10

-0.15

*

-0.20

-0.25

0.0 0

1

2

3

4 5 Time (hour)

6

7

8

Figure 2 Serum insulin concentrations of eight subjects after a low dairy calcium and vitamin D breakfast (LCB J) and high dairy calcium and vitamin D breakfast (HCB K), each followed by an SL. Values are means with their standard error of mean shown by vertical bars. *P ¼ 0.011, lunch vs respective breakfast (Wilcoxon test).

2006; Clark et al., 2006) The other is the change in insulin sensitivity that may occur following a second meal, when two meals are ingested in quick succession. While some authors discuss this aspect (Frape et al., 1997a, b; Robertson et al., 2002), others do not (Jenkins et al., 1982; Brighenti et al., 2006; Clark et al., 2006). In this paper, we use the term ‘sequential effects’ to discriminate this from the former. The present study was primarily designed to explore second meal effects. The test breakfasts were equivalent in caloric intake, macronutrient composition, weight and volume through the use of the same food ingredients (Table 2). Calcium and vitamin D were increased through the use of specially formulated UHT (ultra-heat treatment) milks. As part of a mixed meal, these milks did not influence gastric emptying as judged by the paracetamol absorption test (Cummings et al., 2004). Hence differences in GI or gastric emptying of the test breakfasts provided are not expected to significantly influence events in the immediate and subsequent postprandial period. In addition, the potential metabolic effects of the overnight meal on breakfast the subsequent day (Robertson et al., 2002) were controlled through the provision of the same standard dinner (low fat, o50 mg calcium) to each subject before each trial.

Second meal effects of calcium and vitamin D Zemel et al. (2004) have opined that low calcium intake, via the calcitrophic hormones, would increase insulin secretion and/or action and thus lead to fuel storage. The provision of dietary calcium is hence expected to act in reverse. We assessed the relative bioavailability of calcium by monitoring the change in iCa2 þ , iPTH and through timed uCa. We did not detect any significant difference between breakfasts in European Journal of Clinical Nutrition

Figure 3 Change in NEFA concentrations of eight subjects after a low dairy calcium and vitamin D breakfast (LCB &) and a high dairy ), each followed by an SL. calcium and vitamin D breakfast (HCB Bars are means with their standard error of mean shown by vertical lines. *P ¼ 0.036 vs lunch following LCB (Wilcoxon test).

any of the measures used (Table 3). Following each breakfast, however, there was a significant rise in iCa2 þ with a reciprocal suppression of iPTH and an increased uCa excretion. At 4 h following each breakfast, iCa2 þ was still elevated, and iPTH was still significantly suppressed; suggesting that absorption of calcium from breakfast was not complete. This is supported by the observations that the fractional absorption of calcium at 3 and 7 h after a load is about 80 and 95%, respectively (Barger-Lux et al., 1989). Second meal effects, in part, may be due to spillover of calcium from breakfast to lunch. We observed similar glucose and insulin responses between the two test breakfasts and between the two identical lunches (Figures 1 and 2). While NEFA was significantly suppressed following both breakfasts, the magnitude of the suppression was lesser following HCB and was significantly carried over to lunch (Table 3, Figure 3). The pattern of change in glycerol mimicked that of NEFA (Spearman’s r ¼ 0.52, P ¼ 0.04) but did not achieve statistical significance because of the small sample (Figures 3 and 4). Overall, they confirm our earlier observations that a high calcium breakfast meal produced a lesser suppression of NEFA without affecting insulinaemia (Cummings et al., 2006). In addition, they provide newer evidence to suggest that dietary calcium and vitamin D may also contribute to second meal effects (Figures 3 and 4). Coppack et al. (1994) and Frayn (1998) have argued that following ingestion of a mixed meal, chylomicron triglyceride is preferentially acted upon by adipose tissue lipoprotein lipase to release NEFA. While re-esterification into triacylglycerol does occur, much of this NEFA fails to be ‘trapped’ within adipose tissue and finds itself back in circulation. The data presented here are consistent with this scheme. Whether dietary calcium may bring into play subtle changes to the antilipolytic effect of insulin remains to be determined. Alternatively, calcium may influence the sympa-


877

Serum Glycerol concentrations (mmol /L)

0.025 0.020 0.015 0.010 0.005 0.000 -0.005 -0.010 -0.015 Lunch

Breakfast -0.020

Time (hour)

Figure 4 Change in glycerol concentrations of eight subjects after a low dairy calcium and vitamin D breakfast (LCB &) and a high dairy ), each followed by an SL. calcium and vitamin D breakfast (HCB Bars are means with their standard error of mean shown by vertical lines.

thetic nervous system (SNS) in mediating postprandial events. SNS activity stimulates adipose tissue lipolysis, and contributes to meal-induced thermogenesis (Landsberg, 2006). Ingestion of calcium is expected to lower intracellular calcium and a reduction in intracellular calcium concentrations has been shown to stimulate lipolysis in human adipocytes (Xue et al., 1999). We acknowledge that systemic NEFA and glycerol concentrations are not the best quantitative indices of adipose tissue lipolysis (Coppack et al., 1994). However, higher postprandial NEFA (Cummings et al., 2004) and higher glycerol following calcium (Zemel et al., 2005a, b) are emerging as consistent findings in humans. Clearly, the role of dietary calcium and vitamin D in adipose tissue lipolysis is a potential area of interest.

