Assessment of Postprandial Glucose Excursions Throughout the Day in Newly Diagnosed Type 2 Diabetes

Abstract

Background:

A growing body of evidence suggests that postprandial glucose (PPG) is independently linked to multiple complications and that testing of PPG should be added to hemoglobin A1c (HbA1c) and fasting glucose measurements in the evaluation of glycemic control of type 2 diabetes patients. An ongoing debate is questioning how to assess PPG. This observational study looks further into this question in a cohort of newly diagnosed type 2 diabetes patients.

Subjects and Methods:

PPG characteristics and intra-/intersubject variations post-breakfast, -lunch, and -dinner, obtained from continuous glucose monitoring (CGM), were retrospectively analyzed in 86 newly diagnosed non–insulin-treated type 2 diabetes patients.

Results:

In total, 462 recorded meals were analyzed. The area under the curve 1–4 h postmeal was significantly larger after breakfast compared with both lunch and dinner (P<0.001). Time to peak was approximately 90 min and did not differ significantly between meals. However, the distribution of the blood glucose peaks was only normally distributed among breakfasts, and time to peak had a day-to-day correlation coefficient of 0.60, compared with a nonsignificant result for lunch and dinner. Breakfast PPG peaks were highly correlated to HbA1c (P<0.05, r=0.64) and had a day-to-day correlation coefficient of 0.86 compared with 0.44 for lunch and 0.74 for dinner.

Conclusions:

Self-monitoring of blood PPG should be evaluated with care. From our data, monitoring of PPG patterns in newly diagnosed type 2 diabetes patients should preferably be obtained following breakfast for a more consistent assessment, reducing day-to-day variations.

Introduction

T he association between hemoglobin A1c (HbA1c) and diabetes complications has led to the use of HbA1c as the gold standard for assessing glycemic control in diabetes.^1,2 Reducing fasting plasma glucose has been a primary target for lowering HbA1c.¹ However, Monnier et al.³ showed that loss of postprandial glucose (PPG) control precedes deterioration of fasting glycemic control in type 2 diabetes patients. Subsequently, a growing body of evidence from clinical studies suggests that postprandial hyperglycemia in type 2 diabetes patients is independently associated with cardiovascular risk, oxidative stress, retinopathy, cognitive dysfunction, and certain types of cancer.^4

–8

In type 2 diabetes, an ongoing debate is how to assess PPG.^4,9,10 The latest guidelines from the American Diabetes Association recommend assessment of PPG in a wide window 1–2 h postmeal.¹¹ Standl et al.¹² likewise concluded that the absence of a generally accepted standard of estimation of postprandial hyperglycemia/variability adds to the challenge. The magnitude and time to peak for PPG excursions depend on a variety of factors, including pathophysiology, timing, quantity, and composition of the meal. Generally, 2-h measurements of PPG are considered an approximation of peak glucose and provide a reasonable assessment of postprandial hyperglycemia, although this is largely based on expert opinions rather than on solid evidence.^4,10 It is not surprising that 2-h post meal measurement is often recommended to assess blood glucose as this value will be within the normal reference interval in the population without diabetes.¹³ Also, numerous studies have demonstrated a relation between 2-h PPG levels and cardiovascular disease.^14
–16

In type 2 diabetes patients there is evidence that blood PPG peak values (PPG peaks) are associated with development of long-term cardiovascular complications, independent of fasting blood glucose and HbA1c.¹⁷ Furthermore, oxidative stress is activated by acute glucose fluctuations,¹³ and measures of endothelial dysfunction and oxidative stress seem to be closely related to glucose peaks.^6,12 This adds to the interest of assessing more than general postprandial hyperglycemia.

The assessment of PPG may be influenced by the time of the day, and the optimal time for evaluating PPG is still unclear. The aim of this study was to investigate the difference in PPG excursions throughout the day in newly diagnosed non–insulin-treated type 2 diabetes patients.

Subjects and Methods

In total, 52 men and 48 women with newly diagnosed type 2 diabetes (defined as diabetes known for <5 years) were consecutively recruited from the outpatient clinics at Aarhus University Hospital, Aarhus, Denmark.

Inclusion criteria were age >18 years and diagnosis of type 2 diabetes according to World Health Organization criteria.¹⁸

Exclusion criteria were acute or chronic infectious disease, end-stage renal failure, pregnancy or lactation, prior or present cancer, contraindications to magnetic resonance imaging, and use of insulin.

