Tuesday March 17 2015
Eating lots of sugar is associated with a higher BMI
"Obese people are 'in denial' about the amount of sugar they eat," the Mail Online reports. Researchers looking into the link between sugar consumption and obesity found a "huge gap" between overweight people's self-reported sugar consumption and the reality, according to the news story.
Researchers assessed the self-reported sugar consumption (based on food diaries) and sugar levels in urine samples in about 1,700 people in Norfolk. After three years, they had their body mass index (BMI) measured.
The researchers found those whose urine test suggested they actually consumed the most sugar were more likely to be overweight after three years compared with those who consumed the least. However, the opposite was true for self-reported sugar intake.
The specific role of sugar (rather than calorie intake as a whole) in obesity is unclear, and previous studies have had inconsistent results.
One limitation of this study is that the spot-check urinary sugar test may not be representative of sugar intake over the whole study period. Also, the results may be affected by factors not taken into account by the analyses.
Although the news story focuses on the suggestion that overweight people are "in denial" about what they eat, this study itself did not attempt to explain the discrepancy between diet diaries and urine sugar measurements.
Overall, the main conclusion of this study is that more objective measures, rather than subjective diet-based records, may help future studies to better disentangle the effects of sugar on outcomes such as being overweight.
Where did the story come from?
The study was carried out by researchers from the universities of Reading and Cambridge in the UK and Arizona State University in the US.
It was funded by the World Cancer Research Fund, Cancer Research UK, and the Medical Research Council.
The study was published in the peer-reviewed medical journal Public Health Nutrition. It is available on an open-access basis, so is available to download for free.
The Mail focuses on the suggestion that overweight people are "in denial" about what they eat. But this study did not assess why the discrepancies between diet diaries and urine sugar measurements exist. It also does not question some potential problems with the urine tests, which could undermine the results.
What kind of research was this?
This was a prospective cohort study, part of the European Prospective Investigation into Cancer and Nutrition (EPIC), a long-running investigation. It aimed to see whether people who ate more sugar were more likely to be overweight using two different ways of measuring sugar intake.
Observational studies assessing whether total sugar intake is linked to obesity have had conflicting findings. Such studies usually ask people to report what they eat using food frequency questionnaires or a food diary, and then use this information to calculate sugar intake.
However, there is concern that people under-report their food intake. Therefore, the researchers in this study used both food diaries and an objective measure (the level of sugar in urine) to assess sugar intake. They wanted to see if there was any difference in results with the two approaches.
The main limitation of observational studies such as this is that it is difficult to prove that a single factor, such as a particular type of food, directly causes an outcome such as being overweight. This is because other differences between people may be affecting the results.
However, it would not be ethical to expose people to potentially unhealthy diets in a long-term randomised controlled trial, so this type of observational study is the best practical way of assessing the link between diet and weight.
What did the research involve?
Researchers recruited adults aged 39 to 79 in Norfolk in the UK. They took measurements including their body mass index (BMI), lifestyle information, and tested their urine for sugar levels. Participants were also asked to record their diet over seven days.
Three years later, the participants were invited back and measured again for BMI and waist circumference. Researchers looked for links between people's sugar levels as shown in urine samples, the amount of sugar they reported eating based on their diet records, and whether they were overweight at this three-year assessment.
The entire EPIC study included more than 70,000 people, but researchers took a single urine sample from around 6,000 people as a "spot check" biomarker on sugar levels.
These single spot check samples measured recent sugar intake, and may be a less reliable measure of overall sugar intake than the more expensive and difficult test of collecting urine over a 24-hour period for analysis.
Almost 2,500 people did not come back for the second health check, and 1,367 people's urine tests were either not possible to analyse or the results were outside the standard range and so discarded.
This means only 1,734 of the original sample could be included in the final analysis. Because the people finally included were not randomly selected, it's possible that their results are not representative of all the people in the study.
The researchers ranked both the urine sugar results and sugar based on the dietary record results into five groups, from lowest to highest sugar intake. The specific sugar they were assessing was sucrose, found in normal table sugar.
