Methods and validity of dietary assessments in four Scandinavian populations

Traditional Self-Reported Dietary Instruments Are Prone to Inaccuracies and New Approaches Are Needed

Background: Diet is a modifiable behavior that influences an individual's health. Because of this, diet assessment is an important component of public health surveillance, evaluating response to community health interventions, and monitoring individual compliance to medical interventions. Diet assessments are usually performed using one of three basic methods: diet recall, diet diaries, or food frequency questionnaires. Although these three assessment instruments have displayed a strong agreement between themselves, when reported intake is compared with intake measured using quantitative nutrient biomarkers, investigators have identified systematic misreporting errors for all three of these self-reported dietary instruments.

Aims: This work aims to summarize the state of knowledge regarding misreporting and why it impedes diet–health research and to introduce advances in the collection and the treatment of dietary data.

Methods: This work reviews and summarizes published data on misreporting and the recent efforts to reduce such errors.

Results: The evidence demonstrates a strong and consistent systematic underreporting of energy intake (EIn) across adults and children studies. Underreporting of EIn has been found to increase with body mass index (BMI), and the differences between macronutrient reports indicate that not all foods are underreported equally. Protein is least underreported, but which specific foods are commonly underreported are not known.

Conclusions: Because energy underreporting varies as a function of BMI, self-reported EIn should not be used for the study of energy balance in the study of obesity. The between-individual variability in the underreporting of self-reported intake of energy and other nutrients attenuates diet–disease relationships. Recent efforts to correct for underreporting have reduced misreporting of diet outcomes, but improvements have been incremental in nature and more research is needed to validate and extend these efforts.

The influence of systemic inflammation, dietary intake and stage of disease on rate of weight loss in patients with gastro-oesophageal cancer

Although weight loss is often a dominant symptom in patients with upper gastrointestinal malignancy, there is a lack of objective evidence describing changes in nutritional status and potential associations between weight loss, food intake, markers of systemic inflammation and stage of disease in such patients. Two hundred and twenty patients diagnosed with gastric/oesophageal cancer were studied. Patients underwent nutritional assessment consisting of calculation of body mass index, measurement of weight loss, dysphagia scoring and estimation of dietary intake. Serum acute-phase protein concentrations were determined by enzyme-linked immunosorbent assay. In all, 182 (83%) patients had lost weight at diagnosis (median loss, 7% body weight). Weight loss was associated with poor performance status, advanced disease stage, dysphagia, reduced dietary intake and elevated serum C-reactive protein (CRP) concentrations. Multiple regression identified dietary intake (estimate of effect, 38%), serum CRP concentrations (estimate of effect, 34%) and stage of disease (estimate of effect, 28%) as independent variables in determining degree of weight loss. Mechanisms other than reduced dietary intake or mechanical obstruction by the tumour appear to be involved in the nutritional decline in patients with gastro-oesophageal malignancy. Recognition that systemic inflammation plays a role in nutritional depletion may inform the development of appropriate therapeutic strategies to ameliorate weight loss, making patients more tolerant of cancer-modifying treatments such as chemotherapy.

Definition of cancer cachexia: effect of weight loss, reduced food intake, and systemic inflammation on functional status and prognosis

BACKGROUND

Cancer cachexia is a multifactorial syndrome that is poorly defined.

OBJECTIVE

Our objective was to evaluate whether a 3-factor profile incorporating weight loss (> or = 10%), low food intake (< or = 1500 kcal/d), and systemic inflammation (C-reactive protein > or = 10 mg/L) might relate better to the adverse functional aspects of cachexia and to a patient's overall prognosis than will weight loss alone.

DESIGN

One hundred seventy weight-losing (> or = 5%) patients with advanced pancreatic cancer were screened for nutritional status, functional status, performance score, health status, and quality of life. Patients were followed for a minimum of 6 mo, and survival was noted. Patients were characterized by using the individual factors, > or = 2 factors, or all 3 factors.

RESULTS

Weight loss alone did not define a population that differed in functional aspects of self-reported quality of life or health status and differed only in objective factors of physical function. The 3-factor profile identified both reduced subjective and objective function. In the overall population, the 3 factors, > or = 2 factors, and individual profile factors (except weight loss) all carried adverse prognostic significance (P < 0.01). Subgroup analysis showed that the 3-factor profile carried adverse prognostic significance in localized (hazard ratio: 4.9; P < 0.001) but not in metastatic disease.

CONCLUSIONS

Weight loss alone does not identify the full effect of cachexia on physical function and is not a prognostic variable. The 3-factor profile (weight loss, reduced food intake, and systemic inflammation) identifies patients with both adverse function and prognosis. Shortened survival applies particularly to cachectic patients with localized disease, thereby reinforcing the need for early intervention.