All patients were equipped with a continuous glucose monitoring (CGM) sensor, the CGMS^® iPro^™ continuous glucose recorder (Medtronic Minimed, Northridge, CA), for 3 days. The CGM system was inserted at Day 0 in the morning and removed at Day 3. The CGM readings were blinded for the patients during data collection. Insertion of the sensor occurred at Aarhus University Hospital. The patients were instructed to maintain their usual diet and treatment as well as instructed not to modify their meal timing over the period with CGM. The patients were also instructed to perform self-monitoring of blood glucose (SMBG) measurements, using the OneTouch^® Ultra^® 2 from LifeScan (Milpitas, CA), before breakfast, lunch, and dinner and at bedtime. The CGM and SMBG data collection software versions were synchronized at insertion of the CGM sensor. Upon return of the equipment, data were downloaded to a computer for further analysis of the glucose profile and postprandial characteristics. Clinical data were collected prior to CGM at a preliminary examination. Clinical data included the patient's medical history, lifestyle information, blood pressure, electrocardiogram, and fasting blood samples, including plasma glucose, HbA1c, and lipid profile. The study protocols were approved by the local ethical committee according to the principles of the “Helsinki Declaration II.” All patients gave their written informed consent.

Analysis of CGM data

Premeal SMBG measurements were assumed to be in direct relation with food ingestion and were thus used for meal recordings. Meals recorded during CGM were divided into groups: breakfast, lunch, and dinner, depending on the time of day of the recording. Breakfast was defined as the first meal of the day within the time span from 5 a.m. to 11 a.m. Lunch was defined as the first meal from 11 a.m. to 3 p.m., and dinner was defined as the first meal between 5 p.m. and 9 p.m. Other meals were not used for calculations in this study. The rationale for defining these time spans was based on the eating pattern seen in Nordic people.¹⁹ Based upon these definitions an automated computer program, designed in MATLAB^® R2011b (MathWorks, Natick, MA), calculated postprandial measures. These measures are described below.

Every 30 min until 4 h absolute PPG values were obtained. Also, postmeal relative areas under the CGM curve (AUCpp) for h 1–4 were calculated as a measure for postprandial excursion. The AUCpp was calculated as glucose excursion relative to baseline (premeal) value. PPG peaks were defined as the highest glucose value within a 3-h window from meal start. Meals with adjoining meals 2 h before or after were excluded from further calculations.

Statistical analysis

Results are given as mean (SD) or as median (25^th percentile; 75^th percentile) according to whether the measure were normally distributed or not. Normal distribution was tested with a Lilliefors tests for normality.²⁰ Comparison between mean glucose values in different meals were tested with a one-way analysis of variance or Kruskal–Wallis test for nonparametric measures. Tukey's least significant difference procedure was used to compare individual differences among groups, and the χ² test was used to test for significant difference between proportions. Correlation between glucose measures and HbA1c was calculated using Pearson's correlation.

For subjects with repeated measures for a given meal type (breakfast, lunch, or dinner) during the 3 days of monitoring, day-to-day variations were analyzed as intraclass correlation coefficient (ICC) using a two-factor mixed effects model and the type absolute agreement.

Statistical comparisons were considered significant when P values were<0.05. For statistical data processing, MATLAB R2011b (version 7.13.0.564) and SPSS^® version 19 (IBM, Armonk, NY) were used as appropriate.

Results

In total, 86 of the 100 patients enrolled were used in this study. All had at least one successful day of CGM recording as well as four corresponding SMBG values. Data from 14 patients were excluded from this study because of use of insulin or missing CGM data (less than 24 h of successful CGM recording combined with less than four SMBG calibrations). CGM needle rip outs with duration less than 10 min were not treated as missing data. The patients' characteristics are listed in Table 1. In total, 462 meals distributed among breakfast, lunch, and dinner were recorded and used for calculations. Six meals were excluded from calculations because of close adjoining meals. In Figure 1A the average 24-h profile is shown.

FIG. 1.

(A) The 24-h mean (SD) glycemic profiles from 86 type 2 diabetes patients. (B) Mean postprandial value (with SEM) for breakfast, dinner, and lunch. (C–E) Breakfast, lunch, and dinner histograms (from left to right) for postprandial peak time.

Table 1.

