Nonalcoholic Fatty Liver Disease and Bariatric Surgery
The most prominent histopathological feature of NAFLD is the deposition of the lipids in the hepatocytes. The diagnosis of steatosis is made when lipid deposition is visible in more than 5% of hepatocytes. NASH is diagnosed when, in addition to hepatic steatosis, both inflammatory infiltrates as well as ballooning and liver cell injury are present. Current discussions among pathologists include differentiation of the hepatic pathology of NASH with a concern for its progression, interobserver variation in the liver biopsy and appropriateness of the current nomenclature. For example, Brunt has developed criteria to grade NASH using separate assessments for necroinflammatory lesions (grade) and fibrosis (stage). More recently, the NAFLD Activity Score (NAS) was developed to provide a numerical score for patients who most likely have NASH. Accordingly, NAS is the sum of the separate scores for steatosis (0–3), hepatocellular ballooning (0–2) and lobular inflammation (0–3), with the majority of patients with NASH having a NAS score of ≥5. This system was developed as a tool to quantify changes in NAFLD during therapeutic trials. Importantly, the cutoff NAS score ≥5 cannot be used as a surrogate of histological diagnosis of NASH that is usually performed by expert-based evaluation of pathologic patterns for NASH.
Liver biopsy is the standard for diagnosis in comparative studies of NAFLD; however, sampling errors can limit its diagnostic accuracy. Even paired liver biopsies taken from the same subject at the same time can sometimes be discordant. One study of paired liver biopsies in individuals undergoing gastric bypass demonstrated a relatively high rate of discordance for portal fibrosis (26%), followed by zone 3 fibrosis (13%) and ballooning degeneration (3%). In another study of paired biopsies, no features displayed high agreement. The discordance rate for the presence of hepatocyte ballooning was 18%, and ballooning would have been missed in 24% of patients had only one biopsy been performed. Importantly, in cases of patients with bridging fibrosis, 35% of the patients would be understaged if they had only one biopsy sample. Other researchers identified that the biopsy concordance rates for necroinflammatory features are the lowest, even for biopsies evaluated by the same histopathologist.
From the observations described above, it is obvious that sample variability inherent to the liver biopsy procedure may substantially hinder interpretation of the results obtained in interventional or natural history studies of NAFLD. In fact, a careful retrospective study performed by Ratziu et al. showed that only improvements in steatosis significantly exceed sampling variability for this feature (47 vs 8%; p < 0.0001), while neither activity grade nor ballooning features improve beyond random fluctuation owing to sampling error. Effects on fibrosis were at best marginal.
In our opinion, in bariatric surgery-based interventional studies, the biopsy-based evaluations of NAFLD-related parameters should be augmented by repeated measurements of the noninvasive biomarkers. Importantly, the current NAFLD diagnostic panels have focused on concerted evaluations of NAFLD features that got summed up in disease-predicting regression models. For the feature-dissecting purposes, a combination of biomarkers, rather than a biomarker panel, should be used, with each biomarker aiming to evaluate a single feature: fibrosis, steatosis or necroinflammatory features of NAFLD. Repeated measurements of these biomarkers may aid in establishing multipoint trends in the physiological state of the liver diseases after bariatric surgery.
Histological Diagnosis of NAFLD
The most prominent histopathological feature of NAFLD is the deposition of the lipids in the hepatocytes. The diagnosis of steatosis is made when lipid deposition is visible in more than 5% of hepatocytes. NASH is diagnosed when, in addition to hepatic steatosis, both inflammatory infiltrates as well as ballooning and liver cell injury are present. Current discussions among pathologists include differentiation of the hepatic pathology of NASH with a concern for its progression, interobserver variation in the liver biopsy and appropriateness of the current nomenclature. For example, Brunt has developed criteria to grade NASH using separate assessments for necroinflammatory lesions (grade) and fibrosis (stage). More recently, the NAFLD Activity Score (NAS) was developed to provide a numerical score for patients who most likely have NASH. Accordingly, NAS is the sum of the separate scores for steatosis (0–3), hepatocellular ballooning (0–2) and lobular inflammation (0–3), with the majority of patients with NASH having a NAS score of ≥5. This system was developed as a tool to quantify changes in NAFLD during therapeutic trials. Importantly, the cutoff NAS score ≥5 cannot be used as a surrogate of histological diagnosis of NASH that is usually performed by expert-based evaluation of pathologic patterns for NASH.
Liver biopsy is the standard for diagnosis in comparative studies of NAFLD; however, sampling errors can limit its diagnostic accuracy. Even paired liver biopsies taken from the same subject at the same time can sometimes be discordant. One study of paired liver biopsies in individuals undergoing gastric bypass demonstrated a relatively high rate of discordance for portal fibrosis (26%), followed by zone 3 fibrosis (13%) and ballooning degeneration (3%). In another study of paired biopsies, no features displayed high agreement. The discordance rate for the presence of hepatocyte ballooning was 18%, and ballooning would have been missed in 24% of patients had only one biopsy been performed. Importantly, in cases of patients with bridging fibrosis, 35% of the patients would be understaged if they had only one biopsy sample. Other researchers identified that the biopsy concordance rates for necroinflammatory features are the lowest, even for biopsies evaluated by the same histopathologist.
From the observations described above, it is obvious that sample variability inherent to the liver biopsy procedure may substantially hinder interpretation of the results obtained in interventional or natural history studies of NAFLD. In fact, a careful retrospective study performed by Ratziu et al. showed that only improvements in steatosis significantly exceed sampling variability for this feature (47 vs 8%; p < 0.0001), while neither activity grade nor ballooning features improve beyond random fluctuation owing to sampling error. Effects on fibrosis were at best marginal.
In our opinion, in bariatric surgery-based interventional studies, the biopsy-based evaluations of NAFLD-related parameters should be augmented by repeated measurements of the noninvasive biomarkers. Importantly, the current NAFLD diagnostic panels have focused on concerted evaluations of NAFLD features that got summed up in disease-predicting regression models. For the feature-dissecting purposes, a combination of biomarkers, rather than a biomarker panel, should be used, with each biomarker aiming to evaluate a single feature: fibrosis, steatosis or necroinflammatory features of NAFLD. Repeated measurements of these biomarkers may aid in establishing multipoint trends in the physiological state of the liver diseases after bariatric surgery.
SHARE
