The close relationship between excess body weight, insulin resistance, and type 2 diabetes is well known (1). A decrease in weight improves insulin sensitivity, metabolic control, and the necessity for oral or insulin treatment (2). Dietary monitoring is a cornerstone of diabetes treatment, being essential to both improve current food intake and aid in the explanation of any disturbance in metabolic control. Dietary monitoring of food and energy intake (EI) therefore assumes that the dietary reports of diabetic patients are valid. Meanwhile, underreporting of food intake is widely acknowledged in obese individuals (3,4) yet is generally disregarded in the type 2 diabetic population, despite the fact that a significant percentage of this population is obese. The issue of EI underreporting and diabetes was first raised by Prentice et al. (5) and has since been discussed in only a few studies conducted in type 2 diabetic patients (68). The doubly labeled water (DLW) method is a state-of-the-art technique used to estimate total energy expenditure (TEE) (9). TEE refers to the energy requirement of an individual, which corresponds to EI under conditions of weight stability. The comparison of the ratio of EI to resting energy expenditure (REE) and the ratio of TEE to REE (i.e., the physical activity level) is the tool conventionally used to detect for misreporting (4). To date, the DLW method has been used in the measurement of TEE in only one study (10), which involved a small number of subjects. However, the DLW method has not been used to quantify misreporting in the type 2 diabetic population, and current evidence relies on the results of proxy measurements of TEE, such as REE prediction equations and physical activity monitors (68). We have therefore undertaken this study to directly measure and compare TEE and REE with EI in a group of 21 obese patients, 12 of whom have type 2 diabetes.

Twelve weight-stable, obese, type 2 diabetic patients treated with metformin only (11) and 9 obese nondiabetic patients participated in this study, which was approved by the hospital’s ethical committee. Height was measured to the nearest 0.1 cm (SECA, Hamburg, Germany). Weight was measured to the nearest 0.01 kg (DS Medica, Rome, Italy). Percentage of fat and fat-free mass were determined using a four-compartmental model to measure body volume and bone mineral content, with the estimation of total body water conducted using bioelectrical impedance (12,13). An experienced dietitian estimated usual food intake using a 3-day food recall and calculated the total daily EI, i.e., the energy from all reported meals and snacks combined, using computerized French food composition tables (BILNUT 4.0. S.C.D.A. Nutrisoft, 1995). Protein, lipid, carbohydrate, and alcohol intake were calculated in the same way. REE was measured by indirect calorimetry using a ventilated hood system (Vmax Spectra; SensorMedics, Yorba Linda, CA) as already described (14). TEE was measured using the DLW method as previously described by Ritz et al. (15). Briefly, subjects consumed a weight-dependent dose of DLW. Urine samples were collected before dosing and everyday for the following 14 days. Isotopic analyses were performed as described by Ripoche et al. (16), using the multipoint method (15). Statistical analysis was performed using Statview software (Abacus Concept, Cary, NC). Results are expressed as means ± SD. Cross-sectional differences between categories were assessed by ANOVA.

Table 1 summarizes the characteristics of the two patient groups who were matched for weight, body composition, and BMI. The diabetic patients were older and had a higher HbA1c than their nondiabetic counterparts. TEE and REE did not significantly differ between the groups. Reported EI was, however, significantly lower in the diabetic group. The ratio of TEE to REE fell within the expected range (4) and was similar between the diabetic and nondiabetic groups. Although the ratio of EI to REE was low in nondiabetic patients, it was significantly lower in the diabetic patients, suggesting that these patients reported eating 22% less energy than is necessary for them to maintain even basic functions to live. Furthermore, given that the range of EI to REE ratios was between 0.5 and 1.23 in this diabetic group, all 12 of these patients underreported (4). All but one patient in the nondiabetic group had an EI to TEE ratio that fell below 0.79, which represents the lower 95% CI of the cutoff value for underreporting (6). In both patient groups, women underreported to a greater extent than men (data not shown), as has been shown in previous studies (8).

Underreporting food intake has been suggested as a problem in diabetic patients. The present data has validated this in an unquestionable way, given that we used a state-of-the-art tool (DLW) to measure TEE. In this study, reported EI was just adequate to match REE values in the nondiabetic patient group but was wholly insufficient to meet the REE values in the diabetic group. This suggests that in this group we have to multiply reported food intake by ∼2.5 to attain a credible EI. This factor, however, varies considerably between patients and probably does not affect carbohydrate and fat intake to the same extent. Further studies are necessary to understand the reasons why diabetic patients underreport more than their obese counterparts. Dietary educational programs should consider this aspect when dealing with patients to better understand how their feeding behavior can be improved.

