Dear Editor,
We read with great interest the article by Song et al. [
1], which systematically compared the characteristics and outcomes of normal-weight and overweight/obese individuals with metabolic dysfunction-associated steatotic liver disease (MASLD) across multiple cohorts. The study provides valuable insights into the reclassification discrepancies between non-alcoholic fatty liver disease (NAFLD) and MASLD definitions, particularly in normal-weight populations. However, several methodological issues need further discussion, as they may significantly impact the interpretation and generalizability of the findings.
First, selection bias in initial inclusion within cross-cohort studies poses a significant challenge to the validity of subsequent findings. The Clinical Data Analysis and Reporting System (CDARS) database comprises patients diagnosed with NAFLD based on International Classification of Diseases, 9th revision (ICD-9) codes from hospital patients, rather than a population-based sample. As shown in Table 1 of Song et al. [
1], the prevalence of diabetes (95.4%) and hypertension (86%) in this cohort was significantly higher than in the general population. This suggests that the sample exhibited a significant number of comorbidities prior to the diagnosis of NAFLD, which could lead to an overestimation of all-cause mortality. Moreover, the mean age of the normal-weight MASLD group in CDARS is 62.9 years, with only 34.9% males, whereas the Korean magnetic resonance elastography (MRE) cohort has a mean age of 51.2 years and 90.6% males. Such demographic and clinical heterogeneity, combined with variations in steatosis diagnostic methods (e.g., ICD coding, controlled attenuation parameter [CAP], MRE, biopsy), introduces spectrum bias and limits the comparability of outcomes across cohorts.
Second, residual confounding and methodological limitations in modeling may obscure the true relationship between body mass index (BMI) and mortality. We appreciate the authors’ adjustments for extrahepatic diseases and suggest that additional clarification regarding potential residual confounding could further strengthen the interpretation. Conditions such as stroke, chronic kidney disease, and potentially unaccounted for malignancies or other wasting diseases may lead to weight normalization and increased mortality risk, rather than MASLD itself. This potential confounding could lead to an overestimation of mortality risk to MASLD itself in normal-weight individuals. Furthermore, while BMI categorization is practical, categorizing a continuous variable would lead to information loss and oversimplify its association with outcomes. From a methodological perspective, the employment of restricted cubic spline analysis as a supplementary method could be insightful for future research, as it models BMI continuously and helps explore potential non-linear associations (e.g., U-shaped relationships), thereby providing a more detailed understanding of the BMI-mortality relationship in MASLD [
2].
Third, waist circumference is a key component of cardiometabolic risk factors in MASLD and a critical indicator of visceral fat deposition [
3]. Although this study applied different waist circumference thresholds for Asian and Western cohorts, additional stratification by sex and race could be highly informative. For instance, the Korean MRE cohort comprised only 9.4% females, while the CDARS normal-weight MASLD group was 65.1% female. This notable sex imbalance raises concerns about whether the results are generalizable across sexes, especially given potential sex-specific risk factors (e.g., postmenopausal metabolic changes). Similarly, the multi-ethnic National Health and Nutrition Examination Survey (NHANES) cohort lacked racial stratification, potentially leading to misclassification bias if waist circumference thresholds are not equally applicable across racial groups. Additionally, according to Supplementary Table 4 of Song et al. [
1], the waist circumference data was missing for as much as 37.6% of the CDARS cohort. We therefore suggest that subgroup stratifications by sex and race are key to accurately interpreting the findings for specific populations
In summary, we commend the authors for their timely and important study on the implications of MASLD redefinition, especially in the clinically relevant normal-weight population. However, methodological considerations, including potential selection bias, residual confounding, and limited subgroup analyses, suggest that the results should be interpreted with caution. It is important to avoid overgeneralizing the conclusion that normal-weight MASLD uniformly confers excess mortality risk across all heterogeneous populations. To further advance this field, future research should place greater emphasis on prospective designs, consistent inclusion criteria, and rigorous control of confounding factors.
FOOTNOTES
-
Authors’ contribution
Qingshang Li wrote the manuscript; Xiaona Hu, Yuanwen Chen, and Zhijun Bao provided methodological and revised the manuscript.
-
Conflicts of Interest
No conflicts of interests for this study.
Abbreviations
controlled attenuation parameter
Clinical Data Analysis and Reporting System
International Classification of Diseases
metabolic dysfunction-associated steatotic liver disease
magnetic resonance elastography
non-alcoholic fatty liver disease
National Health and Nutrition Examination Survey
REFERENCES
- 1. Song SJ, Yoon EL, Wong VW, Jo AJ, Wong GL, Lai JC, et al. Normal-weight metabolic dysfunction-associated steatotic liver disease: reclassification, characteristics, and adverse liver outcomes across diverse populations. Clin Mol Hepatol 2026;32:646-660.
- 2. Lopez-Ayala P, Riley RD, Collins GS, Zimmermann T. Dealing with continuous variables and modelling non-linear associations in healthcare data: practical guide. BMJ 2025;390:e082440.
- 3. Choe HJ, Després JP, Wilding JPH, Ryan DH, Lim S. Clinical and preclinical obesity in Korean adults from 2014 to 2023. Diabetes Metab J 2026;50:127-138.
Citations
Citations to this article as recorded by