Insulin sensitivity following sequential meals The response to both lunches differed markedly from breakfast, with significant increases in both glucose and insulin (Figures 1 and 2, Table 3). Such an outcome is suggestive of a decrease in insulin sensitivity following lunch, and has been indicated by other authors (Frape et al., 1997a, b; Robertson et al., 2002). To ascribe this phenomenon to sequential meal effects per se, would require close standardization of test meals. While the energy and macronutrient content of breakfast and lunch in this study were the same (Table 2), the use of different food ingredients led to some variations between the two. Lunch had a slightly higher estimated GI than breakfast, although both were in the moderate range (Table 2). This could have contributed to differences in glycaemia at that meal. The calcium content of lunch was very low (B50 mg), and this was reflected in a significantly lower DiCa2 þ and higher DiPTH relative to

breakfast (Table 3). Although we did not observe an inverse relation between measures of bioavailability and glucose/ insulin, the time course and direction of change follow the predicted effect of calcium ingestion on metabolism (Zemel et al., 2004). Lastly, our volunteers rated the taste of lunch as significantly better than breakfast. Improved meal palatability has been shown to increase insulin concentrations and affect postprandial metabolism (Sawaya et al., 2001). In other studies that have demonstrated raised glucose/ insulin following sequential meals, the second meal differed markedly in composition from the first (Frape et al., 1997a, b; Robertson et al., 2002). Hence, second meal factors may have also contributed to their observations. Burdge et al. (2003) circumvented this issue by providing the same meal at breakfast and lunch on two occasions. However, they too observed a greater insulin response, post lunch. A potential explanation for such findings also comes from the study of Burdge et al. (2003). A standard meal served at lunchtime, but in the absence of breakfast, resulted in the same higher glucose and insulin profile as when lunch was preceded by breakfast. This would argue in favour of a ‘time of day’ effect (Burdge et al., 2003), rather than a sequential meal phenomenon. There is evidence to suggest that insulin secretion rates and insulin concentrations have a circadian rhythm, being lowest early morning and gradually rising to a peak between noon and 1800 hours (Boden et al., 1996). This diurnal change was evident at all levels of circulating glucose, although the amplitude was greater at higher glucose concentrations (Boden et al., 1996). Other investigators have also reported circadian variations in insulin sensitivity, glucose tolerance and postprandial metabolism (Lee et al., 1992; Morgan et al., 2003). Given the variety of study designs and the lack of meal standardization, the basis and magnitude of sequential effects deserves further investigation. Deliberate manipulation of second meal composition, and lengthening the interval between meals, are areas that could be addressed.

Conclusion Increasing calcium and vitamin D intake resulted in a reduced postprandial suppression of NEFA. The rank order of responses in NEFA and glycerol were strongly indicative of second meal effects, but would require validation. We observed a higher glucose and insulin response following a second meal. Variations in GI, calcium content and palatability are partly responsible for such effects.

Acknowledgements We acknowledge the skilled assistance of Dr AP James, and thank Murray Goulburn Co-operative Company Ltd for provision of the specialized milk. The study was funded by Dairy Australia, and conducted under the aegis of the ATN Centre for Metabolic Fitness. European Journal of Clinical Nutrition


