- Email: [email protected]
Underreporting of energy intake in an elderly German population
APPLIED NUTRITIONAL INVESTIGATION
Underreporting of Energy Intake in an Elderly German Population Petra M. Lu¨hrmann, PhD, Birgit M. Herbert, PhD, and Monika Neuha¨user-Berthold, PhD From the Institute of Nutritional Science, University of Giessen, Giessen, Germany OBJECTIVES: Within the longitudinal study on nutrition and health status in an aging population in Giessen, Germany (GISELA), the underreporters of energy intake (EI) were identified and characterized. METHODS: EI was assessed in 238 female and 105 male participants of the GISELA study (age range ⫽ 60 – 89 y) by means of a 3-day estimated dietary record developed especially for this study. Resting metabolic rate (RMR) was measured by indirect calorimetry after an overnight fast. EI was expressed as a multiple of RMR and subjects with an EI:RMR ratio below 1.073 were classified as underreporters. RESULTS: Mean EI:RMR was 1.62 ⫾ 0.46 in females and 1.53 ⫾ 0.46 in males; 7.6% of the females and 16.2% of the males were identified as underreporters. They showed lower levels of education and significantly greater body weight, body mass index, and fat mass than the adequate reporters. Further, underreporters stated more often than adequate reporters that they want to lose weight. Except for -carotene in males, reported nutrient intakes were significantly lower in underreporters than in adequate reporters. Carbohydrate and fat intake in both sexes, protein intake in females calculated as a percentage of EI, and vitamin and mineral densities were not affected by underreporting. CONCLUSIONS: The results indicate that underreporting of EI is related to a low educational level and greater body weight, body mass index, and fat mass and affects all nutrients. These findings should be considered when the association between nutrition and health status is investigated in the elderly. Nutrition 2001;17:912–916. ©Elsevier Science Inc. 2001 KEY WORDS: elderly, dietary survey, 3-d estimated dietary record, ratio of energy intake to resting metabolic rate, underreporting
INTRODUCTION Since the beginning of the 19th century in Germany and other Western countries, the proportion of elderly people has continuously increased.1 Because of this demographic trend, more information concerning the relation between nutrition and health status of older age groups is needed. Dietary surveys are important for the assessment of nutrition status. However, from several studies comparing total energy expenditure measured by double-labeled water with reported energy intake (EI), it is well known that self-reported food intake underestimates habitual food intake.2– 4 Due to this common bias in nutritional epidemiology, there is increasing scientific interest in underreporting and underreporters. Numerous investigations with young and middle-aged people have shown that underreporting is more likely to occur in overweight individuals, especially in females and people with low educational levels.4 – 8 Only few studies have been done in older individuals with regard to underreporting and findings are inconsistent. For example, in some investigations9,10 with elderly people, underreporting was related to a heavy body weight and large body mass index (BMI), whereas in other reports11,12 no relation between underreporting and body weight or BMI could be detected. Therefore, within the scope of the longitudinal study on nutrition and health status in an aging population in Giessen, Germany (GISELA), elderly subjects with low reported relative EI should be identified and characterized. Food intake was measured by means of a 3-d estimated dietary record developed especially for the GISELA-study.13 To identify underreporting, reported EI was ex-
Correspondence to: Monika Neuha¨user-Berthold, PhD, Institute of Nutritional Science, University of Giessen, Goethestr. 55, D-35390 Giessen, Germany. E-mail: [email protected] Nutrition 17:912–916, 2001 ©Elsevier Science Inc., 2001. Printed in the United States. All rights reserved.
pressed as a multiple of measured resting metabolic rate (RMR) and compared with the cutoff 2 derived from Goldberg et al.14 Underreporters were then compared with adequate reporters with regard to sociodemographic characteristics, body weight, BMI, fat mass, and food consumption. We also investigated the effects of underreporting on nutrient intake.
SUBJECTS AND METHODS Study Design The present investigation is part of the longitudinal study on an aging population in Giessen, Germany (GISELA), in which the nutrition and health status of free-living elderly people have been investigated at yearly intervals since 1994. Within the scope of that study, anthropometric data, body composition, RMR, and various biochemical parameters in blood and food intake, and corresponding energy and nutrient intakes of the study participants were examined. All measurements take place annually at the Institute of Nutritional Science in Giessen between June and November, from 6:00 to 10:00 AM after an 8- to 12-h overnight fast. The study protocol was approved by the Ethical Committee of the Faculty of Medicine at the Justus-Liebig University Giessen, and written informed consent was obtained from each participant. Subjects The subjects were recruited by physicians, notices, senior citizens’ meetings, advertisements in local newspapers, and by recruitment of friends through subjects who already were participants. Subjects had to be at least 60 y old, physically mobile, and living near Giessen on a long-term basis. During the first 4 y of the GISELA study (1994 to 1997), 320 women and 133 men participated in the 0899-9007/01/$20.00 PII S0899-9007(01)00664-5
Nutrition Volume 17, Number 11/12, 2001
Underreporting of Energy Intake in an Elderly German Population
investigations. The present report includes cross-sectional data from the first examination of 238 females and 105 males, with complete data on anthropometric measurements, body composition, RMR, and energy and nutrient intakes.
TABLE I. AGE, ANTHROPOMETRIC DATA, EI, AND RMR OF THE SUBJECTS*
3-d Estimated Dietary Record To determine food intake, a 3-d estimated dietary record was developed especially for the GISELA-study. The development and validity of the dietary record have been described in detail elsewhere.13 The dietary record consists of 146 food items, subdivided into 16 food groups. For every food item, typical household measures (e.g., slice, cup, spoon) and the appropriate weights were given; with this information, the subjects were supposed to estimate the amount of their food consumption. Further, participants could note any consumed food they were unable to classify under the heading “others.” Subjects were instructed to record their complete food intakes directly after consumption over 3 consecutive days in their dietary records, starting on a Sunday. To analyze the 3-d dietary record, foods recorded under the heading “others” were classified among given food items according to their energy and nutrient intake. Energy and nutrient contents of the food items then were calculated with the nutrient calculation program CALORA version II (Richard Thein, Wettenberg, Germany) and the Federal Nutrient Data Base version II.2 (Bundesgesundheitsamt, Berlin, Germany). Cutoff Limits to Identify Underreporting To identify underreporting, EI was expressed as a multiple of measured RMR and compared with the cutoff 2 derived from Goldberg et al.14 Because this cutoff limit differs according to sample size and number of days recorded by the dietary record, the cutoff value was calculated especially for the GISELA-study as follows: cutoff 2 ⫽ PAL ⫻ exp(SDmin ⫻ [共S/100兲/n1/ 2]) where PAL is the assumed average physical activity level of the population studied (1.55), SDmin is ⫺2 for 95% confidence limits, n is the number of subjects, and S is the overall coefficient of variation for physical activity level after considering the variability in EI and RMR. S was calculated from the equation: S ⫽ (CVIw2/k ⫹ CVB2 ⫹ CVP2)1/2 where CVIw is the within-subject variation in EI (23%), k is the number of recording days, CVB is the variation in repeated RMR measurement (2.5%), and CVP is the between-subject variation in physical activity level. In the present study, calculated cutoff values were 1.51 for females and 1.50 for males. To classify individuals (sample size ⫽ 1), we used a cutoff value of 1.073. Participants with an EI:RMR ratio below 1.073 were considered underreporters and those with a ratio of at least 1.073 were classified as adequate reporters. Resting Metabolic Rate RMR was determined with an open-circuit indirect calorimeter (Deltatrac MBM-100, Hoyer, Bremen, Germany). Oxygen uptake and carbon dioxide production were measured for 35 min at 1-min intervals by respiratory gas analysis using a ventilated-hood system, with the subjects in a supine position and completely at rest in a thermoneutral environment. Calibrations of the gas analyzer were performed immediately before every measurement. Each participant was allowed to become acclimated before the ventilated hood was placed over his or her head, and measurements were started. Data collected during the initial 10 min were discarded. RMR was calculated by using the equation of Weir.15
913
Age (y) Body height (cm) Body weight (kg) BMI (kg/m2) Reported EI (kJ/d) Measured RMR (kJ/d) EI:RMR Respiratory quotient
Females (n ⫽ 238)
Males (n ⫽ 105)
67.5 ⫾ 5.5 160.5 ⫾ 5.5 69.0 ⫾ 11.8 26.8 ⫾ 4.3 8311 ⫾ 2186 5212 ⫾ 659 1.62 ⫾ 0.46 0.84 ⫾ 0.07
67.0 ⫾ 4.8 173.3 ⫾ 6.4 80.0 ⫾ 10.4 26.8 ⫾ 3.2 9705 ⫾ 2579 6434 ⫾ 836 1.53 ⫾ 0.46 0.85 ⫾ 0.06
* Values are presented as mean ⫾ standard deviation. BMI, body mass index; EI, energy intake; RMR, resting metabolic rate.
