• Potential conflict of interest: Nothing to report.


We appreciate the interest and comments by Prati and colleagues regarding our recent publication. The main objectives of our study were to determine the spectrum of nonalcoholic fatty liver disease (NAFLD) using the proposed “healthy” limits in morbidly obese women. The authors referred to our study population as “very peculiar.” Webster's dictionary defines “peculiar” as “distinctive.” We agree that our cohort is distinctive and that the US prevalence of at least 10 million individuals1 with morbid obesity underscores its importance. Also, our cohort was being evaluated for bariatric surgery, thereby avoiding alanine aminotransferase (ALT) referral bias. Our study demonstrated that 63% of individuals had “unhealthy” ALT values (>19 U/L), compared with only 27% (>30 U/L, our present ULN), and the majority of these patients had early stages of NAFLD. Approximately 20% of our cohort had minimal fat on biopsy (<5%, considered normal by many experts), and another 36% had fatty liver only; therefore, significant chronic liver injury was not 100%, as suggested by the authors.

The authors questioned the clinical benefit of detecting a biochemical silent liver disease among morbidly obese individuals. Would they support the same statement for individuals that are simply overweight or less obese? Our data demonstrated that the prevalence of nonalcoholic steatohepatitis (NASH) (23%) and initial poor function (IPF) (37%) was identified in subjects with “healthy” ALT values (<19 U/L), supporting the notion that the newly proposed “healthy range” for ALT is not very useful in predicting the more progressive form of NAFLD. Elevated ALT has been suggested to be an independent risk factor for advanced fibrosis2; we expected that the proposed ALT “unhealthy” cutoff would increase the prevalence of detecting none or mild fibrosis, and our data supported that assumption.

Our study never addressed whether any ALT value was “cost effective,” but it did address the potential health care costs if the proposed “healthy” range is adopted. We must accept the reality that many more individuals will have an (H) next to their ALT level. This (H) represents a statistically higher value than the “healthy” proposed limit. The authors suggested that diagnostic algorithms and thresholds should be individualized depending upon clinical circumstances. We suspect the reality will be the following: the astute clinician will repeat “higher than healthy” ALT values while counseling overweight and obese subjects to lose weight and exercise. The clinician will eventually have to deal with scenarios in which the patient does not lose weight3 or regains weight and the ALT remains “higher than healthy.” The clinician can either ignore the “unhealthy” ALT or obtain an ultrasound to confirm hepatic steatosis and rule out etiologies that are less likely (viral hepatitis, iron overload, autoimmune hepatitis) but clearly in the differential diagnosis. Finally, the necessity to obtain a liver biopsy to diagnose NASH continues to be debated.4 Unfortunately, our current medical-legal society encourages defensive medicine and we posit that the proposed “healthy” ALT cutoff will result in more diagnostic tests, thereby increasing health care expenditures. Our data support the authors' contention that an ALT level that is statistically “healthy” does not affirm that an individual is healthy, further highlighting the limited diagnostic utility of ALT in NAFLD.

Gary A. Abrams M.D.*, Ronald H. Clements M.D.*, Audrey J. Lazenby M.D.*, Sachin S. Kunde M.D., M.P.H.†, * University of Alabama at Birmingham, Birmingham, AL, † Wayne State University Detroit, MI.