The present study showed that hs-CRP values were associated with biopsy-proven liver steatosis, NASH, and fibrosis. A hs-CRP cutoff < 7 mg/L had a reasonable specificity; that is, it allows for correct reporting of approximately 80% of patients without documented liver steatosis or fibrosis at biopsy as true negative.

Adipose tissue dysfunction associated with obesity, adipocyte hypertrophy, and hyperplasia causes a low-grade systemic inflammation characterised by increased pro-inflammatory molecules. Accordingly, many of our patients showed increased hs-CRP levels (78% above 2 mg/L, 70% above 3 mg/L, and 38% above 10 mg/L). Previous studies reported that low-grade inflammation and increased circulating concentrations of proinflammatory cytokines are associated with visceral adipose tissue [28]. Similarly, in our studied population, hs-CRP values and body fat percentage were positively correlated.

NAFLD is the result of hepatic lipid accumulation due to increased free fatty acids (FFA) derived from one or more of the following mechanisms; (1) increased lipolysis in adipose tissue, (2) reduced FFA oxidation, (3) enhanced de novo hepatic lipogenesis, and (4) reduced hepatic very-low-density lipoprotein-triglyceride secretion [4]. Multiple insults (adipokine secretion, inflammation, lipotoxicity, deregulation in glucose and lipid metabolism) triggered by liver lipid accumulation might act synergistically to determine the progression from NAFLD to NASH/cirrhosis [4, 29]. The progression to NASH is associated with systemic inflammation and underpinned by many processes, such as endoplasmic reticulum stress, adipocytokine deregulation, mitochondrial dysfunction, alterations in innate immunity, and toll-like receptor signalling, and intestinal dysbiosis. These processes lead to the accumulation of extracellular matrix, fibrosis development, and liver function deterioration [4, 29]. These pathological conditions (dysbiosis, inflammation, insulin resistance, adipocytokine deregulation) are common in severe obesity. Accordingly, almost 60% of our patients showed some degree of histologically-diagnosed liver damage.

Inflammation can potentially be perpetuated through a vicious cycle that causes further hepatocyte damage. Consistently, we found a significant association between the acute-phase protein hs-CRP and every degree of biopsy-proven liver damage (steatosis, NASH, and fibrosis), in line with the literature [5, 7,8,9, 11,12,13,14,15,16,17,18,19]. Intriguingly, these associations remained significant after adjusting for waist circumference and the insulin resistance index (HOMA-IR), suggesting that hs-CRP is a marker of liver damage independent of obesity and insulin resistance. However, a study carried out in individuals with type-2 diabetes did not show an association between liver damage and CRP [22]; indeed, treatments, or the underlying conditions, might have masked the relationship. Other previous studies have identified different inflammatory markers (i.e., VCAM-1, IL-6, IL-8, chemokines) with better performance than hs-CRP in distinguishing advanced fibrosis from milder stages [6, 18, 20, 21, 23]. However, these markers are not assayed in all laboratories, are rarely used in clinical practice, and are very expensive.

This is the first study that simultaneously compared different methods to define the diagnostic accuracy of hs-CRP in the detection of liver disease. In addition to the most widely used technique in clinical practice (ultrasonography) and the ‘gold standard’ (liver biopsy), 2D-SWE was employed. This diagnostic tool was recently shown to have good accuracy in hepatic function and fibrosis assessment in severely obese candidates for bariatric surgery [30, 31].

In our studied population, hs-CRP values showed an acceptable specificity towards liver steatosis or fibrosis. This association may be attributed to inflammatory process and insulin resistance along with dyslipidemia, excess weight, and increased liver enzymes [32, 33].

These results are intriguing considering the difficulty of obtaining morphological liver assessments in patients with severe obesity [34]. However, liver biopsy is a high-cost, invasive technique, afflicted by sampling error, inter-operational variability, and complications because a thick subcutaneous layer of fat and poor mechanical beam transmission make non-invasive imaging methods less accurate in the presence of severe obesity [34, 35]. Furthermore, the measurement of other non-invasive biomarkers (e.g., blood transaminases) provides non-specific indications regarding the type and extent of liver damage. To overcome the limitations of these methods, simple non-invasive fibrosis scores employing readily available laboratory parameters have been developed (i.e., AST/ALT ratio, APRI, FIB-4, NFS, BARD, etc.) [36]. However, their usefulness in predicting liver fibrosis has recently been questioned, especially in severe obesity [36, 37].

In our patients, hs-CRP showed a lower specificity towards NASH prediction, in line with two other studies, showing that this protein could not differentiate steatohepatitis from simple steatosis [9, 14]. Therefore, inflammation was hypothesised to precede liver steatosis, and the common pathogenesis might decrease the discriminatory role of hs-CRP for steatohepatitis [4, 9]. However, CRP mRNA expression in the liver is significantly elevated in NASH patients compared to patients with simple steatosis, thus, suggesting a pathogenetic implication of this protein in steatohepatitis [7].

Finally, around 20% of our patients without the disease were incorrectly identified to test positive. Genetic/epigenetic/environmental factors, and several protective mechanisms, including liver triglycerides, increased adiponectin levels, and hyperleptinemia, might potentially be involved in protecting the liver from toxic lipid insults [4]. The pathogenesis of NAFLD is a complex clinical course involving different processes and pathways that need to be better characterised to develop new non-invasive markers of NASH and fibrosis, which can simultaneously take into account the numerous implicated factors in the pathogenesis of NAFLD.

NAFLD is associated with an increased risk for cirrhosis, cancer, and cardiometabolic diseases [1,2,3, 13, 38]. Correct and accurate evaluation of the histological stage is crucial since fibrosis represents an adverse prognostic factor. Therefore, it is essential to identify non-invasive biomarkers for predicting the severity of NASH and liver fibrosis. Hs-CRP is a readily-available and straightforward marker but showed low sensitivity in our patients with severe obesity and other cohorts from previous studies [10, 19, 21, 23]. From a clinical perspective, there probably is no single biomarker alone that can differentiate the severity of the liver disease. New insights into the pathogenesis and progression of NAFLD will help search for a combination of predictive non-invasive biomarkers, which must be validated as accurate indicators for both disease prognosis and the response to treatment.

It is worth noting that the availability of histology may be considered a strength of the study because we obtained liver samples during the surgery, which were larger than percutaneous samples (50 > mm vs. > 25 mm). Furthermore, an expert surgeon obtained the liver specimen under gross examination during surgery to take the most optimal specimen. One limitation to our study was that our sample size was small; specifically, only a few patients showed fibrosis at the time of the histological exam/assessment. This may be due to their relatively young age since age is one of the most important variables influencing NASH onset and its development [37]. The observational nature of our study also precludes interpretations of causality.