Because measurements were made on an outpatient basis, inpatient RMR was calculated according to the method of Berke et al.16 by multiplying measured RMR by 0.93. Anthropometric Data and Body Composition Body weight was measured on a calibrated digital scale (Seca, Vogel & Halke, Hamburg, Germany) to the nearest 0.1 kg after shoes, coats, and sweaters had been removed; 0.5 to 1.0 kg for remaining clothes were subtracted from the measured weight. Body height without shoes was determined to the nearest 0.005 m with a height measurement device integrated into the scale. Body height was measured with the subject’s back and heels against the upright bar of the height scale, with the subject looking straight ahead and standing erect, with assistance when necessary. Body composition was investigated with the use of bioelectrical impedance (Akern-RJL BIA 101/S, Data Input, Frankfurt, Germany) in a supine position according to the manufacturers‘ instruction. Fat-free mass and fat mass were calculated by applying the equation suggested by Deurenberg et al.17 Non-Dietary Data Non-dietary data, such as age and educational level, were obtained from the subjects by questionnaire. Statistical Analysis Statistical analyses were done with the SPSS/PC Statistical Package version 6.1.3 (SPSS Inc, Chicago, IL, USA). Results are presented as mean and standard deviation. Data were checked with regard to normal distribution with the Kolmogorow–Smirnow test. Because not all variables were normally distributed, the unpaired Mann–Whitney rank sum test (two-tailed) was used to evaluate differences between underreporters and adequate reporters. Results were considered statistically significant at P ⬍ 0.05.
RESULTS Characteristics of the subjects are listed in Table I. The age range of the study group was 60 to 85 y; most of the females (69.7%) and males (77.1%) were between 60 and 70 y old. The mean BMI of females and males was 26.8 kg/m2. According to the International Obesity Task Force18 42.4% of the females and 55.2% of the males were overweight (BMI ⫽ 25.0 –29.9 kg/m2) and 18.9% of the females and 13.3% of the males were obese (BMI ⱖ 30 kg/m2).
914
Lu¨ hrmann et al.
Nutrition Volume 17, Number 11/12, 2001 TABLE II.
AGE, ANTHROPOMETRIC DATA, AND BODY COMPOSITION OF UNDERREPORTERS AND ADEQUATE REPORTERS* Females
Age (y) Body height (cm) Body weight (kg) BMI (kg/m2) Fat-free mass (kg) Fat mass (kg)
Males
Underreporters (n ⫽ 18)
Adequate reporters (n ⫽ 220)
Underreporters (n ⫽ 17)
Adequate reporters (n ⫽ 88)
65.9 ⫾ 5.4 160.9 ⫾ 5.4 75.2 ⫾ 13.4 29.0 ⫾ 4.8 39.5 ⫾ 6.3 35.6 ⫾ 7.8
67.6 ⫾ 5.5 160.5 ⫾ 5.6 68.5 ⫾ 11.6† 26.6 ⫾ 4.2† 37.8 ⫾ 5.4 30.6 ⫾ 6.9‡
66.5 ⫾ 4.8 175.4 ⫾ 6.8 88.1 ⫾ 11.7 28.7 ⫾ 3.9 57.9 ⫾ 6.2 30.3 ⫾ 7.0
67.1 ⫾ 4.8 172.9 ⫾ 6.3 78.9 ⫾ 9.4‡ 26.4 ⫾ 2.9† 53.2 ⫾ 5.4† 25.7 ⫾ 5.4‡
* Values are presented as mean ⫾ standard deviation. † P ⬍ 0.05, versus underreporters. ‡ P ⬍ 0.01, versus underreporters. BMI, body mass index.
Most of the participants in our group had an intermediate school qualification (females: 85.1%, males: 68.0%) and only a few subjects (females: 0.9%, males: 1.9%) had no school qualification; 13.9% of females and 30.1% of males had university entrance qualification. EI:RMR ratios were 1.62 in females and 1.53 in males. According to the cutoff 2 derived from Goldberg et al.14 7.6% of the females (n ⫽ 18) and 16.2% (n ⫽ 17) of the males were underreporters. The level of education was significantly lower in underreporters than in adequate reporters (P ⬍ 0.05, 2). None of the female underreporters and 18.8% of the male underreporters had a university entrance qualification in contrast to 15.1% and 32.1%, respectively, of the adequate reporters. Underreporters of both sexes had significantly greater body weight, BMI, and fat mass that the adequate reporters (Table II). Compared with adequate reporters, underreporters stated significantly more often that they wanted to lose weight (females: 88.9% versus 52.8%, males: 66.7% versus 42.5%). In comparison with the adequate reporters, the underreporters recorded significantly lower consumption of all the food groups listed in Table III, with the exception of meat and fish in females and alcoholic drinks in both sexes. Therefore, in both sexes calculated nutrient intakes were
significantly lower in underreporters than in adequate reporters, with the exception of -carotene in males (Table IV). Expressed as a percentage of EI, carbohydrate and fat intakes in both sexes and protein intake in females did not differ significantly between groups. With regard to vitamin and mineral densities, there were no significant differences between underreporters and adequate reporters.
DISCUSSION The aim of the present investigation was to identify and characterize underreporters by means of a newly developed 3-d estimated dietary record in a sample of older German women and men. The best biological marker for validation of EI and thus for identifying underreporting is the double-labeled water method.14,19 However, because this method is very expensive and time consuming, it is unsuitable for large population studies. As proposed by Goldberg et al.,14 the ratio between the protocol EI and RMR measured by indirect calorimetry served as a marker for reported EI in our study. This ratio correlates positively with the ratio of EI to energy
TABLE III. FOOD CONSUMPTION (G/D) OF UNDERREPORTERS AND ADEQUATE REPORTERS* Females
Bread, buns, and cakes Milk and dairy products Meat and fish Fruits and vegetables Potatoes, rice, pasta, and muesli Sugar and sweets Spread and cooking fat Alcoholic drinks
Males
Underreporters (n ⫽ 18)
Adequate reporters (n ⫽ 220)
Underreporters (n ⫽ 17)
Adequate reporters (n ⫽ 88)
106 ⫾ 43 125 ⫾ 153 115 ⫾ 46 307 ⫾ 130 112 ⫾ 67 32.6 ⫾ 50.0 9.1 ⫾ 5.1 99 ⫾ 121
208 ⫾ 80§ 271 ⫾ 183§ 138 ⫾ 69 435 ⫾ 209‡ 153 ⫾ 68‡ 46.4 ⫾ 39.9‡ 15.1 ⫾ 10.5‡ 82 ⫾ 139
179 ⫾ 79 176 ⫾ 214 114 ⫾ 91 300 ⫾ 166 120 ⫾ 53 23.1 ⫾ 25.9 10.2 ⫾ 6.8 230 ⫾ 287
267 ⫾ 98‡ 252 ⫾ 179† 177 ⫾ 83‡ 421 ⫾ 236† 177 ⫾ 90‡ 53.5 ⫾ 50.6‡ 15.0 ⫾ 9.7† 214 ⫾ 269
* Values are presented as mean ⫾ standard deviation. † P ⬍ 0.05, versus underreporters. ‡ P ⬍ 0.01, versus underreporters. § P ⬍ 0.001, versus underreporters.