Patients' Characteristics

	Value
Number of type 2 diabetes patients	86
Females (%)	55
Age (years)	61 (10)
Hemoglobin A1c (%)	6.5 (0.65)
Weight (kg)	87 (16)
BMI (kg/m²)	30 (4.5)
Current smokers (%)	22
Blood pressure (mm Hg)
Diastolic	79 (8)
Systolic	126 (12)
Pulse pressure (mm Hg)	45 (14)
Heart rate (bpm)	66 (10)
Total cholesterol (mmol/L)	4.4 (0.78)
LDL cholesterol (mmol/L)	2.3 (0.67)
HDL cholesterol (mmol/L)	1.4 (0.32)
Triglycerides (mmol/L)	1.4 (1; 1.9)
Medication (%)
Metformin	62
Blood pressure	63
Lipid	77

Measures are shown as mean (SD) values or median (25^th percentile; 75^th percentile) as indicated. Normality was evaluated by the Liliefors test.²⁰

BMI, body mass index; HDL, high-density lipoprotein; LDL, low-density lipoprotein.

The postprandial characteristics for breakfast, lunch, and dinner are shown in Table 2 and Figure 1. AUCpp values 1–4 h postmeal were larger for breakfast compared with both lunch (P<0.001) and dinner (P<0.001). This is also illustrated in Figure 1B, showing that the postprandial excursion remains larger until 150 min postmeal.

Table 2.

Summary of Meal Characteristics and Correlations

Postprandial measure	Breakfast	Lunch	Dinner	P value ^a
Number	161	142	159
AUCpp CGM (h/mmol/L)
1 h	13.0 (12.1)	4.3 (9.3)	4.1 (10.2)	<0.001
2 h	36.4 (32.2)	14.2 (27.4)	14.0 (30.0)	<0.001
4 h	46.2 (48.7)	19.9 (44.9)	22.7 (54.0)	<0.001
4 h	47.5 (67.1)	23.4 (61.1)	28.1 (76.2)	<0.001
Premeal (mmol/L)	6.9 (1.2)	6.6 (1.4)	6.6 (1.5)	NS
PPG peak (mmol/L)	10.2 (2.5)	8.4 (1.6)	8.8 (2.1)	<0.001
Peak height (mmol/L)^b	3.2 (2.1)	1.8 (1.5)	2.1 (1.8)	<0.001
Time to peak (min)	88 (64; 105)	91 (55; 135)	91 (45; 140)	NS
Peak time 60–120 min postprandial (%)^c	55	36	25	<0.001
Correlation with HbA1c ^d
AUCpp CGM
1 h	NS	NS	NS
2 h	0.25	NS	NS
3 h	0.32	NS	0.52
4 h	0.33	NS	NS
Premeal	0.59	0.29	0.58
PPG peak	0.64	0.43	NS
Peak height	0.41	0.20	NS
Time to peak	0.15	NS	0.26
Intraclass correlation ^{d e}
AUCpp CGM
1 h	0.58	NS	0.28
2 h	0.63	NS	0.42
3 h	0.68	NS	0.43
4 h	0.69	NS	0.42
Premeal	0.64	0.49	0.46
PPG peak	0.86	0.44	0.74
Peak height	0.77	NS	0.51
Time to peak	0.60	NS	NS

Measures are shown as mean (SD) values or median (25^th percentile; 75^th percentile) for nonparametric distributed measures (normality evaluation by Liliefors' test²⁰) as indicated.

One-way analysis of variance test is used for parametric measures, and the Kruskal–Wallis test is for nonparametric measures.

Relative postprandial glucose (PPG) peak height, calculated as (PPG peak – premeal glucose).

Proportion of PPG peaks within a window 1–2 h from ingestion.

Correlations are shown if α is below 0.05.

Presented as intraclass correlation coefficient.

AUCpp CGM, postmeal relative areas under the continuous glucose monitoring curve; HbA1c, hemoglobin A1c; NS, nonsignificant correlation.

Average time to peak was not significantly different between meals (P=0.979), but the PPG peak time was only normally distributed among breakfasts as illustrated in Figure 1C–E. In addition, a relatively larger percentage of peaks occurred postprandial after breakfast within 60–120 min (P<0.001; breakfast 55%, lunch 36%, dinner 25%). Relative and absolute breakfast peaks were larger compared with lunch (relative and absolute, P<0.001) and dinner (relative and absolute, P<0.001). When comparing glucose measures between lunch and dinner, a significant difference was seen only in the absolute peak (P=0.028). As shown in Table 2, breakfast measures also showed a significantly higher correlation to HbA1c (i.e., the absolute peak revealed the highest correlation) (P<0.05, r=0.64).

