Sugar intake, soft drink consumption and body weight among British children: further analysis of National Diet and Nutrition Survey data with adjustment for under-reporting and physical activity.

We investigated associations between body mass index (BMI) and intake of non-milk extrinsic sugars (NMES) and caloric soft drinks using weighed 7-day food records, nutrient intakes, BMI measurements and 7-day physical activity (PA) diaries from the UK National Dietary and Nutritional Survey of Young People (n=1,294 aged 7-18 years). NMES and caloric soft drinks (excluding 100% fruit juice) were quantified by their contribution to energy intake. BMI z-scores were calculated from UK reference curves and International Obesity Task Force (IOTF) cut-off values were used to define overweight. The BMI z-score was weakly inversely correlated with percentage energy from NMES after adjustment for under-reporting and dieting (r=-0.06, P=0.03). The percentage of energy from soft drinks was not associated with the BMI z-score or PA. After excluding under-reporters and dieters, the heaviest children (top quintile: Q5 of BMI z-scores) consumed more total energy (+1,220 kJ/day) than those in the lowest quintile (Q1), but only 60 kJ (5%) was from soft drinks. In logistic regression (adjusted for age and gender, under-reporting, and dieting), overweight was positively associated with energy intake (MJ) (odds ratio [OR]=1.58, confidence interval [CI]=1.42-1.77) and sedentary activity (h) (OR=1.11, CI=1.01-1.23), and inversely associated with moderate/vigorous activity (h) (OR=0.71, CI=0.58-0.86). In the macronutrient model, high fat and protein intake (top tertile vs lowest tertile, g/day) were positively associated with overweight (OR>2.5, P<0.001) while starch had less impact (OR=1.60, CI=1.0-2.55, P<0.05). Top tertile intakes of caloric soft drinks were weakly associated with overweight (OR=1.39, CI=0.96-2.0, P=0.08), while other sources of NMES showed no association (OR=0.81, CI=0.52-1.27, P=0.4). Risk associated with caloric soft drinks appeared non-linear with an increase in odds only for very high consumers (top quintile, mean 870 kJ/day; OR=1.67, CI=1.04-2.66, P=0.03). These data are not consistent with any specific role for NMES or caloric soft drinks in obesity among British children and adolescents, but point instead to a general role of overeating and physical inactivity. Evidence of successful interventions is urgently needed but these must use reliable measurements of exposure (diet and PA) and outcome (BMI z-score, body fat, waist circumference) and have a sufficient timescale.