Nutrition Volume 17, Number 11/12, 2001
Underreporting of Energy Intake in an Elderly German Population
915
TABLE IV. DAILY NUTRIENT INTAKES OF UNDERREPORTERS AND ADEQUATE REPORTERS* Females
Carbohydrates (g) Carbohydrates (% energy) Fat (g) Fat (% energy) Protein (g) Protein (% energy) Dietary fiber (g) Dietary fiber (g/MJ) Cholesterol (mg) Cholesterol (mg/MJ) Vitamin A† (mg) Vitamin A† (mg/MJ) Vitamin E (mg) Vitamin E (mg/MJ) Vitamin C (mg) Vitamin C (mg/MJ) -Carotene (mg) -Carotene (mg/MJ) Calcium (mg) Calcium (mg/MJ) Iron (mg) Iron (mg/MJ)
Males
Underreporters (n ⫽ 18)
Adequate reporters (n ⫽ 220)
Underreporters (n ⫽ 17)
Adequate reporters (n ⫽ 88)
133 ⫾ 37 44.9 ⫾ 7.1 45.7 ⫾ 12.3 35.1 ⫾ 6.1 53.3 ⫾ 16.1 18.0 ⫾ 3.6 14.8 ⫾ 5.1 2.96 ⫾ 1.06 199 ⫾ 73 39.3 ⫾ 14.9 1.13 ⫾ 1.92 0.26 ⫾ 0.52 6.8 ⫾ 2.5 1.38 ⫾ 0.52 77.5 ⫾ 33.3 16.1 ⫾ 8.2 2.33 ⫾ 1.44 0.49 ⫾ 0.39 674 ⫾ 271 138 ⫾ 66 7.8 ⫾ 1.9 1.55 ⫾ 0.34
238 ⫾ 60㛳 47.9 ⫾ 5.6 77.9 ⫾ 26.1㛳 35.2 ⫾ 5.5 83.6 ⫾ 21.4㛳 16.9 ⫾ 3.1 24.2 ⫾ 7.8㛳 2.85 ⫾ 0.68 308 ⫾ 120㛳 35.9 ⫾ 11.1 1.29 ⫾ 1.55㛳 0.15 ⫾ 0.15 12.9 ⫾ 4.8㛳 1.50 ⫾ 0.38 117.2 ⫾ 54.2㛳 13.7 ⫾ 5.5 3.24 ⫾ 2.10‡ 0.38 ⫾ 0.25 1010 ⫾ 331㛳 120 ⫾ 35 12.6 ⫾ 3.2㛳 1.48 ⫾ 0.23
178 ⫾ 52 45.6 ⫾ 10.5 56.7 ⫾ 16.3 33.4 ⫾ 8.0 70.5 ⫾ 20.8 18.0 ⫾ 3.7 20.4 ⫾ 6.4 3.12 ⫾ 0.99 230 ⫾ 88 35.4 ⫾ 15.0 0.95 ⫾ 0.36 0.14 ⫾ 0.06 9.7 ⫾ 3.0 1.49 ⫾ 0.51 83.8 ⫾ 48.1 12.6 ⫾ 6.4 2.70 ⫾ 1.69 0.42 ⫾ 0.27 739 ⫾ 339 111 ⫾ 43 10.8 ⫾ 2.9 1.63 ⫾ 0.35
285 ⫾ 71㛳 47.7 ⫾ 6.0 90.9 ⫾ 27.7㛳 34.4 ⫾ 5.5 96.3 ⫾ 26.6㛳 16.2 ⫾ 3.0‡ 27.7 ⫾ 9.5§ 2.70 ⫾ 0.73 357 ⫾ 163§ 34.3 ⫾ 11.3 1.53 ⫾ 1.32§ 0.15 ⫾ 0.11 13.5 ⫾ 4.5§ 1.31 ⫾ 0.33 120.5 ⫾ 60.6§ 11.7 ⫾ 5.0 3.86 ⫾ 4.20 0.36 ⫾ 0.30 997 ⫾ 361§ 98 ⫾ 32 15.2 ⫾ 4.0㛳 1.48 ⫾ 0.25
* Values are presented as mean ⫾ standard deviation. † Retinol equivalents. ‡ P ⬍ 0.05, versus underreporters. § P ⬍ 0.01, versus underreporters. 㛳 P ⬍ 0.001, versus underreporters.
expenditure as measured by the double-labeled water method. Therefore, it is a useful indicator for the analysis of underreporting in epidemiologic studies.20 To test whether reported EI was a plausible measure of food consumption during the current measurement period, the EI:RMR-ratio was compared with the cutoff 2 derived from Goldberg et al.14 This cutoff is based on the assumed average physical activity level of the study group and makes allowances for the number of subjects and recorded days and for the errors associated with variations in EI, RMR measurements, and physical activity. Therefore, this cutoff limit is studyspecific and must be calculated for each investigation. Our results showed that in both sexes the mean EI:RMR ratio was above the cutoff limits calculated for females (1.51) and males (1.50). These values for mean EI:RMR ratio are in accordance with findings from studies measuring total energy expenditure by double-labeled water and RMR by indirect calorimetry in elderly subjects.21–23 Thus, results indicate that the 3-d estimated dietary record is suitable to assess EI validly on the group level. Nevertheless, on the individual level in our sample, 7.6% of the females and 16.2% of the males were classified as underreporters. In general, the comparison of different investigations with regard to the proportion of underreporters is limited because the study-specific Goldberg cutoff limits were not used in all studies. Further, the prevalence of underreporting depends on the validity of the dietary assessment method applied in the various studies. However, results indicate that, in comparison with other investigations, the proportion of underreporters in our study group was low, especially in women. Further, the male subjects were more likely than female subjects to underreport their usual EI, in contrast
to other studies. For example, in the third National Health and Nutrition Examination Survey, 28% of the adult women and 18% of the adult men were classified as underreporters according to the cutoff limit of 0.9 for the EI:RMR ratio.5 Price et al.,6 who used a cutoff limit of 1.1, determined that 23% of the females and 19% of the males, age 43 y, underestimated their usual EI. Pryer et al.24 found that 46% of the women and 29% of the men, age 16 to 64 y, showed an EI:RMR ratio below 1.2. Rothenberg et al.9 classified 28% of the elderly women and 22% of the elderly men as underreporters with a cutoff limit of 1.2. An explanation for the low proportion of subjects underreporting their EI in our study might be that the participants were self-selected, highly motivated, and compliant volunteers. The relatively high level of education in our study group compared with the average elderly German population might be another explanation. A reason for the low proportion of female underreporters might be that the older females are more motivated and willing to record their dietary intakes than older males in comparison with younger adults because they are more interested in their nutrition. Our finding that underreporters had a lower level of education than adequate reporters is consistent with the results from previous studies with young and middle-aged adults.5,6 In contrast, Tomoyasu et al.10 did not find such an association in their investigation in elderly subjects. The greater body weight, BMI, and fat mass and the more frequent desire for weight reduction of underreporters versus adequate reporters are well known from numerous studies with younger individuals.4,5,7,8,25 In the few investigations available with older subjects, results are inconsistent. Rothenberg et al.9
916
Lu¨ hrmann et al.