The validity of self-reported energy intake as determined using the doubly labelled water technique

In the 1980s the development of the doubly labelled water (DLW) technique made it possible to determine the validity of dietary assessment methods using external, independent markers of intake in free-living populations. Since then, the accuracy of self-reported energy intake (EI) has been questioned on a number of occasions as under-reporting has been found to be prevalent in many different populations. This paper is a review of investigations using the DLW technique in conjunction with self-reported EI measures in groups including adults, children and adolescents, obese persons, athletes, military personnel and trekking explorers. In studies where a person other than the subject is responsible for recording dietary intake, such as parents of young children, EI generally corresponds to DLW determined energy expenditure. However, in instances where the subjects themselves report their intake, EI is generally under-reported when compared with energy expenditure. It was originally believed that this phenomenon of under-reporting was linked to increased adiposity and body size, however, it is now apparent that other factors, such as dietary restraint and socio-economic status, are also involved. This paper therefore aims to present a more comprehensive picture of under-reporting by tying in the findings of many DLW studies with other studies focusing particularly on the characteristics and mechanisms for under-reporting. Awareness of these characteristics and mechanisms will enable researchers to obtain more accurate self-reports of EI using all dietary recording techniques.

Validation of dietary intakes measured by diet history against 24 h urinary nitrogen excretion and energy expenditure measured by the doubly-labelled water method in middle-aged women

A diet history method for estimating energy and N intakes was validated against 24 h urinary N excretion and energy expenditure measured by the doubly-labelled water (DLW) method. Forty-eight women aged 50-65 years were studied over 1 year. Weighed diet records from 4 d and two 24 h urine collections, for measurement of urinary N excretion, were obtained in each of four seasons. At the end of the year, a diet history was obtained, BMR was measured by whole-body calorimetry, and, in sixteen women, total energy expenditure (EE) was measured by DLW. Energy intake (EI) and N intake (NI) were calculated using food tables. Using weighed records and diet history respectively mean NI were 11.21 (SD 2.09) g and 11.47 (SD 2.40) g (NS) and EI were 8.08 (SD 1.54) MJ and 8.20 (SD 1.86) MJ (NS). Mean urine N:NI and EI:BMR values indicated bias to under-reporting by weighed record and diet history techniques in some individuals, but there was no significant difference between these measures at the group level. The Pearson correlation coefficient (r) for urine N v. NI was 0.81 for the weighed record and 0.38 for the diet history. The correlation of EE v. EI was r 0.48 for weighed record and r 0.11 for diet history. In this study the diet history gave the same estimate of mean intake, but the weighed record appeared to perform better in ranking individuals.

Validation with biological markers for food intake of a dietary assessment method used by Swedish women with three different dietary preferences

OBJECTIVES

To validate a dietary assessment method, a 4-day food record together with a duplicate portion technique, with biological markers for food intake.

DESIGN

Four days of duplicate portions were collected in parallel with food recording. A 24-h urine sample and the faeces corresponding to the food intake (using a coloured marker) were collected. Completeness of urine and faeces collections was assessed using para-aminobenzoic acid (PABA) in urine and cadmium in faeces, respectively. Biomarkers of food intake (energy, protein, fibre, sodium, potassium, calcium) were measured in urine and faeces.

SETTING

Swedish west coast.

SUBJECTS

Non-smoking Swedish women, 20-50 years of age, consuming a mixed diet (n = 34), a mixed diet rich in shellfish (n = 17) or a vegetarian/high-fibre diet (n = 23).

RESULTS

The average ratio (food intake according to the dietary assessment methods/ biological marker) for protein, sodium, potassium and calcium was 0.86. This indicates an underestimation of the food intake by approximately 15%. The ratio of stated fibre intake to biological marker was 1.20 for the mixed diet and the vegetarian diet group, indicating an overestimation by approximately 20%.

CONCLUSIONS

The underestimation of the intake of protein, sodium, potassium and calcium by all three groups and the overestimation of the fibre intake by two groups indicate that underreporting is selective to certain nutrients and foods and to various groups of people. The two dependent dietary assessment methods were equally good in measuring protein intake, which indicates that the women recorded what they actually duplicated.

Relations between nutrient intake and serum lipid and apolipoprotein levels

OBJECTIVE

To study the relation between diet and serum lipid and apolipoprotein levels in a large, homogeneous group of Spanish children.

DESIGN

Survey.

SETTING

Eleven schools chosen at random in Madrid City and the surrounding area.

SUBJECTS AND METHODS

Subjects comprised 1682 children, 2 to 12 years of age. Dietary data were obtained with a 24-hour record performed by the child's main caregiver. For statistical analysis the sample was divided into tertiles on the basis of calorie-adjusted consumption of dietary components.