For the analyses of people's self-reported sugar intake based on dietary record, the researchers took into account how many calories each person ate so this did not affect the analysis.
They then looked at how well the two types of sugar consumption measurement compared, and how likely people at the five different levels of sugar consumption were to be overweight or obese after three years, based on their BMI and waist circumference.
What were the basic results?
Results showed a striking difference between the urine sugar measurements and the sugar intake based on the diet diaries.
People who had the highest levels of sugar in their urine were more likely to be overweight after three years than those with the lowest levels.
The reverse was true when researchers looked at the people whose diet diaries suggested they ate the most sugar relative to their overall calorie intake compared with the least.
Using the urine sugar measurement, 71% of people with the highest concentration were overweight three years later, compared to 58% of people with the lowest concentration.
This meant that having the highest urinary levels of sugar was associated with a 54% increase in the odds of being overweight or obese after three years (odds ratio [OR] 1.54, 95% confidence interval [CI] 1.12 to 2.12).
Using people's seven-day diet diaries, 61% of people who said they ate the most sugar relative to their overall calorie intake were overweight, compared to 73% of people who said they ate the least sugar.
This meant those who reported the highest sugar intake relative to their overall calorie intake were 44% less likely to be overweight or obese after three years (OR 0.56, 95% CI 0.40 to 0.77).
How did the researchers interpret the results?
The researchers conclude that, "Sucrose measured by objective biomarker, but not self-reported sucrose intake, is positively associated with BMI."
They say there are "several possible reasons" for the discrepancies between the methods used to assess sugar intake. They admit the spot check urinary sugar marker may have disadvantages, but conclude that under-reporting of foods with high sugar content, particularly among those who are overweight or obese, may be a contributing factor.
As a result, they say future researchers looking at sugar as part of diet should consider using an "objective biomarker" such as urinary sugar, rather than relying on people's own estimates of what they have consumed.
This study has found conflicting associations between an objective measure of sugar intake and a subjective measure of sugar intake based on food diaries, and the risk of a person becoming overweight.
While more sugar in urine samples was associated with a greater risk of becoming overweight, consuming more sugar (based on food diary records) was actually associated with a reduced risk.
If the urine biomarker is a more accurate reflection of sugar consumed than diet diaries, then this research may explain why some previous diet studies have failed to show a link between sugar and being overweight.
However, there are some limitations to consider with the urine biomarker. Because the test used was a one-off snapshot of sugar intake, it can only show us how much sugar was in the person's urine at the time they were tested. Similar to a short-term food diary, we don't know whether that is representative of their sugar consumption over time.
The urine test is also not able to measure very high or very low sugar levels. The analyses of urine sugar levels did not adjust for overall calorie intake, while those for self-reported sugar intake did. It would have been interesting to see whether the association between urinary sugar levels remained once calorie intake was taken into account.
The current study did not assess why the dietary records and urinary measures of sugar differed. It also did not assess whether the discrepancies were larger among people who were overweight or obese at the start of the study – only how these measures were related to the outcomes at the end.
So it is not possible to say from this study alone that people who were overweight or obese had greater discrepancies between what they reported eating and their urinary sugar measurements.
However, the authors report that other studies have shown overweight people, especially women, are prone to under-reporting diet, particularly between-meal snacks.
As with all observational studies, it is difficult to rule out that factors other than those being assessed might be having an effect on the results. The researchers adjusted their analyses for age and gender, and say that results "did not change materially" after they adjusted the figures to take account of people's physical activity levels.
The results do not appear to have been adjusted to take account of other factors, such as people's level of education, income or other components of their diet, which may have an effect on weight.
The effect of sugar on health, independent of calorie intake, is still being debated by health organisations. If the findings of the current study are correct, using objective measures of sugar intake could help assess its effect on obesity and more widely on health.
Analysis by Bazian. Edited by NHS Choices. Follow Behind the Headlines on Twitter. Join the Healthy Evidence forum.