showed a heavier body weight in elderly women and men who underreported their EI in comparison with adequate reporters and a positive correlation between BMI and underreporting. In accordance with those results, Tomoyasu et al.10 associated an increase in BMI with an increase in the magnitude of energy underreporting in older people. In contrast, other reports11,12 found no relation between underreporting and body weight or BMI in the elderly. With regard to fat mass, Tomoyasu et al.10 found a positive correlation between fat mass and the magnitude of underreporting, whereas Visser et al.12 did not. Johnson et al.11 negatively correlated fat mass with underreporting in older females but not in males. The underlying reasons concerning the greater body weight, fat mass, and BMI of the underreporters are still unknown, although several hypotheses have been discussed in the literature.6,14,26 Underreporters might consciously reduce their typical level of food intake because they want to pretend they eat less or they might use recording days for losing weight. Another explanation could be that persons with high BMI are more conscious of body shape and nutrition, resulting in eating patterns oscillating between restrained and unlimited periods. Those subjects probably restrain and consequently underreport their food consumption unconsciously if they have to record their food intake. In a recent investigation with highly motivated, lean women, underreporting was attributed to undereating because body weight declined significantly over the recording week.27 Some investigators in studies of different age groups have found that underreporters specifically exclude energy-rich, micronutrient-poor foods such as sugar, sweets, spreads, and cooking fat from their recorded diets. This results in a lower fat intake expressed as a percentage of EI and in a higher vitamin and mineral density of the diet.6,9,28 In contrast to those observations, our results indicated that underreporting concerns all foods consumed. Therefore, underreporters showed a lower absolute intake of most nutrients and nearly the same nutrient density of the diet than adequate reporters. The observation that alcohol consumption was not affected by underreporting confirms results from other investigations.9,27 An explanation might be that alcoholic drinks were not perceived as foods and thus neither consciously nor unconsciously undereaten or underreported, respectively. In summary, our observations confirm the findings of previous studies with younger individuals and those of two previous investigations9,10 with elderly subjects, i.e., that volunteers with a low educational level, heavy body weight, and large BMI and fat mass underreport their EI. Whereas absolute nutrient intake was strongly affected by underreporting, nutrient density did not seem to be affected in our study. The question of how to deal with underreporters in nutritional epidemiology cannot be answered at present. In general, however, it is not possible to exclude underreporters from further investigations because it has been shown that underreporting occurs not only at the lower end of EI but also throughout all levels of EI and energy expenditure.29 Therefore, it is important to characterize the underreporters and keep those characteristics in mind when interpreting the association between dietary data and health outcomes.
Nutrition Volume 17, Number 11/12, 2001
4. 5.
6. 7.
8.
9.
10. 11. 12.
13.
14.
15. 16. 17.
18.
19. 20.
21. 22. 23. 24.
25.
26.
REFERENCES 27. 1. Kinsella KG. Changes in life expectancy 1900 –1990. Am J Clin Nutr 1992;55: 1196S 2. Black AE, Coward WA, Cole TJ, Prentice AM. Human energy expenditure in affluent societies: an analysis of 574 doubly-labelled water measurements. Eur J Clin Nutr 1996;50:72 3. Lichtman SW, Pisarska K, Raynes Berman E, et al. Discrepancy between
28.
29.
self-reported and actual caloric intake and exercise in obese subjects. N Engl J Med 1992;327:1893 Prentice AM, Black AE, Coward WA, et al. High levels of energy expenditure in obese women. BMJ 1986;292:983 Briefel RR, Sempos CT, McDowell MA, Chien S, Alaimo C. Dietary methods research in the third National Health and Nutrition Examination Survey: underreporting of energy intake. Am J Clin Nutr 1997;65(suppl 4):S1203 Price GM, Paul AA, Cole TJ, Wadsworth MEJ. Characteristics of the low-energy reporters in a longitudinal national dietary survey. Br J Nutr 1997;77:833 Stallone DD, Brunner EJ, Bingham SA, Marmot MG. Dietary assessment in Whitehall II: the influence of reporting bias on apparent socioeconomic variation in nutrient intakes. Eur J Clin Nutr 1997;51:815 Voss S, Kroke A, Klipstein-Grobusch K, Boeing H. Is macronutrient composition of dietary intake data affected by underreporting? Results from the EPICPotsdam study. Eur J Clin Nutr 1998;52:119 Rothenberg E, Bosaeus I, Steen B. Evaluation of energy intake estimated by a diet history in three free-living 70 year old populations in Gothenburg, Sweden. Eur J Clin Nutr 1997;51:60 Tomoyasu NJ, Toth MJ, Poehlmann ET. Misreporting of total energy intake in older men and women. J Am Geriatr Soc 1999;47:710 Johnson RK, Goran MI, Poehlman ET. Correlates of over- and underreporting of energy intake in healthy older men and women. Am J Clin Nutr 1994;59:1286 Visser M, de Groot LCPGM, Deurenberg P, van Staveren WA. Validation of dietary history method in a group of elderly women using measurements of total energy expenditure. Br J Nutr 1995;74:775 Lu¨ hrmann P, Herbert B, Gaster C, Neuha¨ user-Berthold M. Validation of a self-administered 3-day estimated dietary record for use in the elderly. Eur J Nutr 1999;38:235 Goldberg GR, Black AE, Jebb SA, et al. Critical evaluation of energy intake data using fundamental principles of energy physiology: 1. Derivation of cut-off limits to identify under-recording. Eur J Clin Nutr 1991;45:569 Weir JB de V. New methods for calculating metabolic rate with special reference to protein metabolism. J Physiol 1949;109:1 Berke EM, Gardner AW, Goran MI, Poehlman ET. Resting metabolic rate and the influence of the pretesting environment. Am J Clin Nutr 1992;55:626 Deurenberg P, van der Kooy K, Leenen R, Weststrate JA, Seidell JC. Sex and age specific prediction formulas for estimating body composition from bioelectrical impedance: a cross-validation study. Int J Obes 1991;15:17 International Obesity Task Force, World Health Organization. Obesity: preventing and managing the global epidemic. Geneva: World Health Organization, 1998 Speakman JR. The history and theory of the doubly labeled water technique. Am J Clin Nutr 1998;68:932 Black AE, Bingham SA, Johansson G, Coward WA. Validation of dietary intakes of protein and energy against 24 hour urinary N and DLW energy expenditure in middle-aged women, retired men and post-obese subjects: comparisions with validation against presumed energy requirements. Eur J Clin Nutr 1997;51:405 Carpenter WH, Fonong T, Toth MJ, et al. Total daily energy expenditure in free-living older African-Americans and Caucasians. Am J Physiol 1998;274:E96 Goran MI, Poehlman ET. Total energy expenditure and energy requirements in healthy elderly persons. Metabolism 1992;41:744 Pannemans DLE, Westerterp KR. Energy expenditure, physical activity and basal metabolic rate of elderly subjects. Br J Nutr 1995;73:571 Pryer JA, Vrijheid M, Nichols R, Kiggins M, Elliott P. Who are the “low energy reporters” in the dietary and nutritional survey of British adults? Int J Epidemiol 1997;26:146 Heerstrass DW, Ocke´ MC, Bueno-de-Mesquita HB, Peeters PHM, Seidell JC. Underreporting of energy, protein and potassium intake in relation to body mass index. Int J Epidemiol 1998;27:186 Black AE, Goldberg GR, Jebb SA, et al. Critical evaluation of energy intake data using fundamental principles of energy physiology: 2. Evaluating the results of published surveys. Eur J Clin Nutr 1991;45:583 Goris AHC, Westerterp KR. Underreporting of habitual food intake is explained by undereating in highly motivated lean women. J Nutr 1999;129:878 Bingham SA, Cassidy A, Cole TJ, et al. Validation of weighed records and other methods of dietary assessment using the 24 h urine nitrogen technique and other biological markers. Br J Nutr 1995;73:531 Black AE. Under-reporting of energy intake at all levels of energy expenditure: evidence from doubly labelled water studies. Proc Nutr Soc 1997;56:121A
