In a recent letter1 concerning the applicability of noninvasive fibrosis markers in patients coinfected with human immunodeficiency virus/hepatitis C virus, Schiavon et al. indicated that serum aminotransferase (AT) levels, expressed as international units per liter (U/L), exhibited considerable interlaboratory variability. To correct this insufficiency, they suggested expressing alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels as multiples of the upper limit of normality when these markers are integrated into FIB-4, a multivariable score of fibrosis proposed by Sterling et al.2 Interlaboratory variability for ALT and AST results was very large when patients' results were expressed in U/L. That created a problem for the applicability of these markers for the fibrosis evaluation as observed by Schiavon et al.1 The expression as multiples of upper reference limit proposed by Schiavon et al. did not solve this problem but increased the scattering of ALT and AST patients' results. The latter result suggests that interlaboratory variability was due to 2 additional causes: the use of distinct limits of normality as pointed out by Schiavon et al. and the differences in the analytical procedures used by the laboratories. Indeed, 2 types of procedures are used for the determination of AT, depending on whether the used reagents are supplemented with pyridoxal phosphate (PP) or not. The addition of PP, a cofactor necessary for the expression of an AT activity, determined the sum of the 2 forms existing in serum: the apoenzyme (without PP) and the complex apoenzyme-PP.4 In other words, the 2 kinds of measurement procedures do not exhibit the same analytical specificity, and the procedures supplemented with PP are recommended.5, 6 When the analytical procedures (with a PP supplementation) were calibrated with the same material in each laboratory, patients' results became very similar and not statistically significant. More precisely, median values obtained in 6 laboratories did not vary by more than 10%.3 Reagents containing PP Procedures calibrated with a common material whose ALT and AST values were assigned by the corresponding analytical reference method,5, 6 thus ensuring a metrological traceability, as recommended at the international level.7 The validity of this approach was confirmed in another study performed on patients suffering from hepatitis B.8 These data indicated that a standardization of enzyme measurement, both by the use of PP for reason of analytical specificity and by a common calibration, improved the interlaboratory comparability for enzyme results. In addition, this approach decreased the number of patient misclassifications when ALT was integrated into a multivariable score of fibrosis, such as Fibrotest,9 and made possible the use of the same reference values in all laboratories.3, 9 Interest should be also observed for FIB-4,2 and APRI score,10 which integrates AST level, as well as in Forns' equation,11 which takes into account gamma glutamyl transpeptidase concentration. In conclusion, the use of measurement procedures exhibiting the same analytical specificity coupled with a calibration with an adequate material makes possible the development of multivariable scores containing enzymes. Anne Myara*, Jérôme Guechot , Elisabeth Lasnier , Françoise Imbert-Bismut , Hélène Voitot§, Georges Ferard Ph.D.¶, * Biologie, Hôpital Paris-Saint-Joseph, Paris, France, AP-HP, Biochimie, Hôpital Saint-Antoine, Paris, France, AP-HP, Biochimie, Hôpital Pitié-Salpêtrière, Paris, France, § AP-HP, Biochimie, Hôpital Beaujon, Clichy, France, ¶ Université Louis Pasteur, Strasbourg, France.