A recent population-based study described a borderline significantly higher carbohydrate intake for heterozygote MC4R
V103I genotype carriers (7
). Since a carbohydrate-rich diet is associated with a leaner body composition (12
), it was speculated that this polymorphism plays a role in appetite regulation in humans by reducing the quantity or modifying the quality of food intake. This might modulate central obesity, as well as lipid and glucose metabolism. In line with these findings (7
) we observed a significantly higher carbohydrate intake in rare allele carriers in our study with severely obese subjects. However, these genotype carriers had also a general increase in energy intake which might explain why we did not find a lower BMI as reported in the literature. Previous studies in population-based subjects or spanning the entire range of BMI showed a lower BMI in rare allele carriers (2
). This association might be disturbed in severe obese subjects as seen in our study as well as in the only other large studies investigating severely obese persons of obesity class II and III (5
It might be speculated that the rare allele of the MC4R
V103I may modulate appetite towards higher carbohydrate intake, which may have strongly differential effects on different population groups. For the general population, there may be a tendency towards a leaner phenotype being associated with increased carbohydrate intake, while for a subgroup in already a severe obesity stage, this appetite modulation is connected with higher total energy intake counteracting a leaner phenotype. According to mediator analysis proposed by Prentice et al. (14
), the to some part mediating role of carbohydrate intake is supported by the analysis of BMI adjusted for carbohydrate intake. This analysis yielded decreased association estimates for BMI which supports the idea that the influence of this polymorphism on body size measures is mediated by the nutritional behavior.
Even though our study included 1029 mostly severely obese patients, one might assume that our findings might be limited due to the small number of 103I genotype carriers as the variant is rather rare. While we had 70% power for the observed higher intake of carbohydrates, our power to detect a BMI difference of 1.5 kg/m² was only 20%. Furthermore, the nutritional variables are very difficult to assess and usually involve substantial uncertainty (15
), which aggravates the power issue. However, statistical theory teaches that such random uncertainty in the regression outcome variable would not impose a bias on the estimate, as the uncertainty can be considered to be undifferential between variant carriers and non-carriers. Therefore, the association estimate with the true underlying carbohydrate intake could be expected to be the same as observed but with smaller confidence intervals and therefore better p values.
Underreporting of dietary intake in obese individuals is also a well-known problem. However, as we analyze a “case-only group”, systematic underreporting could be accounted for by subtracting a constant from the outcome variable values and would neither result in biased estimates nor in a loss of precision, as the underreporting can be considered to be independent of the genotype status (15
Finally, it is highly unlikely that our finding is a technical artifact of genotyping since we confirmed each 103I genotype carrier and a similar number of wildtype genotype carriers by sequencing.
We propose that future large-scale studies may especially analyze the association of MC4R polymorphisms stratified by severity of obesity. If our preliminary finding can be confirmed, we propose a differential influence of this polymorphism in severely obese persons compared to the general population. It is conceivable that several interconnected mediators in appetite regulation are heavily disturbed in severely obese individuals which changes the influence of this polymorphism on the ingested nutritional spectrum.
In summary, the higher dietary intake of carbohydrates in severely obese individuals with the MC4R 103I variant is in line with previous findings and might indicate a differential consequence on body size measures in extremely obese subjects when compared to the general population which, however, needs confirmation in independent studies or meta-analyses.