Underreporting of Energy Intake in an Elderly German Population Petra M. Lu¨hrmann, PhD, Birgit M. Herbert, PhD, and Monika Neuha¨user-Berthold, PhD From the Institute of Nutritional Science, University of Giessen, Giessen, Germany OBJECTIVES: Within the longitudinal study on nutrition and health status in an aging population in Giessen, Germany (GISELA), the underreporters of energy intake (EI) were identified and characterized. METHODS: EI was assessed in 238 female and 105 male participants of the GISELA study (age range ⫽ 60 – 89 y) by means of a 3-day estimated dietary record developed especially for this study. Resting metabolic rate (RMR) was measured by indirect calorimetry after an overnight fast. EI was expressed as a multiple of RMR and subjects with an EI:RMR ratio below 1.073 were classified as underreporters. RESULTS: Mean EI:RMR was 1.62 ⫾ 0.46 in females and 1.53 ⫾ 0.46 in males; 7.6% of the females and 16.2% of the males were identified as underreporters. They showed lower levels of education and significantly greater body weight, body mass index, and fat mass than the adequate reporters. Further, underreporters stated more often than adequate reporters that they want to lose weight. Except for -carotene in males, reported nutrient intakes were significantly lower in underreporters than in adequate reporters. Carbohydrate and fat intake in both sexes, protein intake in females calculated as a percentage of EI, and vitamin and mineral densities were not affected by underreporting. CONCLUSIONS: The results indicate that underreporting of EI is related to a low educational level and greater body weight, body mass index, and fat mass and affects all nutrients. These findings should be considered when the association between nutrition and health status is investigated in the elderly. Nutrition 2001;17:912–916. ©Elsevier Science Inc. 2001 KEY WORDS: elderly, dietary survey, 3-d estimated dietary record, ratio of energy intake to resting metabolic rate, underreporting
INTRODUCTION Since the beginning of the 19th century in Germany and other Western countries, the proportion of elderly people has continuously increased.1 Because of this demographic trend, more information concerning the relation between nutrition and health status of older age groups is needed. Dietary surveys are important for the assessment of nutrition status. However, from several studies comparing total energy expenditure measured by double-labeled water with reported energy intake (EI), it is well known that self-reported food intake underestimates habitual food intake.2– 4 Due to this common bias in nutritional epidemiology, there is increasing scientific interest in underreporting and underreporters. Numerous investigations with young and middle-aged people have shown that underreporting is more likely to occur in overweight individuals, especially in females and people with low educational levels.4 – 8 Only few studies have been done in older individuals with regard to underreporting and findings are inconsistent. For example, in some investigations9,10 with elderly people, underreporting was related to a heavy body weight and large body mass index (BMI), whereas in other reports11,12 no relation between underreporting and body weight or BMI could be detected. Therefore, within the scope of the longitudinal study on nutrition and health status in an aging population in Giessen, Germany (GISELA), elderly subjects with low reported relative EI should be identified and characterized. Food intake was measured by means of a 3-d estimated dietary record developed especially for the GISELA-study.13 To identify underreporting, reported EI was ex-
Correspondence to: Monika Neuha¨user-Berthold, PhD, Institute of Nutritional Science, University of Giessen, Goethestr. 55, D-35390 Giessen, Germany. E-mail: [email protected] Nutrition 17:912–916, 2001 ©Elsevier Science Inc., 2001. Printed in the United States. All rights reserved.
pressed as a multiple of measured resting metabolic rate (RMR) and compared with the cutoff 2 derived from Goldberg et al.14 Underreporters were then compared with adequate reporters with regard to sociodemographic characteristics, body weight, BMI, fat mass, and food consumption. We also investigated the effects of underreporting on nutrient intake.
SUBJECTS AND METHODS Study Design The present investigation is part of the longitudinal study on an aging population in Giessen, Germany (GISELA), in which the nutrition and health status of free-living elderly people have been investigated at yearly intervals since 1994. Within the scope of that study, anthropometric data, body composition, RMR, and various biochemical parameters in blood and food intake, and corresponding energy and nutrient intakes of the study participants were examined. All measurements take place annually at the Institute of Nutritional Science in Giessen between June and November, from 6:00 to 10:00 AM after an 8- to 12-h overnight fast. The study protocol was approved by the Ethical Committee of the Faculty of Medicine at the Justus-Liebig University Giessen, and written informed consent was obtained from each participant. Subjects The subjects were recruited by physicians, notices, senior citizens’ meetings, advertisements in local newspapers, and by recruitment of friends through subjects who already were participants. Subjects had to be at least 60 y old, physically mobile, and living near Giessen on a long-term basis. During the first 4 y of the GISELA study (1994 to 1997), 320 women and 133 men participated in the 0899-9007/01/$20.00 PII S0899-9007(01)00664-5
Nutrition Volume 17, Number 11/12, 2001
Underreporting of Energy Intake in an Elderly German Population
investigations. The present report includes cross-sectional data from the first examination of 238 females and 105 males, with complete data on anthropometric measurements, body composition, RMR, and energy and nutrient intakes.
TABLE I. AGE, ANTHROPOMETRIC DATA, EI, AND RMR OF THE SUBJECTS*
3-d Estimated Dietary Record To determine food intake, a 3-d estimated dietary record was developed especially for the GISELA-study. The development and validity of the dietary record have been described in detail elsewhere.13 The dietary record consists of 146 food items, subdivided into 16 food groups. For every food item, typical household measures (e.g., slice, cup, spoon) and the appropriate weights were given; with this information, the subjects were supposed to estimate the amount of their food consumption. Further, participants could note any consumed food they were unable to classify under the heading “others.” Subjects were instructed to record their complete food intakes directly after consumption over 3 consecutive days in their dietary records, starting on a Sunday. To analyze the 3-d dietary record, foods recorded under the heading “others” were classified among given food items according to their energy and nutrient intake. Energy and nutrient contents of the food items then were calculated with the nutrient calculation program CALORA version II (Richard Thein, Wettenberg, Germany) and the Federal Nutrient Data Base version II.2 (Bundesgesundheitsamt, Berlin, Germany). Cutoff Limits to Identify Underreporting To identify underreporting, EI was expressed as a multiple of measured RMR and compared with the cutoff 2 derived from Goldberg et al.14 Because this cutoff limit differs according to sample size and number of days recorded by the dietary record, the cutoff value was calculated especially for the GISELA-study as follows: cutoff 2 ⫽ PAL ⫻ exp(SDmin ⫻ [共S/100兲/n1/ 2]) where PAL is the assumed average physical activity level of the population studied (1.55), SDmin is ⫺2 for 95% confidence limits, n is the number of subjects, and S is the overall coefficient of variation for physical activity level after considering the variability in EI and RMR. S was calculated from the equation: S ⫽ (CVIw2/k ⫹ CVB2 ⫹ CVP2)1/2 where CVIw is the within-subject variation in EI (23%), k is the number of recording days, CVB is the variation in repeated RMR measurement (2.5%), and CVP is the between-subject variation in physical activity level. In the present study, calculated cutoff values were 1.51 for females and 1.50 for males. To classify individuals (sample size ⫽ 1), we used a cutoff value of 1.073. Participants with an EI:RMR ratio below 1.073 were considered underreporters and those with a ratio of at least 1.073 were classified as adequate reporters. Resting Metabolic Rate RMR was determined with an open-circuit indirect calorimeter (Deltatrac MBM-100, Hoyer, Bremen, Germany). Oxygen uptake and carbon dioxide production were measured for 35 min at 1-min intervals by respiratory gas analysis using a ventilated-hood system, with the subjects in a supine position and completely at rest in a thermoneutral environment. Calibrations of the gas analyzer were performed immediately before every measurement. Each participant was allowed to become acclimated before the ventilated hood was placed over his or her head, and measurements were started. Data collected during the initial 10 min were discarded. RMR was calculated by using the equation of Weir.15
913
Age (y) Body height (cm) Body weight (kg) BMI (kg/m2) Reported EI (kJ/d) Measured RMR (kJ/d) EI:RMR Respiratory quotient
Females (n ⫽ 238)
Males (n ⫽ 105)
67.5 ⫾ 5.5 160.5 ⫾ 5.5 69.0 ⫾ 11.8 26.8 ⫾ 4.3 8311 ⫾ 2186 5212 ⫾ 659 1.62 ⫾ 0.46 0.84 ⫾ 0.07
67.0 ⫾ 4.8 173.3 ⫾ 6.4 80.0 ⫾ 10.4 26.8 ⫾ 3.2 9705 ⫾ 2579 6434 ⫾ 836 1.53 ⫾ 0.46 0.85 ⫾ 0.06
* Values are presented as mean ⫾ standard deviation. BMI, body mass index; EI, energy intake; RMR, resting metabolic rate.
