Assessing self-reported green tea and coffee consumption by food frequency questionnaire and food record and their association with polyphenol biomarkers in Japanese women.

BACKGROUND AND OBJECTIVES Despite the demonstrated protective effects of green tea and coffee intake against several chronic diseases, finding between studies have not been consistent. One potential reason of this discrepancy is the imprecision in the measurement of tea or coffee consumption using food frequency questionnaire (FFQ) and food record (FR) in epidemiological studies. METHODS AND STUDY DESIGN In a sample of 57 healthy Japanese women, intake of green tea and coffee was estimated by a validated FFQ and a 3-day FR, while their plasma and urine concentrations of polyphenol biomarkers were measured by HPLC. The polyphenols assessed included (-)-epigallocatechin gallate (EGCG), (-)-epicatechin gallate (ECG), (-)-epigallocatechin (EGC) and (-)- epicatechin (EC), caffeic acid (CA) and chlorogenic acid (CGA). RESULTS Green tea consumption estimated by FFQ and FR showed moderate association, while strong association was detected for coffee consumption. Urinary green tea polyphenol concentrations were moderately-strongly associated with FR-estimated intake, while the associations were weak with FFQ. Similarly, coffee polyphenols in urine were moderately associated with FR-estimated coffee intake, whereas FFQ showed poor correlation. The associations between urinary and plasma polyphenols ranged from moderate to high. CONCLUSIONS The results indicated that firstly, the FFQ tends to overestimate green tea intake. Secondly, the urinary polyphenols are preferred over plasma polyphenols as a potential surrogate marker of the short-term green tea and coffee intake, while their use as an indicator of long-term consumption is not reliable.