878

References Barger-Lux MJ, Heaney RP, Recker RR (1989). Time course of calcium absorption in humans: evidence for a colonic component. Calcif Tissue Int 44, 308–311. Boden G, Ruiz J, Urbain JL, Chen X (1996). Evidence for a circadian rhythm of insulin secretion. Am J Physiol Endo Metab 271, E246–E252. Brighenti F, Benini L, Del Rio D, Casiraghi C, Pellegrini N, Scazzina F et al. (2006). Colonic fermentation of indigestible carbohydrates contributes to the second-meal effect. Am J Clin Nutr 83, 817–822. Burdge GC, Jones AE, Frye SM, Goodson L, Wootton SA (2003). Effect of meal sequence on postprandial lipid, glucose and insulin responses in young men. Eur J Clin Nutr 57, 1536–1544. Clark CA, Gardiner J, McBurney MI, Anderson S, Weatherspoon LJ, Henry DN et al. (2006). Effects of breakfast meal composition on second meal metabolic responses in adults with type 2 diabetes mellitus. Eur J Clin Nutr 60, 1122–1129. Coppack SW, Jensen MD, Miles JM (1994). In vivo regulation of lipolysis in humans. J Lipid Res 35, 177–193. Cummings NK, Soares MJ, James AP (2006). The acute effects of different sources of dietary calcium on postprandial energy metabolism. Br J Nutr 96, 138–144. Cummings NK, Soares MJ, Chan She Ping-Delfos W, James AP, Sivakumar P, Mamo JC et al. (2004). Gastric emptying of high calcium breakfast meals: potential effect of paracetamol (acetaminophen) on substrate oxidation rates. Austr Soc Study Obes 79, P25. English R, Lewis J (1991). Nutritional Values of Australian Foods. Australian Government Publishing Service: Canberra. Ercan N, Gannon MC, Nuttall FQ (1994). Effect of added fat on the plasma glucose and insulin response to ingested potato given in various combinations as two meals in normal individuals. Diabetes Care 17, 1453–1459. Foster-Powell K, Holt SH, Brand-Miller JC (2002). International table of glycemic index and glycemic load values. Am J Clin Nutr 76, 5–56. Frape DL, Williams NR, Scriven AJ, Palmer CR, O’Sullivan K, Fletcher RJ (1997a). Diurnal trends in responses of blood plasma concentrations of glucose, insulin, and C-peptide following high- and low-fat meals and their relation to fat metabolism in healthy middle-aged volunteers. Br J Nutr 77, 523–535. Frape DL, Williams NR, Scriven AJ, Palmer CR, O’Sullivan K, Fletcher RJ (1997b). Effects of high- and low-fat meals on the diurnal response of plasma lipid metabolite concentrations in healthy middle-aged volunteers. Br J Nutr 77, 375–390. Frayn KN (1998). Non-esterified fatty acid metabolism and postprandial lipaemia. Atherosclerosis 141, S41–S46. Gunther CW, Legowski PA, Lyle RM, McCabe GP, Eagan MS, Peacock M et al. (2005). Dairy products do not lead to alterations in body weight or fat mass in young women in a 1-y intervention. Am J Clin Nutr 81, 751–756.

European Journal of Clinical Nutrition

Jenkins DJ, Wolever TM, Taylor RH, Griffiths C, Krzeminska K, Lawrie JA et al. (1982). Slow release dietary carbohydrate improves second meal tolerance. Am J Clin Nutr 35, 1339–1346. Landsberg L (2006). Feast or famine: the sympathetic nervous system response to nutrient intake. Cell Mol Neurobiol 26, 495–506. Lee A, Ader M, Bray GA, Bergman RN (1992). Diurnal variation in glucose tolerance: cyclic suppression of insulin action and insulin secretion in normal-weight, but not obese, subjects. Diabetes 41, 750–758. Mathews JN, Altman DG, Campbell MJ, Royston P (1990). Analysis of serial measurements in medical research. Br Med J 300, 230–235. Morgan L, Hampton S, Gibbs M, Arendt J (2003). Circadian aspects of postprandial metabolism. Chronobiol Int 20, 795–808. Norton K, Olds T (2000). Anthropometrica. University of New South Wales Press: Sydney. Robertson MD, Jackon KG, Fielding BA, Williams CM, Frayn KN (2002). Acute effects of meal fatty acid composition on insulin sensitivity in healthy post-menopausal women. Br J Nutr 88, 635–640. Sawaya AL, Fuss PJ, Dallal GE, Tsay R, McCrory MA, Young V et al. (2001). Meal palatability, substrate oxidation and blood glucose in young and older men. Physiol Behav 72, 5–12. Soares MJ, Cummings S, Mamo JC, Kenrick M, Piers LS (2004). The acute effects of olive oil v. cream on postprandial thermogenesis and substrate oxidation in postmenopausal women. Br J Nutr 91, 245–252. Weihrauch JL, Tamaki J (1999). Provisional Table on the Vitamin D Content of Foods. Human Nutrition Information Service, United States Department of Agriculture: Beltsville, USA. Wolever TM, Jenkins DJ, Ocana AM, Rao VA, Collier GR (1988). Second-meal effect: low-glycemic-index foods eaten at dinner improve subsequent breakfast glycemic response. Am J Clin Nutr 48, 1041–1047. Xue BZ, Wilkison WO, Mynatt RL, Moustaid N, Goldman M, Zemel MB (1999). The agouti gene product stimulates pancreatic b-cell Ca2 þ signalling and insulin release. Physiol Genomics 1, 11–19. Zemel MB (2002). Regulation of adiposity and obesity risk by dietary calcium: mechanisms and implications. J Am Coll Nutr 21, S146–S151. Zemel MB, Richards J, Mathis S, Milstead A, Gebhardt L, Silva E (2005a). Dairy augmentation of total and central fat loss in obese subjects. Int J Obes Relat Metab Disord 29, 391–397. Zemel MB, Richards J, Milstead A, Campbell P (2005b). Effects of calcium and dairy on body composition and weight loss in African-American adults. Obes Res 13, 1218–1225. Zemel MB, Shi H, Greer B, Dirienzo D, Zemel PC (2000). Regulation of adiposity by dietary calcium. FASEB J 14, 1132–1138. Zemel MB, Thompson W, Milstead A, Morris K, Campbell P (2004). Calcium and dairy acceleration of weight and fat loss during energy restriction in obese adults. Obes Res 12, 582–590.