Because measurements were made on an outpatient basis, inpatient RMR was calculated according to the method of Berke et al.16 by multiplying measured RMR by 0.93. Anthropometric Data and Body Composition Body weight was measured on a calibrated digital scale (Seca, Vogel & Halke, Hamburg, Germany) to the nearest 0.1 kg after shoes, coats, and sweaters had been removed; 0.5 to 1.0 kg for remaining clothes were subtracted from the measured weight. Body height without shoes was determined to the nearest 0.005 m with a height measurement device integrated into the scale. Body height was measured with the subject’s back and heels against the upright bar of the height scale, with the subject looking straight ahead and standing erect, with assistance when necessary. Body composition was investigated with the use of bioelectrical impedance (Akern-RJL BIA 101/S, Data Input, Frankfurt, Germany) in a supine position according to the manufacturers‘ instruction. Fat-free mass and fat mass were calculated by applying the equation suggested by Deurenberg et al.17 Non-Dietary Data Non-dietary data, such as age and educational level, were obtained from the subjects by questionnaire. Statistical Analysis Statistical analyses were done with the SPSS/PC Statistical Package version 6.1.3 (SPSS Inc, Chicago, IL, USA). Results are presented as mean and standard deviation. Data were checked with regard to normal distribution with the Kolmogorow–Smirnow test. Because not all variables were normally distributed, the unpaired Mann–Whitney rank sum test (two-tailed) was used to evaluate differences between underreporters and adequate reporters. Results were considered statistically significant at P ⬍ 0.05.
RESULTS Characteristics of the subjects are listed in Table I. The age range of the study group was 60 to 85 y; most of the females (69.7%) and males (77.1%) were between 60 and 70 y old. The mean BMI of females and males was 26.8 kg/m2. According to the International Obesity Task Force18 42.4% of the females and 55.2% of the males were overweight (BMI ⫽ 25.0 –29.9 kg/m2) and 18.9% of the females and 13.3% of the males were obese (BMI ⱖ 30 kg/m2).
914
Lu¨ hrmann et al.
Nutrition Volume 17, Number 11/12, 2001 TABLE II.
AGE, ANTHROPOMETRIC DATA, AND BODY COMPOSITION OF UNDERREPORTERS AND ADEQUATE REPORTERS* Females
Age (y) Body height (cm) Body weight (kg) BMI (kg/m2) Fat-free mass (kg) Fat mass (kg)
Males
Underreporters (n ⫽ 18)
Adequate reporters (n ⫽ 220)
Underreporters (n ⫽ 17)
Adequate reporters (n ⫽ 88)
65.9 ⫾ 5.4 160.9 ⫾ 5.4 75.2 ⫾ 13.4 29.0 ⫾ 4.8 39.5 ⫾ 6.3 35.6 ⫾ 7.8
67.6 ⫾ 5.5 160.5 ⫾ 5.6 68.5 ⫾ 11.6† 26.6 ⫾ 4.2† 37.8 ⫾ 5.4 30.6 ⫾ 6.9‡
66.5 ⫾ 4.8 175.4 ⫾ 6.8 88.1 ⫾ 11.7 28.7 ⫾ 3.9 57.9 ⫾ 6.2 30.3 ⫾ 7.0
67.1 ⫾ 4.8 172.9 ⫾ 6.3 78.9 ⫾ 9.4‡ 26.4 ⫾ 2.9† 53.2 ⫾ 5.4† 25.7 ⫾ 5.4‡
* Values are presented as mean ⫾ standard deviation. † P ⬍ 0.05, versus underreporters. ‡ P ⬍ 0.01, versus underreporters. BMI, body mass index.
Most of the participants in our group had an intermediate school qualification (females: 85.1%, males: 68.0%) and only a few subjects (females: 0.9%, males: 1.9%) had no school qualification; 13.9% of females and 30.1% of males had university entrance qualification. EI:RMR ratios were 1.62 in females and 1.53 in males. According to the cutoff 2 derived from Goldberg et al.14 7.6% of the females (n ⫽ 18) and 16.2% (n ⫽ 17) of the males were underreporters. The level of education was significantly lower in underreporters than in adequate reporters (P ⬍ 0.05, 2). None of the female underreporters and 18.8% of the male underreporters had a university entrance qualification in contrast to 15.1% and 32.1%, respectively, of the adequate reporters. Underreporters of both sexes had significantly greater body weight, BMI, and fat mass that the adequate reporters (Table II). Compared with adequate reporters, underreporters stated significantly more often that they wanted to lose weight (females: 88.9% versus 52.8%, males: 66.7% versus 42.5%). In comparison with the adequate reporters, the underreporters recorded significantly lower consumption of all the food groups listed in Table III, with the exception of meat and fish in females and alcoholic drinks in both sexes. Therefore, in both sexes calculated nutrient intakes were
significantly lower in underreporters than in adequate reporters, with the exception of -carotene in males (Table IV). Expressed as a percentage of EI, carbohydrate and fat intakes in both sexes and protein intake in females did not differ significantly between groups. With regard to vitamin and mineral densities, there were no significant differences between underreporters and adequate reporters.
DISCUSSION The aim of the present investigation was to identify and characterize underreporters by means of a newly developed 3-d estimated dietary record in a sample of older German women and men. The best biological marker for validation of EI and thus for identifying underreporting is the double-labeled water method.14,19 However, because this method is very expensive and time consuming, it is unsuitable for large population studies. As proposed by Goldberg et al.,14 the ratio between the protocol EI and RMR measured by indirect calorimetry served as a marker for reported EI in our study. This ratio correlates positively with the ratio of EI to energy
TABLE III. FOOD CONSUMPTION (G/D) OF UNDERREPORTERS AND ADEQUATE REPORTERS* Females
Bread, buns, and cakes Milk and dairy products Meat and fish Fruits and vegetables Potatoes, rice, pasta, and muesli Sugar and sweets Spread and cooking fat Alcoholic drinks
Males
Underreporters (n ⫽ 18)
Adequate reporters (n ⫽ 220)
Underreporters (n ⫽ 17)
Adequate reporters (n ⫽ 88)
106 ⫾ 43 125 ⫾ 153 115 ⫾ 46 307 ⫾ 130 112 ⫾ 67 32.6 ⫾ 50.0 9.1 ⫾ 5.1 99 ⫾ 121
208 ⫾ 80§ 271 ⫾ 183§ 138 ⫾ 69 435 ⫾ 209‡ 153 ⫾ 68‡ 46.4 ⫾ 39.9‡ 15.1 ⫾ 10.5‡ 82 ⫾ 139
179 ⫾ 79 176 ⫾ 214 114 ⫾ 91 300 ⫾ 166 120 ⫾ 53 23.1 ⫾ 25.9 10.2 ⫾ 6.8 230 ⫾ 287
267 ⫾ 98‡ 252 ⫾ 179† 177 ⫾ 83‡ 421 ⫾ 236† 177 ⫾ 90‡ 53.5 ⫾ 50.6‡ 15.0 ⫾ 9.7† 214 ⫾ 269
* Values are presented as mean ⫾ standard deviation. † P ⬍ 0.05, versus underreporters. ‡ P ⬍ 0.01, versus underreporters. § P ⬍ 0.001, versus underreporters.