Intrapatient findings

ICC analyses indicated that breakfast excursions were most consistent from day to day; results are presented in Table 2. Time to peak for breakfast had an ICC of 0.60 (P<0.05), whereas correlations for lunch and dinner were not significant. The absolute breakfast PPG peak also had an ICC of 0.86 (P<0.05) compared with 0.44 (lunch, P<0.05) and 0.74 (dinner, P<0.05). For the purpose of calculating ICC, 24 meals (six breakfasts, eight lunches, and 10 dinners) were excluded because of lack of matching meals.

Discussion

This study shows that breakfast is associated with larger and more consistent PPG excursions intrapatient as well as interpatient than those seen after lunch and dinner. Our results seem to be in line with previous findings, pointing at breakfast as the largest contributor to PPG excursions.³

ICC analysis between meals revealed that PPG after breakfast was strongly correlated between days but poorly correlated with PPG after lunch and dinner. In combination with findings of Monnier et al.,³ this raises the question of whether patients with newly diagnosed type 2 diabetes could benefit from initial CGM profiling and subsequent SMBG post-breakfast (based on the profile) in order to evaluate glycemic control.

The current recommendations for SMBG are mainly based on expert opinions, and the 2-h time frame for SMBG is often used for evaluation of PPG control.⁴ The intersubject variation in time to peak seen at lunch/dinner indicates that SMBG should be evaluated with care and might in fact not reflect the true glycemic control but rather a complex combination of the underlying insulin resistance, β-cell dysfunction, meal composition/timing, previous meals, and daily activities. In a study of type 1 diabetes patients Johansen et al.²¹ found significant interindividual and intraindividual variations in postprandial glycemic peak time but did not investigate meal variations throughout the day. This study concludes that simple and general recommendations regarding postprandial SMBG for detection of maximum values is problematic because of these variations.

Compared with lunch and dinner, AUCpp 1–4 h reveals that breakfast is associated with relatively larger excursions from premeal values. This may indicate that lunch and dinner meals induce lower glucose excursions and/or have a larger degree of glycemic interpatient variability than breakfast. The high degree of intrapatient PPG fluctuations found in this study clearly demonstrates the challenges in interpretation of post-lunch/dinner excursions based upon a single measurement of blood glucose. Zaccardi et al.²² suggested that SMBG may not adequately reflect glycemic excursions because each value reflects only a single time point in 24-h glucose dynamics and should therefore be avoided. Several studies have shown association between SMBG in non–insulin-treated type 2 diabetes patients and outcome.^23

–28 Based on these studies there is little evidence to support SMBG for this group. The lack of significant associations in these studies can perhaps be attributed to assessing PPG throughout the day and not breakfast PPG alone, as supported by our study.

Bonora et al.²⁹ found in outpatients with type 2 diabetes that HbA1c showed better correlation with preprandial glucose values than with postprandial levels. Our results did not find similar tendencies, but in both studies single measures of fasting and PPG were relatively poor predictors of HbA1c. Additionally, Bonora et al.²⁹ revealed that neither preprandial levels nor HbA1c provided accurate information regarding postprandial peaks and also that postprandial peaks were poorly correlated throughout the day. Our results confirm this complexity of assessing postprandial peaks. Monnier et al.³ have shown how postprandial control is lost before it can be detected in fasting glucose values. In this context the potential implication from our data could be that assessing breakfast postprandial peaks seems suitable to detect initial loss of postprandial control in otherwise well-regulated type 2 diabetes outpatients.

It remains speculative why post-breakfast measurements are more consistent than other postmeal measurements regarding PPG excursions, in both inter- and intraindividual variability. It may be due to the fact that the real fasting state occurs in the early morning,³⁰ and breakfast therefore is the only meal not affected by previous meals and daily activities. Hence, breakfast PPG may reflect the response to intestinal carbohydrate absorption, insulin secretion, hepatic glucose output, and the insulin resistance state without confounding from previous meals and daily activities. Furthermore, hepatic glucose output reaches peak production at the end of the night and decline until late afternoon, which may partly explain why breakfast excursions are larger than those seen at lunch and dinner.³⁰

Data from this study have been obtained in real-life settings, and the patients were responsible for SMBG recordings before eating. Based on the distribution of interpatient PPG peak time (Fig. 1C–E) it seems reasonable to consider if delay of recording SMBG has caused the early postprandial peak times in the lunch and dinner group. This is probably the case in a fraction of the meals, but there is no reason to believe that recordings at lunch or dinner are more biased compared with breakfast recordings in this regard.