RESULTS

Children in the highest tertile of total fat consumption, compared with children in the lowest tertile, had significantly higher mean serum levels of total cholesterol (TC), low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), apolipoprotein A-I (Apo A) and apolipoprotein B-100 (Apo B). When compared with children in the lowest tertile, children in the highest tertile of saturated fat consumption had significantly higher mean levels of TC, LDL-C, and Apo B, and lower mean levels of HDL-C and Apo A. Children in the highest tertile of monounsaturated fat consumption, compared with children in the lowest tertile, had significantly higher mean levels of HDL-C and Apo A, and lower mean levels of TC, LDL-C, and Apo B.

CONCLUSIONS

Our findings suggest that diet composition strongly influences lipid profile in children and point out the importance of monounsaturated fatty acids as modulators of serum lipid and apolipoprotein levels.

The effects on resting energy expenditure of different tumor types

The presence of a malignant tumor is said to influence the resting energy expenditure (REE) of the host. This study assesses the hypothesis that different tumor types exert differing effects on REE. REE was measured using indirect calorimetry in 84 cancer patients. Fifty-one patients had colorectal cancer, 22 had gastric cancer, and 11 had non-small cell bronchial cancer. For each tumor type, REE correlated significantly with body weight and lean body mass (LBM). The slope of the regression line for the bronchial cancer patients was significantly different from the colorectal (P less than 0.02) and gastric (P less than 0.02) cancer patients when REE was related to LBM. When related to body weight, the slope of the line of the bronchial cancer patients was significantly different from that of the gastric cancer patients only (P less than 0.05). The bronchial cancer patients had a significantly higher REE than the colorectal (P less than 0.005) and gastric (P less than 0.02) cancer patients when REE was expressed in terms of body weight. However, when corrected for LBM, no significant differences in REE were found between the groups. The presence of hepatic metastases did not influence REE in any of the tumor groups. Differing relationships between REE and indices of body size have been detected in differing tumor types. Consequently, the use of heterogeneous cancer groups may be inappropriate in studies of REE and cancer.

Effects of hepatic metastases on resting energy expenditure in patients with colorectal cancer

There are frequent reports that cancer causes an increase in resting energy expenditure (REE), and the presence of hepatic metastases is said to increase REE further. The effect of tumour burden on REE has been measured in 24 patients before and following surgery for colorectal cancer. Fifteen patients had a 'curative' resection whereas nine patients were found to have hepatic metastases at the time of surgery. No significant differences in REE were found between the groups either before surgery or at follow-up, by which time all 15 curative resection patients were tumour free and the remaining nine had progression of their hepatic disease. It is concluded that neither surgical removal of the primary tumour nor progression of metastatic hepatic disease significantly alters REE in patients with colorectal cancer.

Dietary fibre consumption and its association with large bowel cancer in man

The hypothesis that lack of fibre in the diet is responsible for a variety of large bowel problems, including cancer, has stimulated much discussion and research over the past 15 years. However, the epidemiological examination of this hypothesis has been hampered by the absence of data on the fibre content of most of the world's foods. In studies in Britain and Scandinavia where consumption of the chemical fraction of dietary fibre, the non-starch polysaccharides, has been determined using accurate methods, significant negative association between colon cancer occurrence and NSP consumption have been shown. Fibre may therefore be protective to populations otherwise assumed to be at risk from a westernised type of diet. At present, methodological problems preclude the use of case-control studies in confirming or refuting these associations.

Diet and cancer: value of different types of epidemiological studies

Diet and nutrition are increasingly recognized as likely to be major determinants of cancer, notably cancers of the gastrointestinal tract, breast, endometrium, ovary, and prostate. Dietary factors may collectively account for a greater proportion of all cancers that occur in contemporary Western society than does any other category of environmental exposure (1). With the development of knowledge of the protective properties of certain components of food, links with diet have been suggested for other cancer sites (2). The epidemiological evidence for the association of diet and cancer is, however, not uniformly convincing; also, the likely biological pathways are not always clear. In this paper, we comment on some current hypotheses in this area and examine the best epidemiological methods to test them.

Epidemiology of dietary fibre and colorectal cancer: current status of the hypothesis

The hypothesis that lack of dietary fibre in the diet is responsible for a variety of large bowel problems, including cancer, has stimulated much discussion and research over the past 15 years. However, the epidemiological examination of this hypothesis has been hampered by the absence of data on the fibre content of most of the world's foods. In Scandinavia and Britain where the consumption of the major chemical fraction of dietary fibre, the non-starch polysaccharides, has been measured using accurate methods, significant negative associations have been shown with colon cancer occurrence. These studies suggest that non-starch polysaccharides may be protective in populations at otherwise high risk of colon cancer from an excess of meat and fat. However, methodological problems in the assessment of non-starch polysaccharide consumption in individuals preclude the use of case control studies in verifying these associations within a single homogeneous population.