Coffee consumption and liver health

Abstract

Beneficial associations between coffee drinking and a range of liver outcomes have been consistently reported in observational research, yet no randomised controlled trial has been conducted to investigate whether drinking more coffee might reduce the risk of progression of Non-Alcoholic Fatty Liver Disease (NAFLD). NAFLD is an umbrella term for a pathological pathway that includes steatosis, steatohepatitis, fibrosis, cirrhosis, and hepatocellular carcinoma, where no other aetiology is identified such as alcohol or viral hepatitis. NAFLD is an important public health issue with a general population prevalence of approximately 25% that has risen in parallel with that of obesity, and as such represents a significant burden to individuals and health systems. NAFLD has few treatment options and current best advice is to lose weight through healthy diet and exercise. If coffee was shown to have benefit in reducing the risk of NAFLD progression it would be a valuable addition to the current management of the condition.
The methodological approach of a randomised controlled study could be shaped by addressing a number of current knowledge gaps. Firstly, could increasing coffee intake cause additional non-liver harm in people. To address this issue an umbrella review, or review of reviews, was conducted to draw together the vast amount of existing research between coffee intake and multiple health outcomes. Reassuringly, outside of pregnancy, drinking coffee was more frequently associated with benefit than harm. For important generic outcomes such as all-cause mortality, cardiovascular mortality, and incident cardiovascular disease, maximum relative risk reduction was associated with intakes of 3-5 cups a day. Some harmful associations, such as between coffee drinking and lung cancer, were nullified by adequate adjustment for smoking, known to be an important confounder. Liver outcomes consistently had the largest magnitude of beneficial associations with coffee drinking. Secondly, in observational research, ascertainment of coffee intake is usually measured in cups a day. This is a heterogeneous measure because of different preparation methods, cup sizes, coffee beans, and roast types, resulting in the risk of misclassification. To overcome this limitation the next stage of the research aimed to create a coffee unit measure, similar in concept to alcohol units, that took preparation method and cup size into account. The unit measure, where 1 coffee unit was equivalent to a 227mL cup of instant coffee, was then applied to a representative UK population using data from the National Diet and Nutrition Survey, and the proportion of misclassified intake, when not accounting for preparation type and cup size was derived. Overall, approximately 1 in 4 participants had misclassified intake, largely under or over estimated by one cup a day. This effect of 25% misclassification of coffee intake in existing research is of uncertain significance, but would generally be non-differential, and therefore more likely to dilute risk estimates of both benefit and harm. The coffee unit measure could be applied to a future experimental study to better quantify coffee intake or allow increases in consumption across preferred preparation types.
Coffee preparation preferences were explored as part of the final element of the research, which was a mixed-methods study designed to explore patterns of coffee consumption in a secondary care population of people with NAFLD, their views about drinking more coffee, and acceptability of a randomised controlled trial in which drinking more coffee was the intervention. The mixed method study included an initial qualitative phase of 17 semi-structured interviews that were used to inform the final design of a questionnaire to explore the same phenomenon in a stratified sample of 393 people with NAFLD recruited from three NHS secondary care sites. In the survey, which was stratified across three liver stiffness groups (<7 KPa, 7-13 KPa, and >13 KPa), 78% of respondents were current coffee drinkers, and 22% non-coffee drinkers. Median coffee consumption was 2 cups a day (interquartile range 1 to 3 cups). The proportion of coffee drinkers reduced as liver stiffness increased but not the median daily cup intake. Nearly half of non-coffee drinkers thought they would be able to start drinking it, and 85% of those drinking <4 cups a day thought they would be able to drink an additional 2 cups a day. These proportions reduced to 38% and 66% respectively when considering those who also expressed an interest, albeit hypothetically, in becoming involved in a randomised controlled trial. In this group of participants, acceptable options for increasing coffee intake included 71% for drinking their own coffee at their own expense, 32% being supplied instant coffee, 27% being given a monetary allowance towards the extra coffee, and 15% being supplied ground coffee. Other aspects of a future experimental study including randomisation, and blood and imaging tests were generally considered acceptable. Importantly this data suggests that recruiting people with NAFLD into a future experimental study would be possible from an NHS secondary care setting. Arguably, now is the time for such a study, in the context of the huge burden of NAFLD, the lack of effective treatments, and the potential coffee has to offer benefit

More information

Identifiers

ORCID for Julie Parkes: orcid.org/0000-0002-6490-395X

Catalogue record

Export record