Bitter-Sweet: Two epidemiological studies on the relationship between diet and health in the UK Women’s Cohort Study

Joshua Lambert, Ph.D., Associate Professor of Food Science, Department of Food Science, The Pennsylvania State University

Date and Location

When (Date/Time)

January 18, 2018, 4:00 PM - 5:00 PM


252 Erickson Food Science Building

Add to calendar


The UK Women’s Cohort Study (UKWCS) is a prospective cohort study of more than 35,000 women in the United Kingdom.  Baseline health, diet, and lifestyle information was collected between 1995 and 1998.  Collection of mortality and incident cancer data has been on-going since the study subjects were recruited.  The focus of my sabbatical was to learn the basics of nutritional epidemiology and how to integrate this research approach into my group’s overall interest in the impact of food and food components on chronic diseases.  To this end, I conducted two studies.  The first investigated the relationship between a phenotypic marker (i.e. phenylthiolcarbamate [PTC] taster status in 5,500 subjects) or a genotypic marker (i.e. TAS2R38 diplotype in 750 subjects), food intake patterns, and risk of malignant cancer.  PTC tasters (hazard ratio [HR] = 1.30, 95% confidence interval [CI]: 1.04, 1.62), but not supertasters (HR = 0.98, CI: 0.76, 1.44), had increased cancer risk compared to nontasters.  An interaction was found between phenotype and age for supertasters (p = 0.019) but not tasters (p = 0.54).  Among women > 60 y, tasters (HR = 1.40 CI: 1.03, 1.90) and supertasters (HR = 1.58, CI: 1.06, 2.36) had increased cancer risk compared to nontasters, but no such association was observed among women ≤ 60 y (tasters HR = 1.16, CI: 0.84, 1.62; supertasters HR = 0.54, CI: 0.31, 0.94).  I found no association between TAS2R38 diplotype and cancer risk.  I observed no major differences in bitter fruit and vegetable intake.  These results suggest that the relationship between PTC taster phenotype and cancer risk may be mediated by factors other than fruit and vegetable intake.  In the second study, I examined the relationship between solid chocolate intake and risk of all-cause (ACM), cardiovascular disease (CVD), and cancer-related mortality in 29,253 women.  Chocolate intake was associated with decreased ACM (HR = 0.79, CI: 0.71, 0.88 for highest category of consumption) and CVD (HR = 0.61, CI: 0.43, 0.87) mortality risk.  I found that prevalent angina at baseline was a significant modifier of the relationship between chocolate consumption and ACM (p = 0.002).  Among women with angina, women in the highest category of chocolate consumption had a 59% lower risk of ACM (HR = 0.41, CI: 0.26, 0.65), whereas among women without prevalent angina, risk was only 19% lower (HR = 0.81, CI: 0.72, 0.90). These results demonstrate that chocolate consumption is associated with reduced risk of ACM and CVD mortality in women.  For ACM, this association is stronger in women with angina.