Does reference value of “Normal Liver Stiffness” using fibroscan have a regional variation?

Abstract

We read with interest the recently published article by Das et al.,1 where the investigators have reported a U-shaped distribution of liver stiffness measure (LSM) among healthy subjects categorized as per body mass index (BMI), and proposed 8.5 KPa as the upper limit of normal (ULN) of LSM in healthy Indian subjects. Such observations can have significant implications in clinical practice. This is a well-conducted study; however, the investigators' interpretation about U-shaped distribution of LSM is not supported by strong data. The mean LSM was similar over a broad range of BMI (18-29.9 kg/m2), and higher values of LSM at one extreme were based on data of obese persons, which were only 4 (<1%) in number. Furthermore, as per the Asian standard of BMI categories, the distribution pattern of LSM would not have been U-shaped. Although underweight subjects had significantly higher LSM values than healthy and preobese subjects, the mean difference in absolute value was only 0.5 KPa, suggesting that the body-size effect is not perceptible using the current FibroScan machines. The ULN of LSM has been reported to vary from 5.3 to 7.0 KPa in various studies,2, 3 including one study based on histopathology.2 Thus, 8.5 kPa as the ULN of LSM seems too high for a healthy liver in any given population across the world. Such a high cutoff may erroneously cause the exclusion of healthy subjects or patients who would require further evaluation. Though this value corresponded to the 95th percentile, the dispersion of data (mean, 95% confidence interval) revealed that 414 of 418 healthy subjects actually had LSM values within 6.5 KPa, as shown in Fig. 2 of the Das et al. study.1 Notably, sufficient investigations were not done to exclude potential underlying liver disease in healthy subjects with high LSM. Also, approximately 16% of LSM results are known to be unreliable. In our experience, in a cohort of 445 healthy adult subjects, the mean LSM value was 5.1 ± 1.1 KPa, and the 95th percentile value was 7.07 KPa. LSM values increased with increasing BMI categories (4.1 ± 0.7, 5.08 ± 0.6, and 6.08 ± 1.2 KPa in healthy BMI, overweight, and obese subjects, respectively).3 None of our patients belonged to the underweight category; hence, the distribution of LSM cannot be compared with the present study. To conclude, the present study results seem to have limited external validity. However, the background population and properly validated regional data are important while interpreting LSM using FibroScan. Ramesh Kumar M.D., D.M.*, Manoj Kumar Sharma M.D., D.M.*, Shiv Kumar Sarin M.D., D.M.*, * Department of Hepatology, Institute of Liver and Biliary Sciences, New Delhi, India.