Table 1—

Characteristics of the diabetic and the nondiabetic patient groups

Diabetic patientsNondiabetic patientsP
Weight (kg) 105.7 ± 17.2 100.3 ± 19.0 0.504 
BMI (kg/m237.1 ± 4.67 37.0 ± 3.40 0.951 
Fat mass (%) 39.7 ± 4.7 40.6 ± 7.6 0.980 
HbA1c (%) 7.52 ± 1.11 5.15 ± 0.40 <0.0001 
TEE (kcal/day) 3,863 ± 1,890 3,389 ± 887 0.496 
REE (kcal/day) 2,020 ± 421 1,805 ± 470 0.283 
EI (kcal/day) 1,520 ± 266 1,934 ± 331 0.005 
TEE to REE 1.857 ± 0.503 1.881 ± 0.198 0.891 
EI to REE 0.778 ± 0.198 1.109 ± 0.251 0.0031 
EI to TEE 0.438 ± 0.129 0.600 ± 0.178 0.026 
Diabetic patientsNondiabetic patientsP
Weight (kg) 105.7 ± 17.2 100.3 ± 19.0 0.504 
BMI (kg/m237.1 ± 4.67 37.0 ± 3.40 0.951 
Fat mass (%) 39.7 ± 4.7 40.6 ± 7.6 0.980 
HbA1c (%) 7.52 ± 1.11 5.15 ± 0.40 <0.0001 
TEE (kcal/day) 3,863 ± 1,890 3,389 ± 887 0.496 
REE (kcal/day) 2,020 ± 421 1,805 ± 470 0.283 
EI (kcal/day) 1,520 ± 266 1,934 ± 331 0.005 
TEE to REE 1.857 ± 0.503 1.881 ± 0.198 0.891 
EI to REE 0.778 ± 0.198 1.109 ± 0.251 0.0031 
EI to TEE 0.438 ± 0.129 0.600 ± 0.178 0.026 

Data are means ± SD. P by ANOVA.

This study was supported by grants from PHRC 2002, Alfediam, and Roche SA.

1.
Resnick HE, Valsania P, Halter JB, Lin X: Relation of weight gain and weight loss on subsequent diabetes risk in overweight adults.
J Epidemiol Community Health
54
:
596
–602,
2000
2.
Goldstein DJ: Beneficial health effects of modest weight loss.
Int J Obes Relat Metab Disord
16
:
397
–415,
1992
3.
Livingstone MB, Black AE: Markers of the validity of reported energy intake.
J Nutr
133(Suppl. 3)
:
895S
–920S,
2003
4.
Black AE, Goldberg GR, Jebb SA, Livingstone MB, Cole TJ, Prentice AM: Critical evaluation of energy intake data using fundamental principles of energy physiology. 2. Evaluating the results of published surveys.
Eur J Clin Nutr
45
:
583
–599,
2003
5.
Prentice AM, Black A, Goldberg G: Diabetes and energy intake.
Am J Clin Nutr
57
:
596
–598,
1993
6.
Samuel-Hodge CD, Fernandez LM, Henriquez-Roldan CF, Johnston LF, Keyserling TC: A comparison of self-reported energy intake with total energy expenditure estimated by accelerometer and basal metabolic rate in African-American women with type 2 diabetes.
Diabetes Care
27
:
663
–669,
2004
7.
Matsushita Y, Yokoyama T, Homma T, Tanaka H, Kawahara K: Relationship between the ability to recognize energy intake and expenditure, and blood sugar control in type 2 diabetes mellitus patients.
Diabetic Res Clin Pract
67
:
220
–226,
2005
8.
Adams SJ: The dietary intake of people with non insulin diabetes (NIDDM): how valid is self-reported intake?
J Hum Nutr Diet
11
:
296
–306,
1998
9.
Schoeller DA: How accurate is self-reported dietary energy intake?
Nutr Rev
48
:
373
–379,
1990
10.
Chong PK, Jung RT, Rennie MJ, Scrimgeour CM: Energy expenditure in type 2 diabetic patients on metformin and sulphonylurea therapy.
Diabet Med
12
:
401
–408,
1995
11.
Expert Committee on the Diagnosis and Classification of Diabetes Mellitus: Report of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus.
Diabetes Care
20
:
1183
–1197,
1997
12.
Sallé A, Ryan M, Guilloteau G, Bouhanick B, Berrut G, Ritz P: “Glucose control-related” and “non-glucose control-related” effects of insulin on weight gain in newly insulin-treated type 2 diabetic patients.
Br J Nutr
94
:
931
–937,
2005
13.
Packianathan IC, Fuller NJ, Peterson DN, Wright A, Coward WA, Finer N: Use of a reference
4 
-compartment model to define the effect of insulin treatment on body composition in type 2 diabetes: the “Darwin study.“
Diabetologia
48
:
222
–229,
2005
14.
Ryan M, Sallé A, Guilloteau G, Genetay M, Livingstone MBE, Ritz P: Resting energy expenditure is not increased in mildly hyperglycaemic obese diabetic patients.
Br J Nutr
In press
15.
Ritz P, Cole TJ, Couet C, Coward WA: Precision of DLW energy expenditure measurements: contribution of natural abundance variations.
Am J Physiol
270
:
E164
–E169,
1996
16.
Ripoche N, Krempf M, Ritz P: Deuterium and 18-oxygen enrichments in biological fluids in continuous flow elemental analyser with isotope ratio mass spectrometer using two configurations.
J Mass Spectrom
41
:
1212
–1218,
2006

A table elsewhere in this issue shows conventional and Système International (SI) units and conversion factors for many substances.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.