The study by Mofrad1 is a laudable effort to better characterize the clinical spectrum of nonalcoholic fatty liver disease (NAFLD). Previous reports also showed that patients with NAFLD-induced cirrhosis have lower aminotransferase values than those with cirrhosis of other etiologies, 32 suggesting that low aminotransferase levels can coexist with severe histological injury.
In our view there are, however, several methodological limitations relative to the selection of patients and to the definition of “normal” alanine aminotransferase (ALT). Lessons from hepatitis C patients with “normal” ALT levels have taught us that repeated measurements of ALT activity are necessary before concluding that ALT levels are persistently low,4 while half of untreated patients experience rises in ALT.5 Sanyal rightly remarks that in NAFLD, ALT levels can fluctuate6; therefore, a single measurement is insufficient to conclude about the normalcy of ALT. Also, patients included here might have a predisposition toward more severe liver disease. Indeed, most of them (31/51) had various abnormal findings suggestive of liver disease. In this regard, they differ from patients with NAFLD and low ALT levels whom we see in our practice and who are asymptomatic. I
|Type of Treatment||Penicillamine Group||Zinc Group|
|Number of patients (males)||86 (54)||21 (12)|
|Median age at diagnosis in years (range)||7.7 (1–18)||5.5 (1–12.3)|
|Patients with liver disease||79 (91%)||15 (71%)|
|Patients with neurological disease||4 (5%)||2 (10%)|
|Pre-symptomatic||3 (4%)||4 (19%)|
|Basal median ALT values (U/L) (range)||274 (54–640)||177 (28–452)|
|Patients with ALT normalization at the end of follow-up||56 (65%)||12 (57%)|
|Patients with persistent hyper-ALT at the end of follow-up||28 (32.5%)||9 (43%)|
|Median ALT values in patients with persistent hyper-ALT (U/L) (range)||76 (46–960)||72 (45–268)|
The most important problem lies with the definition of “normal” ALT. We believe that a threshold of 75 IU/L is abnormally high for defining normal ALT. The range of normal values was determined long ago in blood donors, on the assumption that absence of risk factors for blood-borne infections, together with low-risk sexual behavior, equates an absence of liver disease. It has been since recognized that ALT activity increases with body mass index7 and that a high body mass index is associated with liver disease.6, 8 The same holds true for the relation between serum triglycerides and ALT activity in patients with metabolic syndrome.9 When both metabolic and viral risk factors for liver disease are taken into account, the range of healthy ALT values is considerably reduced.10. By these standards, 71% (36/51) of the patients reported on here would have had elevated ALT, as do most patients with NAFLD seen in hepatology clinics. The argument that, for analytical reasons, the normal reference value used in the institutional laboratory that performed the tests validates the “normality” of ALT needs also to be challenged. We showed that when ALT activity was measured on the same blood sample in nine different laboratories, interlaboratory variability was high when the result was expressed as upper limit of normal, with some laboratories considering it normal and others high, while this variability was significantly reduced when expressed as IU/L.11 Thus, for both epidemiological and analytical reasons, using a “normal” value of ALT, especially as high as the one proposed here, is questionable. It would be more appropriate, especially in NAFLD patients, to express the results in IU/L rather than as upper limit of normal and to compare the results to updated healthy ranges.10
Unquestionably, there is a group of asymptomatic patients with NAFLD who repeatedly have very low levels of ALT. Whether and in what proportion these patients have significant histological lesions remains to be established.