Authors’ reply

We would like to thank Smyk and Krawczyk1 for their valuable comments on our recently published work on insulin resistance markers and their association with nonalcoholic fatty liver disease (NAFLD) in individuals with type 1 diabetes mellitus (T1DM).2 The authors underline the importance of choosing appropriate controlled attenuation parameter (CAP) cutoff values to detect NAFLD.

Indeed, there has been a knowledge gap regarding reliable values of CAP to diagnose NAFLD in patients with T1DM; thus, data from the general population were extrapolated and used. Based on a meta-analysis by Cao et al,3 including adults with NAFLD or nonalcoholic steatohepatitis, it can be concluded that mean CAP cutoff values for detecting steatosis grade 1 are 254.5 (range, 233.5–275) and 278 (range, 271–285) dB/m for transient elastography probe M and XL, respectively. Of note, individuals without obesity (body mass index [BMI] <⁠30 kg/m2) are characterized by lower mean CAP cutoff values (251.2 [range, 233.5–275] dB/m) than those with a BMI greater than or equal to 30 kg/m2 (281.9 [range, 261–304] dB/m).3 The median (IQR) BMI in our study group was 24.5 (21.8–28.2) kg/m2, which justifies the implementation of a lower cutoff value.

In a valuable work by Mertens et al4 published in April 2023, the authors elegantly assessed the accuracy of noninvasive tools, as compared with 1H magnetic resonance spectroscopy (MRS). The optimal CAP cutoff for detecting NAFLD, regardless of the probe, was found to be greater than or equal to 270 dB/m, with 0.90 sensitivity and 0.74 specificity.

In view of the above, we admit that the CAP values greater than 238 dB/m, which we used to diagnose NAFLD, were relatively low and could result in overdiagnosis of the disease. If we had used the value based on the American Association for the Study of Liver Diseases Guidelines (≥288 dB/m),5 as suggested by the authors of the commentary,1 or the value proposed by Mertens et al4 (≥270 dB/m), the percentage of individuals with NAFLD in the analyzed group would have been 18% and 26%, respectively, which is undoubtedly lower than that in the published article (43%).1 The gold standard would be to compare the elastography results with either MRS or liver biopsy findings; however the latter data were not available for the analyzed group.