Understanding the role of bitter taste perception in coffee, tea and alcohol consumption through Mendelian randomization

Abstract

Consumption of coffee, tea and alcohol might be shaped by individual differences in bitter taste perception but inconsistent observational findings provide little insight regarding causality. We conducted Mendelian randomization analyses using genetic variants associated with the perception of bitter substances (rs1726866 for propylthiouracil [PROP], rs10772420 for quinine and rs2597979 for caffeine) to evaluate the intake of coffee, tea and alcohol among up to 438,870 UK Biobank participants. A standard deviation (SD) higher in genetically predicted bitterness of caffeine was associated with increased coffee intake (0.146 [95%CI: 0.103, 0.189] cups/day), whereas a SD higher in those of PROP and quinine was associated with decreased coffee intake (−0.021 [−0.031, −0.011] and −0.081 [−0.108, −0.054] cups/day respectively). Higher caffeine perception was also associated with increased risk of being a heavy (\textgreater4 cups/day) coffee drinker (OR 1.207 [1.126, 1.294]). Opposite pattern of associations was observed for tea possibly due to the inverse relationship between both beverages. Alcohol intake was only negatively associated with PROP perception (−0.141 [−1.88, −0.94] times/ month per SD increase in PROP bitterness). Our results reveal that bitter perception is causally associated with intake of coffee, tea and alcohol, suggesting a role of bitter taste in the development of bitter beverage consumption. Coffee, tea and alcohol are widely consumed beverages with bitter taste 1 and have been implicated in both beneficial and adverse health effects 2,3. Individual differences in metabolizing caffeine 4,5 and ethanol 6-10 present in these beverages determine their consumption level, whereas the influence of taste factors remains unclear. The relationship between the perception of bitter compounds, such as propylthiouracil (PROP), quinine, and caffeine, and the consumption level of these bitter beverages have been inconsistent across observational studies 11-22. As performing taste tests is a resource-intensive process, these investigations likely lacked sufficient power to convincingly rule out moderate effects. Potential confounders or illnesses that both affect taste perception and consumption in these studies limits their ability to provide unbiased causal effects. Fortunately, these issues can be overcome by recent advances in taste genetics and statistical methodology. Heritability for the perceived intensity of PROP, quinine, and caffeine have been estimated to be 0.73, 0.40, and 0.36 in classical twin studies 23,24. Furthermore, candidate gene and genome-wide association studies (GWAS)