Weight loss is a recommended treatment strategy for non-alcoholic fatty liver disease (NAFLD), but is difficult to achieve and sustain in clinical practice. Physical activity (PA) not only causes weight loss but also may improve insulin resistance and reduce total body and intrahepatic fat.
The study by Gerber et al. is the first to use a quantifiable measure of PA in a large population. They have used Triaxial accelerometer which quantifies PA as sedentary, moderate or vigorous as opposed to previous studies based on self-reports of exercise intensity and duration.[3-5] One could argue that they have captured only a small portion of someone's PA (10 h/day), but the measure is objective.
Data from this study confirms the previously known findings that NAFLD subjects have significantly low level of PA than non-NAFLD. Patients carrying out the lowest PA unsurprisingly had higher body mass index (BMI). Their data also suggest reduced PA with dual diagnosis of diabetes mellitus (DM) and NAFLD.
However, the difficulty with interpreting this study is the case definition. The authors rely upon the fatty liver index (FLI) to identify NAFLD patients. This index is based upon the patient's BMI, waist circumference, serum triglycerides and serum gamma glutamyl transferase. The FLI has been associated with increased all-cause, as well as liver-specific mortality. The advantage of using such a generalised definition is that a large cohort of patients can be studied. However, in view of the components of the FLI, it is hardly surprising that NAFLD patients were more likely to be obese and have DM. The observation that obese patients are less active is not a new one, and similarly it is recognised that few diabetic patients achieve recommended levels of activity.
Fundamentally two questions arise. First, has this study observed an association specifically related to NAFLD? A comparison between obese patients with a low FLI and those with a high FLI would be insightful. Second, is the lack of PA the cause or effect of a raised FLI? In this regard, it is interesting to see that there was no correlation between PA and elevated transaminases.
However, the authors are to be applauded for their quantitative assessment of PA in this patient group. Further studies would need to address the exact role of similarly well quantified PA as an intervention in improving histological features of NAFLD.
In the meantime, as the authors rightly conclude, we should encourage PA amongst our NAFLD patients until we get more data considering its benefits on their morbidity and mortality.