We read with interest the paper by Muthiah and colleagues [1] on the differential impact of non-alcoholic fatty liver disease (NAFLD) and metabolic dysfunction associated fatty liver disease (MAFLD) definitions on long-term outcomes including major adverse cardiovascular events, advanced fibrosis, and all-cause and cardiovascular-related mortality in patients with diabetes. We have fundamental concerns with this work.
It is pivotal to remember that unbiased interpretation of data is as important as performing rigorous experiments and studies. The interpretation of data can itself result in bias, so-called “interpretive bias”, which can affect reporting and leads to questionable conclusions [2]. Interpretive bias can be driven by two well-known types of bias, namely the “wish” bias of individuals and confirmation bias [2].
“Wish” bias is the tendency for people to interpret information according to their wishes or beliefs and has been demonstrated in the past for several societal and scientific efforts, irrespective of topic [3,4]. A long recognized phenomenon in human psychology, such bias distorts the results of a study and reduces its reliability[3,4]. Confirmation bias on the other hand is the tendency of people to interpret information in a way that conforms to their expectations. Confirmation bias can be large and can also influence the interpretation of results [5]. Clinical research is particularly susceptible to such systematic biases and lapses in conduct due to the infiltration of financial and intellectual conflicts of interest, among other factors [2].
In this work [1], the authors observed that the MAFLD definition increased fatty liver diagnosis by 68.89%. Patients who were classified as MAFLD (+)/NAFLD(-) were at a higher risk of major adverse cardiovascular events, advanced fibrosis, and all-cause and cardiovascular-related mortality compared to MAFLD(+)/NAFLD(+). Viral hepatitis had a synergistic impact with MAFLD in increasing the risk of advanced fibrosis and all-cause mortality. However, in spite of these results, the authors concluded from their data that (a) it was premature to change to MAFLD as it results in an over-diagnosis of fatty liver, (b) MAFLD exaggerated mortality and morbidity in patients with T2DM, (c) the definition of MAFLD causes further heterogeneity in fatty liver disease by including patients with other liver diseases, and (d) while NAFLD focuses on the liver, a limitation in MAFLD is that it seeks to capture systemic factors associated with hepatic steatosis. They conclude that the increased risk of adverse events might not be an accurate representation of a derivative risk from fatty liver but a result of systemic comorbidities associated with the condition. Going through their claims systematically.
According to the textbook definition “Overdiagnosis” is the diagnosis of disease that will never cause symptoms or death during a patient's ordinarily expected lifetime and thus presents no practical threat regardless of being pathologic” [6]. How can this fit with the authors sentence that patients with MAFLD (+)/NAFLD(-) are at a higher risk of all outcomes compared to MAFLD(+)/NAFLD(+), consistent with other studies [7]. Obviously, it is not overdiagnosis, but rather that MAFLD improves diagnosis by capturing the previously missed cases when using the NAFLD definition
According to the definition “exaggerated” means represented as larger, better, or worse than in reality. Does the MAFLD definition exaggerate the outcome or capture it better? If you have two biomarkers or scores and one of them has a high diagnostic or prognostic utility, does this mean this biomarker exaggerated the outcome or that it improved the diagnostic/prognostic utility? The authors’ data suggests that the MAFLD definition improves diagnostic/prognostic utility.
According to the authors’, viral hepatitis has synergistic effects with MAFLD on the outcomes. They state, “This is unsurprising given the presence of dual liver aetiology resulting in higher mortality”. Is it not a good thing that the MAFLD definition helped identify its coexistence with other liver diseases that were previously missed under the primitive NAFLD definition? Again, is not coexistence of liver diseases a real life entity that we see in clinics on a daily basis?
Ultimately, they claim that MAFLD seeks to capture systemic factors associated with hepatic steatosis, while NAFLD focus on liver. The centre-point of MAFLD is that it identifies pathophysiology and we treat the patient not the organ. If the revolutionary change from NAFLD to MAFLD has not done anything, attracting attention to the systemic nature of the condition is a win for patients and for Hepatology.
Going through other similar articles from the same group of authors tells us that theirs is a systematic pattern, not a random error of interpretation. One wonders whether this bias in the interpretation offers opportunities to study what happens to the process of consensus on fatty liver disease redefinition when self-speaking results are deliberately misinterpreted. This questions the objectivity of the entire discipline of consensus processes [8,9]. Moreover, it can be difficult to discern whether consensus is based on careful consideration of all the evidence and its accurate interpretation, inappropriate entrenchment of old information, lack of dissemination of newer data, or purposeful silencing of their existence.
In conclusion, efforts should be redoubled to increase awareness within the scientific and academic community to improve adherence to transparent reporting and interpretation as a crucial aspect of data integrity, exactly as when performing the study. Accuracy in interpretation should be taken seriously as otherwise it leads to an erosion in the perceived legitimacy of science as an impartial means of finding the truth.