A major strength of this study is the homogeneous cohort of well-controlled type 2 diabetes patients and the use of CGM to assess glucose levels.

A potential limitation of this study is the lack of diet recording. Carbohydrate content and general meal composition can influence postprandial excursions. This is difficult to assess as self-reported food intake is known to be heavily influenced by reporting bias and intervention bias.^31
–33

Ingesting snacks shortly before a meal could possibly alter PPG readings, and snacking was not recorded in this study. At breakfast this limitation can reasonably be assumed not to be of significant influence on the postprandial readings, as breakfast typically is the first food consumption of the day. Regarding post-lunch and -dinner measurements, snacking may explain some of the variation seen in this study.

In perspective, it is crucial to understand the underlying mechanisms for the observed variations in PPG excursions between meals, in order to give the patient an individual tailored treatment and thereby improve glycemic control. Examining diet data in relation to these findings could add additional information. Further studies will clarify these aspects.

Based on our data, monitoring of PPG patterns in newly diagnosed type 2 diabetes patients should preferably be obtained following breakfast for a more consistent feedback, reducing day-to-day variations. Our study suggests that the optimal time to measure peak blood PPG values is around 90 min after the beginning of a meal; this is based on the median time to peak of approximately 90 min (Table 2).

SMBG results should always be evaluated with care, and an optimized protocol for blood glucose measurements should be considered.

Footnotes

Acknowledgments

The authors acknowledge the contribution of Karin Ø. Kristensen who revised the language of the manuscript.

Author Disclosure Statement

No competing financial interests exist.

References

Nathan

, Buse

, Davidson

, Heine

, Holman

, Sherwin

, Zinman

. Management of hyperglycemia in type 2 diabetes: A consensus algorithm for the initiation and adjustment of therapy: a consensus statement from the American Diabetes Association and the European Association for the Study of Diabetes. Diabetes Care, 2006; 29:1963–1972.

Skyler

. Evidence for glycated hemoglobin targets from the Diabetes Control and Complications Trial. Endocr Pract, 2002; 8:27–30.

Monnier

, Colette

, Dunseath

, Owens

. The loss of postprandial glycemic control precedes stepwise deterioration of fasting with worsening diabetes. Diabetes Care, 2007; 30:263–269.

Ceriello

, Colagiuri

, Gerich

, Tuomilehto

. Guideline for management of postmeal glucose. Nutr Metab Cardiovasc Dis, 2008; 18:S17–S33.

Aryangat

, Gerich

. Type 2 diabetes: postprandial hyperglycemia and increased cardiovascular risk. Vasc Health Risk Manag, 2010; 6:145–155.

Monnier

, Mas

, Ginet

, Michel

, Villon

, Cristol

, Colette

. Activation of oxidative stress by acute glucose fluctuations compared with sustained chronic hyperglycemia in patients with type 2 diabetes. JAMA, 2006; 295:1681–1687.

Wright

, Scism-Bacon

, Glass

. Oxidative stress in type 2 diabetes: the role of fasting and postprandial glycaemia. Int J Clin Pract, 2006; 60:308–314.

Cavalot

, Petrelli

, Traversa

, Bonomo

, Fiora

, Conti

, Anfossi

, Costa

, Trovati

. Postprandial blood glucose is a stronger predictor of cardiovascular events than fasting blood glucose in type 2 diabetes mellitus, particularly in women: lessons from the San Luigi Gonzaga Diabetes Study. J Clin Endocrinol Metab, 2006; 91:813–819.

Monnier

, Colette

, Owens

. Postprandial, basal glucose in type 2 diabetes: assessment, respective impacts. Diabetes Technol Ther, 2011; 13,Suppl 1:S-25–S-32.

10.

American Diabetes Association. Postprandial blood glucose. Diabetes Care, 2001; 19:775–778.

11.

American Diabetes Association. Standards of medical care in diabetes—2012. Diabetes Care, 2012; 35,Suppl 1:S11–S63.

12.

Standl

, Schnell

, Ceriello

. Postprandial hyperglycemia and glycemic variability: should we care? Diabetes Care, 2011; 34,Suppl 2:S120–S127.

13.

Monnier

, Colette

, Owens

. Integrating glycaemic variability in the glycaemic disorders of type 2 diabetes: a move towards a unified glucose tetrad concept. Diabetes Metab Res Rev, 2009; 25:393–402.

14.