Nutrition Volume 17, Number 11/12, 2001
Underreporting of Energy Intake in an Elderly German Population
915
TABLE IV. DAILY NUTRIENT INTAKES OF UNDERREPORTERS AND ADEQUATE REPORTERS* Females
Carbohydrates (g) Carbohydrates (% energy) Fat (g) Fat (% energy) Protein (g) Protein (% energy) Dietary fiber (g) Dietary fiber (g/MJ) Cholesterol (mg) Cholesterol (mg/MJ) Vitamin A† (mg) Vitamin A† (mg/MJ) Vitamin E (mg) Vitamin E (mg/MJ) Vitamin C (mg) Vitamin C (mg/MJ) -Carotene (mg) -Carotene (mg/MJ) Calcium (mg) Calcium (mg/MJ) Iron (mg) Iron (mg/MJ)
Males
Underreporters (n ⫽ 18)
Adequate reporters (n ⫽ 220)
Underreporters (n ⫽ 17)
Adequate reporters (n ⫽ 88)
133 ⫾ 37 44.9 ⫾ 7.1 45.7 ⫾ 12.3 35.1 ⫾ 6.1 53.3 ⫾ 16.1 18.0 ⫾ 3.6 14.8 ⫾ 5.1 2.96 ⫾ 1.06 199 ⫾ 73 39.3 ⫾ 14.9 1.13 ⫾ 1.92 0.26 ⫾ 0.52 6.8 ⫾ 2.5 1.38 ⫾ 0.52 77.5 ⫾ 33.3 16.1 ⫾ 8.2 2.33 ⫾ 1.44 0.49 ⫾ 0.39 674 ⫾ 271 138 ⫾ 66 7.8 ⫾ 1.9 1.55 ⫾ 0.34
238 ⫾ 60㛳 47.9 ⫾ 5.6 77.9 ⫾ 26.1㛳 35.2 ⫾ 5.5 83.6 ⫾ 21.4㛳 16.9 ⫾ 3.1 24.2 ⫾ 7.8㛳 2.85 ⫾ 0.68 308 ⫾ 120㛳 35.9 ⫾ 11.1 1.29 ⫾ 1.55㛳 0.15 ⫾ 0.15 12.9 ⫾ 4.8㛳 1.50 ⫾ 0.38 117.2 ⫾ 54.2㛳 13.7 ⫾ 5.5 3.24 ⫾ 2.10‡ 0.38 ⫾ 0.25 1010 ⫾ 331㛳 120 ⫾ 35 12.6 ⫾ 3.2㛳 1.48 ⫾ 0.23
178 ⫾ 52 45.6 ⫾ 10.5 56.7 ⫾ 16.3 33.4 ⫾ 8.0 70.5 ⫾ 20.8 18.0 ⫾ 3.7 20.4 ⫾ 6.4 3.12 ⫾ 0.99 230 ⫾ 88 35.4 ⫾ 15.0 0.95 ⫾ 0.36 0.14 ⫾ 0.06 9.7 ⫾ 3.0 1.49 ⫾ 0.51 83.8 ⫾ 48.1 12.6 ⫾ 6.4 2.70 ⫾ 1.69 0.42 ⫾ 0.27 739 ⫾ 339 111 ⫾ 43 10.8 ⫾ 2.9 1.63 ⫾ 0.35
285 ⫾ 71㛳 47.7 ⫾ 6.0 90.9 ⫾ 27.7㛳 34.4 ⫾ 5.5 96.3 ⫾ 26.6㛳 16.2 ⫾ 3.0‡ 27.7 ⫾ 9.5§ 2.70 ⫾ 0.73 357 ⫾ 163§ 34.3 ⫾ 11.3 1.53 ⫾ 1.32§ 0.15 ⫾ 0.11 13.5 ⫾ 4.5§ 1.31 ⫾ 0.33 120.5 ⫾ 60.6§ 11.7 ⫾ 5.0 3.86 ⫾ 4.20 0.36 ⫾ 0.30 997 ⫾ 361§ 98 ⫾ 32 15.2 ⫾ 4.0㛳 1.48 ⫾ 0.25
* Values are presented as mean ⫾ standard deviation. † Retinol equivalents. ‡ P ⬍ 0.05, versus underreporters. § P ⬍ 0.01, versus underreporters. 㛳 P ⬍ 0.001, versus underreporters.
expenditure as measured by the double-labeled water method. Therefore, it is a useful indicator for the analysis of underreporting in epidemiologic studies.20 To test whether reported EI was a plausible measure of food consumption during the current measurement period, the EI:RMR-ratio was compared with the cutoff 2 derived from Goldberg et al.14 This cutoff is based on the assumed average physical activity level of the study group and makes allowances for the number of subjects and recorded days and for the errors associated with variations in EI, RMR measurements, and physical activity. Therefore, this cutoff limit is studyspecific and must be calculated for each investigation. Our results showed that in both sexes the mean EI:RMR ratio was above the cutoff limits calculated for females (1.51) and males (1.50). These values for mean EI:RMR ratio are in accordance with findings from studies measuring total energy expenditure by double-labeled water and RMR by indirect calorimetry in elderly subjects.21–23 Thus, results indicate that the 3-d estimated dietary record is suitable to assess EI validly on the group level. Nevertheless, on the individual level in our sample, 7.6% of the females and 16.2% of the males were classified as underreporters. In general, the comparison of different investigations with regard to the proportion of underreporters is limited because the study-specific Goldberg cutoff limits were not used in all studies. Further, the prevalence of underreporting depends on the validity of the dietary assessment method applied in the various studies. However, results indicate that, in comparison with other investigations, the proportion of underreporters in our study group was low, especially in women. Further, the male subjects were more likely than female subjects to underreport their usual EI, in contrast
to other studies. For example, in the third National Health and Nutrition Examination Survey, 28% of the adult women and 18% of the adult men were classified as underreporters according to the cutoff limit of 0.9 for the EI:RMR ratio.5 Price et al.,6 who used a cutoff limit of 1.1, determined that 23% of the females and 19% of the males, age 43 y, underestimated their usual EI. Pryer et al.24 found that 46% of the women and 29% of the men, age 16 to 64 y, showed an EI:RMR ratio below 1.2. Rothenberg et al.9 classified 28% of the elderly women and 22% of the elderly men as underreporters with a cutoff limit of 1.2. An explanation for the low proportion of subjects underreporting their EI in our study might be that the participants were self-selected, highly motivated, and compliant volunteers. The relatively high level of education in our study group compared with the average elderly German population might be another explanation. A reason for the low proportion of female underreporters might be that the older females are more motivated and willing to record their dietary intakes than older males in comparison with younger adults because they are more interested in their nutrition. Our finding that underreporters had a lower level of education than adequate reporters is consistent with the results from previous studies with young and middle-aged adults.5,6 In contrast, Tomoyasu et al.10 did not find such an association in their investigation in elderly subjects. The greater body weight, BMI, and fat mass and the more frequent desire for weight reduction of underreporters versus adequate reporters are well known from numerous studies with younger individuals.4,5,7,8,25 In the few investigations available with older subjects, results are inconsistent. Rothenberg et al.9
916
Lu¨ hrmann et al.
