The characteristics of underreporting women in the POWIRS II study

Abstract

It is now widely recognized that reported energy intakes in dietary surveys underestimate usual energy intake (Black et al., 1993). During this study specific subject characteristics contributing to under reporting and the possible association between the metabolic syndrome and under reporting were investigated. A multi-disciplinary cross-sectional case-control study was carried out with 115 apparently healthy Caucasian women. The POWIRS II (Profiles of Obese Women Suffering from the Insulin Resistance Syndrome) study was performed in the Metabolic Unit of the North-West University, Potchefstroom, South Africa. Here, dietary intake was measured using the quantitative food frequency questionnaire (QFFQ). The ratio of energy intake (El) to basal metabolic rate (BMR) was calculated from the reported El and BMR estimated using equations. Under reporters (URs) were identified using the Goldberg equation, which compares El with energy expenditure (EE), both expressed as multiples of the BMR. URs had reported El <1.27 x BMR, non-URs ≥1.27 x BMR. Total energy expenditure (TEE) was calculated (n=115) and measured in a subgroup (n=63) by using the Actical accelerometer. Physical activity was also divided into three groups by using the Yale physical activity questionnaire. Most of the women were overweight (45.8%) and 29.2% were obese but only 20 subjects presented with the metabolic syndrome. The prevalence of under reporting women with a food intake level (FIL; reported El divided by estimated BMR) below 1.27 was relatively low (21%). Over reporting (FIL >2.4) was also low (6.7%). The prevalence of metabolic syndrome did not differ between URs and non-URs (12.5% and 18.7%, respectively). Most subjects were younger than 35 years (mean=33.2), highly educated and received a high monthly income. The subjects fell mostly into the low activity category. URs had a statistically significantly (p<0.05) higher percentage El from fat but lower from carbohydrates (37.4% and 42.g0h, respectively) when compared to non-URs (33.5% and 47%; respectively). The mean intakes of most micronutrients were statistically significantly lower in the URs. However, expressed per energy unit, intakes of URs were significantly higher for potassium, selenium and vitamin E. When the food group intakes of URs and non-URs were compared no significant differences were found. Measured TEE was significantly higher (p<0.01) in URs and non-URs (1 3 036kJ and 13 805kJ; respectively) compared to calculated TEE (10 105.8kJ and 10 220.3kJ; respectively). Measured TEE was also significantly higher than reported El (p<0.01). No specific subject characteristics were related to under reporting. Considering the low reported El compared to the measured TEE, the cut off values to identify URs might have been too low.