Does margarine lower child IQ?

The Daily Mail today reported that “margarine consumption is linked to lower IQs in children”. The newspaper said that a study from New Zealand has found that children who ate margarine every day had IQs up to six points lower than children who did not. The researchers reported that the reason for the link was unclear, but the newspaper suggests that the trans fats in margarine may be to blame.

A spokesperson for the Food Standards Agency is reported as saying that “trans fat consumption in the UK is now below the recommended level”.

This study looked at diet and IQ in children at ages three-and-a-half and seven years. Because the study looked at diet and IQ at the same point of time, it is not possible to say whether the differences seen in diet were a cause of lower IQ.

In addition, the study carried out multiple statistical tests, which increases the possibility of finding statistically significant associations just by chance. The study did not assess the effects of individual dietary components such as trans fats, meaning no conclusions can be drawn about their effects on IQ.

Further studies will be needed to investigate the observed links before firm conclusions can be made.

Where did the story come from?

Dr Reremoana F Theodore and colleagues from the University of Auckland in New Zealand carried out this research. The study was funded by various sources, including the Health Research Council of New Zealand and other research foundations. The study was published in the peer-reviewed journal Intelligence.

What kind of scientific study was this?

This study featured cross-sectional analyses looking at diet and IQ in children taking part in the Auckland Birthweight Collaborative (ABC) study.

The ABC study followed a cohort of babies who were either small for their gestational age (SGA) at birth or of an appropriate weight for their gestational age (AGA). SGA was defined as being in the lowest 10% of expected weight for their gestational age and gender. AGA was defined as being above the lowest 10% of expected weight for their gestational age and gender.

The babies deemed eligible for inclusion were those from full-term births (i.e. not premature) occurring in two areas in New Zealand between October 16 1995 and November 30 1996. Babies with conditions at birth that were likely to affect growth and or development were excluded.

Information on the diets of the children was collected at age three-and-a-half and seven years, using a food frequency questionnaire (FFQ). The FFQ asked how often a child had eaten 88 different foods in the past four weeks: never, one to three times in the past month, one time a week, two to four times a week, five to six times a week, once a day, or two or more times a day.

Based on this information, the researchers calculated how many children’s diets were in line with the 2002 New Zealand Ministry of Health guidelines on fruit, vegetables, breads and cereals (including rice and pasta) meat, fish, chicken and eggs, and milk and dairy products. The researchers suggested that these guidelines were similar to guidelines from other countries.

They also looked at the children’s intake of red meat, and of food categories that contain nutrients that have been associated with cognitive functioning: fish, oily fish, margarine, butter, blended spread, and vitamin and mineral supplements.

Standard tests were used to assess IQ at ages three-and-a-half and seven years. The researchers conducting the IQ tests were not made aware of the children’s diets. They looked at whether various aspects of the children’s diets were related to their IQs at either age. They took into account the fact that their sample contained more SGA children than the general population.

Once the researchers identified foods that showed some association with IQ, they looked at the combined influence of all of these foods together to see which ones remained significant after taking all the others into account. This analysis also adjusted for other factors that could affect the results (potential confounders), such as gestation, mother’s number of previous deliveries, sex, maternal school-leaving age, parental occupation, marital status, maternal body mass index (BMI), children's BMI, and which researchers administered each IQ test.

At the start of the study, 1,714 mothers agreed to be in the study (840 SGA and 877 AGA). For this study, only children of European descent were analysed, as children from other ethnic groups had low response rates during follow-up. There were 871 children of European descent at the start of the study, with 531 (61%) of these participating and providing IQ data at three-and-a-half years, and 589 (68%) participating and providing IQ data at seven years.

Mothers who participated in the study had higher socio-economic status than those who did not. There was no difference between SGA and AGA children in IQ or diet.

What were the results of the study?

The researchers found that at age three-and-a-half, children who ate breads and cereals four or more times a day had IQ scores an average of 3.96 points higher than children who ate less of this food group. However, only a few children ate breads and cereals four or more times a day, and when the researchers looked across all children and adjusted for all potential confounding factors, the link between amount of bread and cereal eaten and IQ was not statistically significant.

Children who ate margarine at least once a day at age three-and-a-half had IQ scores that were an average of 2.81 points lower than children who did not. There was no link between other food groups and IQ at age three-and-a-half.

At age seven years, children who ate fish weekly had IQ scores 3.64 points higher on average than those who did not. Although some other food groups showed an association with IQ at age seven, these associations were no longer significant after potential confounding factors were taken into account.

When looking just at children who were born small for their gestational age, eating margarine daily was associated with lower IQ scores at ages three-and-a-half and seven. Other food groups did not have significant effects after taking into account other food groups and confounding factors.

What interpretations did the researchers draw from these results?

The researchers concluded that eating fish, breads and cereals at levels recommended by New Zealand’s nutritional guidelines “may be beneficial to children's cognitive development”. Eating margarine daily was associated with poorer cognitive functioning in children aged three-and-a-half. They suggested that further research was needed to determine what causes this association.

What does the NHS Knowledge Service make of this study?

There are a number of points to consider when interpreting this study:

• The study looked at diet and IQ at the same point in time. Diet at time of measurement may not have been representative of prior diet, and therefore it's not possible to say whether it could potentially have caused the differences seen. The authors note that the study “does not prove causation”.
• The food frequency questionnaire that was used performed well compared to keeping a short-term food diary, but there may be some inaccuracies in the parent’s recollection or estimation of details of their child’s diet. Their answers are also likely to be most representative of the child’s recent diet and not any diet further in the past.
• The questionnaire did not appear to assess how much of each food a child ate; only how frequently they ate it. Therefore, eating similar amounts of a food could potentially be counted as eating them at different frequency.
• This study only included children of European descent and any results may not apply to children from other ethnic backgrounds.
• A reasonably high proportion of women who enrolled in the study did not take part in follow-up sessions (32% and 39% at the two time points), with those women who did participate having a higher socioeconomic status than those who did not. Therefore, the results may not be representative of what would be seen in the population as a whole.
• The study carried out multiple statistical tests, which increases the possibility of finding statistically significant associations just by chance.
• Although the researchers adjusted for some factors that could potentially affect results, it is possible that these adjustments have not removed the effects of these factors entirely, and other factors could be having an effect.