showed a heavier body weight in elderly women and men who underreported their EI in comparison with adequate reporters and a positive correlation between BMI and underreporting. In accordance with those results, Tomoyasu et al.10 associated an increase in BMI with an increase in the magnitude of energy underreporting in older people. In contrast, other reports11,12 found no relation between underreporting and body weight or BMI in the elderly. With regard to fat mass, Tomoyasu et al.10 found a positive correlation between fat mass and the magnitude of underreporting, whereas Visser et al.12 did not. Johnson et al.11 negatively correlated fat mass with underreporting in older females but not in males. The underlying reasons concerning the greater body weight, fat mass, and BMI of the underreporters are still unknown, although several hypotheses have been discussed in the literature.6,14,26 Underreporters might consciously reduce their typical level of food intake because they want to pretend they eat less or they might use recording days for losing weight. Another explanation could be that persons with high BMI are more conscious of body shape and nutrition, resulting in eating patterns oscillating between restrained and unlimited periods. Those subjects probably restrain and consequently underreport their food consumption unconsciously if they have to record their food intake. In a recent investigation with highly motivated, lean women, underreporting was attributed to undereating because body weight declined significantly over the recording week.27 Some investigators in studies of different age groups have found that underreporters specifically exclude energy-rich, micronutrient-poor foods such as sugar, sweets, spreads, and cooking fat from their recorded diets. This results in a lower fat intake expressed as a percentage of EI and in a higher vitamin and mineral density of the diet.6,9,28 In contrast to those observations, our results indicated that underreporting concerns all foods consumed. Therefore, underreporters showed a lower absolute intake of most nutrients and nearly the same nutrient density of the diet than adequate reporters. The observation that alcohol consumption was not affected by underreporting confirms results from other investigations.9,27 An explanation might be that alcoholic drinks were not perceived as foods and thus neither consciously nor unconsciously undereaten or underreported, respectively. In summary, our observations confirm the findings of previous studies with younger individuals and those of two previous investigations9,10 with elderly subjects, i.e., that volunteers with a low educational level, heavy body weight, and large BMI and fat mass underreport their EI. Whereas absolute nutrient intake was strongly affected by underreporting, nutrient density did not seem to be affected in our study. The question of how to deal with underreporters in nutritional epidemiology cannot be answered at present. In general, however, it is not possible to exclude underreporters from further investigations because it has been shown that underreporting occurs not only at the lower end of EI but also throughout all levels of EI and energy expenditure.29 Therefore, it is important to characterize the underreporters and keep those characteristics in mind when interpreting the association between dietary data and health outcomes.
Nutrition Volume 17, Number 11/12, 2001
4. 5.
6. 7.
8.
9.
10. 11. 12.
13.
14.
15. 16. 17.
18.
19. 20.
21. 22. 23. 24.
25.
26.
REFERENCES 27. 1. Kinsella KG. Changes in life expectancy 1900 –1990. Am J Clin Nutr 1992;55: 1196S 2. Black AE, Coward WA, Cole TJ, Prentice AM. Human energy expenditure in affluent societies: an analysis of 574 doubly-labelled water measurements. Eur J Clin Nutr 1996;50:72 3. Lichtman SW, Pisarska K, Raynes Berman E, et al. Discrepancy between
28.
29.
self-reported and actual caloric intake and exercise in obese subjects. N Engl J Med 1992;327:1893 Prentice AM, Black AE, Coward WA, et al. High levels of energy expenditure in obese women. BMJ 1986;292:983 Briefel RR, Sempos CT, McDowell MA, Chien S, Alaimo C. Dietary methods research in the third National Health and Nutrition Examination Survey: underreporting of energy intake. Am J Clin Nutr 1997;65(suppl 4):S1203 Price GM, Paul AA, Cole TJ, Wadsworth MEJ. Characteristics of the low-energy reporters in a longitudinal national dietary survey. Br J Nutr 1997;77:833 Stallone DD, Brunner EJ, Bingham SA, Marmot MG. Dietary assessment in Whitehall II: the influence of reporting bias on apparent socioeconomic variation in nutrient intakes. Eur J Clin Nutr 1997;51:815 Voss S, Kroke A, Klipstein-Grobusch K, Boeing H. Is macronutrient composition of dietary intake data affected by underreporting? Results from the EPICPotsdam study. Eur J Clin Nutr 1998;52:119 Rothenberg E, Bosaeus I, Steen B. Evaluation of energy intake estimated by a diet history in three free-living 70 year old populations in Gothenburg, Sweden. Eur J Clin Nutr 1997;51:60 Tomoyasu NJ, Toth MJ, Poehlmann ET. Misreporting of total energy intake in older men and women. J Am Geriatr Soc 1999;47:710 Johnson RK, Goran MI, Poehlman ET. Correlates of over- and underreporting of energy intake in healthy older men and women. Am J Clin Nutr 1994;59:1286 Visser M, de Groot LCPGM, Deurenberg P, van Staveren WA. Validation of dietary history method in a group of elderly women using measurements of total energy expenditure. Br J Nutr 1995;74:775 Lu¨ hrmann P, Herbert B, Gaster C, Neuha¨ user-Berthold M. Validation of a self-administered 3-day estimated dietary record for use in the elderly. Eur J Nutr 1999;38:235 Goldberg GR, Black AE, Jebb SA, et al. Critical evaluation of energy intake data using fundamental principles of energy physiology: 1. Derivation of cut-off limits to identify under-recording. Eur J Clin Nutr 1991;45:569 Weir JB de V. New methods for calculating metabolic rate with special reference to protein metabolism. J Physiol 1949;109:1 Berke EM, Gardner AW, Goran MI, Poehlman ET. Resting metabolic rate and the influence of the pretesting environment. Am J Clin Nutr 1992;55:626 Deurenberg P, van der Kooy K, Leenen R, Weststrate JA, Seidell JC. Sex and age specific prediction formulas for estimating body composition from bioelectrical impedance: a cross-validation study. Int J Obes 1991;15:17 International Obesity Task Force, World Health Organization. Obesity: preventing and managing the global epidemic. Geneva: World Health Organization, 1998 Speakman JR. The history and theory of the doubly labeled water technique. Am J Clin Nutr 1998;68:932 Black AE, Bingham SA, Johansson G, Coward WA. Validation of dietary intakes of protein and energy against 24 hour urinary N and DLW energy expenditure in middle-aged women, retired men and post-obese subjects: comparisions with validation against presumed energy requirements. Eur J Clin Nutr 1997;51:405 Carpenter WH, Fonong T, Toth MJ, et al. Total daily energy expenditure in free-living older African-Americans and Caucasians. Am J Physiol 1998;274:E96 Goran MI, Poehlman ET. Total energy expenditure and energy requirements in healthy elderly persons. Metabolism 1992;41:744 Pannemans DLE, Westerterp KR. Energy expenditure, physical activity and basal metabolic rate of elderly subjects. Br J Nutr 1995;73:571 Pryer JA, Vrijheid M, Nichols R, Kiggins M, Elliott P. Who are the “low energy reporters” in the dietary and nutritional survey of British adults? Int J Epidemiol 1997;26:146 Heerstrass DW, Ocke´ MC, Bueno-de-Mesquita HB, Peeters PHM, Seidell JC. Underreporting of energy, protein and potassium intake in relation to body mass index. Int J Epidemiol 1998;27:186 Black AE, Goldberg GR, Jebb SA, et al. Critical evaluation of energy intake data using fundamental principles of energy physiology: 2. Evaluating the results of published surveys. Eur J Clin Nutr 1991;45:583 Goris AHC, Westerterp KR. Underreporting of habitual food intake is explained by undereating in highly motivated lean women. J Nutr 1999;129:878 Bingham SA, Cassidy A, Cole TJ, et al. Validation of weighed records and other methods of dietary assessment using the 24 h urine nitrogen technique and other biological markers. Br J Nutr 1995;73:531 Black AE. Under-reporting of energy intake at all levels of energy expenditure: evidence from doubly labelled water studies. Proc Nutr Soc 1997;56:121A