Misrepresentation of self-reported drinking and genetic studies

It is well known that modifying behaviours and lifestyle can serve as risk or protective factors for common diseases.
 
Alcohol consumption presents one of the most common harms, particularly for increased risk of cardiovascular and metabolic diseases and mental health.  
 
A number of large-scale epidemiological studies on alcohol consumption conclude that no level of alcohol consumption improves health.
 
But according to new research, this conclusion may contradict negative estimates of the genetic link between alcohol consumption and some diseases – including obesity, major depressive disorder, Parkinson’s disease, and type 2 diabetes – as reported in recent genome-wide association studies identifying numerous genetic variants associated with human behavioural traits.
 
It is also contradictory to the protective effects of moderate drinking reported in observational studies.
 
The recently published observational study found genetic studies that rely on self-reported drinking may be biased by misreporting and changes in behaviour. The study found conflicting results from some genetic studies in that people with genetic-based tendencies towards drinking also have lower rates of some diseases, suggesting a possible protective effect of alcohol against disease.
 
One possible explanation the authors found was that patients with a disease may change their alcohol consumption upon diagnosis, or misreport their alcohol consumption in a survey.
 
According to the authors, behavioural traits are subject to misreports and longitudinal changes, and misreports are common in self-reported data sets but often overlooked in genetic analyses. This can cause biases in genome-wide association studies and follow-up analyses.
 
These misreports are mainly attributed to social desirability, when the tendency for participants to answer questions is in ways that make them viewed favourably by others, and recall bias, when the accuracy and completeness of past events recalled by participants are influenced by subsequent events that they experienced.
 
It is hoped the study findings will help explain conflicting reports about the link between alcohol consumption and certain diseases.
 
Using data from 455,607 individuals in the UK Biobank, the researchers investigated biases due to misreports and longitudinal changes in genetic analyses of self-reported behavioural traits, including alcohol consumption, tobacco smoking, and physical activities, without correcting for misreporting or changes in behaviour.
 
Their findings were negative genetic correlations between alcohol consumption and type 2 diabetes, hypertensive disease and iron deficient anaemias. After correcting for misreporting and behavioural changes, the negative correlations disappeared and they found positive correlations with eight different conditions, including cardiovascular disease and total disease count.
 
‘Our studies demonstrate that individuals with higher disease burden in the UK Biobank are more likely to misreport or reduce their alcohol consumption levels, and propose a correction procedure to mitigate the misreports and longitudinal changes-induced biases,’ the authors wrote.
 
‘The alcohol consumption genome-wide association studies signals removed by the misreports and longitudinal changes corrections are enriched in metabolic/cardiovascular traits. Almost all the previously reported negative estimates of genetic correlations between alcohol consumption and common diseases become positive/non-significant after the misreports and longitudinal changes corrections.’
 
Another important source of bias noted by the authors is the change in amount of alcohol consumption during the life course, for reasons such as changes in health status. For example, if people change their level of alcohol consumption because they are affected by a disease, the diagnosis will ‘give rise to a bias in observed or genetic relationship between alcohol consumption and the disease’.
 
The authors anticipate the results of this study will demonstrate the significance of potential biases in genetic studies of behavioural traits, and provide a method to correct them for future research. They do note, however, that this bias does not necessarily apply to all populations or behavioural traits.
 
‘Our findings provide a plausible explanation of the controversy about the effects of alcohol consumption on health outcomes and a caution for future analyses of self-reported behavioural traits in biobank data,’ the authors wrote.
 
Log in below to join the conversation.

alcohol consumption genetics preventive health