Meigs

, Agostino

, Nathan

, Wilson

. Fasting and postchallenge glycemia and cardiovascular disease risk: the Framingham Offspring Study. Diabetes Care, 2002; 25:1845–1850.

15.

The DECODE Study Group. Glucose tolerance and cardiovascular mortality. Arch Intern Med, 2001; 161:397–404.

16.

Saydah

, Miret

, Sung

, Varas

, Gause

, Brancati

. Postchallenge hyperglycemia and mortality in a national sample of U.S. adults. Diabetes Care, 2001; 24:1397–1402.

17.

Nalysnyk

, Hernandez-Medina

, Krishnarajah

. Glycaemic variability and complications in patients with diabetes mellitus: evidence from a systematic review of the literature. Diabetes Obes Metab, 2010; 12:288–298.

18.

World Health Organization: Definition. Definition and Diagnosis of Diabetes Mellitus and Intermediate Hyperglycemia. 2006. www.who.int/diabetes/publications/en/. 2012 June 15.

19.

Kjærnes

. Eating Patterns. A Day in the Lives of Nordic Peoples. SIFO Report. Lysaker, Norway: National Institute for Consumer Research, 2001.

20.

Lilliefors

. On the Kolmogorov-Smirnov test for the exponential distribution with mean unknown. J Am Stat Assoc, 1969; 64:387–289.

21.

Johansen

, Gjerløv

, Christiansen

, Hejlesen

. Interindividual and intraindividual variations in postprandial glycemia peak time complicate precise recommendations for self-monitoring of glucose in persons with type 1 diabetes mellitus. J Diabetes Sci Technol, 2012; 6:356–361.

22.

Zaccardi

, Pitocco

, Ghirlanda

. Glycemic risk factors of diabetic vascular complications: the role of glycemic variability. Diabetes Metab Res Rev, 2009; 25:199–207.

23.

Gerrald

, Malone

, Shilliday

. Clinical benefit of self-monitoring of blood glucose is uncertain for non–insulin-treated patients with type 2 diabetes. Clin Diabetes, 2010; 28:121–123.

24.

Malanda

, Welschen

, Riphagen

, Dekker

, Nijpels

, Bot

. Self-monitoring of blood glucose in patients with type 2 diabetes who are not using insulin. Cochrane Database Syst Rev, 2012; 1:CD005060.

25.

Jansen

. Self-monitoring of glucose in type 2 diabetes mellitus: a Bayesian meta-analysis of direct and indirect comparisons. Curr Med Res Opin, 2006; 22:671–81.

26.

Martin

, Schneider

, Heinemann

, Lodwig

, Kurth

H-J

, Kolb

, Scherbaum

. Self-monitoring of blood glucose in type 2 diabetes and long-term outcome: an epidemiological cohort study. Diabetologia, 2006; 49:271–278.

27.

Simon

, Gray

, Clarke

, Wade

, Neil

, Farmer

. Cost effectiveness of self monitoring of blood glucose in patients with non-insulin treated type 2 diabetes: economic evaluation of data from the DiGEM trial. BMJ, 2008; 336:1177–1180.

28.

O'Kane

, Bunting

, Copeland

, Coates

. Efficacy of self monitoring of blood glucose in patients with newly diagnosed type 2 diabetes (ESMON study): randomised controlled trial. BMJ, 2008; 336:1174–1177.

29.

Bonora

, Calcaterra

, Lombardi

, Bonfante

, Formentini

, Bonadonna

, Muggeo

. Plasma glucose levels throughout the day and HbA1c interrelationships in type 2 diabetes. Diabetes Care, 2001; 24:2023–2029.

30.

Monnier

. Is postprandial glucose a neglected cardiovascular risk factor in type 2 diabetes? Eur J Clin Invest, 2000; 30,Suppl 2:3–11.

31.

Samuel-Hodge

, Fernandez

, Henríquez-Roldán

, Johnston

, Keyserling

. A comparison of self-reported energy estimated by accelerometer and basal metabolic rate in African-American women with type 2 diabetes. Diabetes Care, 2004; 27:663–669.

32.

Adams

. The dietary intake of people with non-insulin-dependent diabetes (NIDDM): how valid is self-reported intake? J Hum Nutr Diet, 1998; 11:295–306.

33.

Lichtman

, Pisarska

, Berman

, Pestone

, Dowling

, Offenbacher

, Weisel

, Heshka

, Matthews

, Heymsfield

. Discrepancy between self-reported and actual caloric intake and exercise in obese subjects. N Engl J Med, 1992; 327:1893–1898.