Impact of bariatric surgery on non-alcoholic fatty liver disease: an integrative review

The present study aims to demonstrate the results of the impact of bariatric surgery on non-alcoholic fatty liver disease through secondary sources available in the literature. A literature review was carried out with the descriptors "Non-alcoholic Fatty Liver Disease" OR "NAFLD" AND "Bariatric Surgery"; “Doença Hepática Gordurosa não Alcoólica” OR “DHGNA” AND “Cirurgia Bariátrica” in the databases: Latin American and Caribbean Health Sciences Literature, through the Virtual Health Library, and Medical Literature Analysis and Retrieval System Online (Medline), through PubMed. Cohort-type studies addressing the effects of bariatric surgery on non-alcoholic fatty liver disease published in the last 5 years were included in the review. Thirty-one studies were identified, of which 10 were excluded for not performing postoperative evaluation and 8 were excluded for not containing the methodological criteria, totaling 13 studies. The current literature presents data suggestive of the association between bariatric surgery and the regression of non-alcoholic fatty liver disease, however, the studies observed present wide variations in methodological applications and samples with heterogeneous profiles, which makes it difficult to generalize the results. More studies are needed so that it is possible to document with greater evidence and reproducibility the action of different types of bariatric surgery on the rate of this liver disease regression. Research, Society and Development, v. 10, n. 13, e397101321437, 2021 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v10i13.21437 2


Introduction
Non-alcoholic Fatty Liver Disease (NAFLD) is defined as an excessive accumulation of fat in the liver, regardless of excessive alcohol consumption or any other secondary cause (Milić, Lulić & Štimac, 2014). It is characterized by an increase in intrahepatic lipid content (steatosis) with or without inflammation and fibrosis (steatohepatitis) (Fabbrini, Sullivan & Klein, 2010). Thus, it comprises a wide spectrum of histological characteristics ranging from simple steatosis to steatohepatitis, fibrosis, and cirrhosis (Milić, Lulić & Štimac, 2014). Once cirrhosis is established, patients are at high risk of developing hepatocellular carcinoma (HCC), which is related to an increase in mortality due to liver causes (Bhala, Younes & Bugianesi, 2013).
NAFLD is considered the most common liver disease worldwide (Milić, Lulić & Štimac, 2014), with percentages ranging from 25 to 30% in the general population (Bhala, Younes & Bugianesi, 2013). It is commonly considered the hepatic manifestation of metabolic syndrome (MS), grouping with obesity and diabetes, for it reflects shared pathogenic factors (Bugianesi, McCullough & Marchesini, 2005). Given the close association with MS, the entire spectrum of NAFLD occurs mainly in patients with obesity (60-95%), type 2 diabetes mellitus (DM2) (28-55%) and hyperlipidemia (27-92%) (Bugianesi, McCullough & Marchesini, 2005). Among populations with obesity, MS and DM2, the prevalence of NAFLD is much higher, ranging from about 50% to 90%. Considering that the prevalence of obesity and overweight has increased to epidemic levels over the last decades, NAFLD has become one of the greatest challenges for public health (Cazzo, Pareja & Chaim, 2017).
The classical pathogenic pathway leading to NAFLD has been described as the "two-hit hypothesis". This theory states that damage to liver tissue begins with lipid accumulation, a consequence of a sedentary lifestyle, high-fat diet and obesityfirst hit. This lipid accumulation works as a sensitization factor for the activation of inflammatory mediators and fibrogenesissecond hit. Secondary aggressions would lead to steatohepatitis and fibrosis (Day & James, 1998). However, the several molecular and metabolic changes that occur in NAFLD cannot be fully explained by the "two-hit hypothesis". Therefore, the "multiple-hit hypothesis" was then considered to explain the pathogenesis of NAFLD. This hypothesis considers multiple factors that, together, act in genetically predisposed individuals to induce the development of the disease.
The hypothesis includes nutritional factors, hormones secreted by adipose tissue, existence of insulin resistance, changes in the intestinal microbiota, and influence of genetic and epigenetic factors.
Changes in the intestinal microbiota, as observed in obesity and insulin resistance, have consequences for both energy homeostasis and systemic inflammation secondary to endotoxemia. At the hepatic level, high levels of free fatty acids can result in stress in the endoplasmic reticulum and mitochondrial dysfunction with consequent activation of inflammatory responses. In addition, genetic factors may explain more progressive courses of the disease in some individuals, comparatively.
Herewith, it is understood that several parallel processes contribute to the development of steatosis and hepatic inflammation (Buzzetti, Pinzani & Tsochatzis, 2016).
Liver biopsy is currently the gold standard for the diagnosis of NAFLD (Chalasani et al, 2018). When at least 5% of hepatocytes have steatosis, patients can be defined as having NAFLDin an appropriate clinical context. When, in addition, there is lobular inflammation and ballooning of liver cells, the lesion is usually defined as steatohepatitis (Milić, Lulić & Štimac, 2014). The analysis of samples collected by biopsy are commonly evaluated using the Non-Alcoholic Liver Disease Activity Score (NAS), which takes into account histological investigation of steatosis, lobular inflammation, and hepatocellular balloonism (Kleiner et al, 2005). The identification of the histological stage is crucial for the long-term prognosis because the population with NAFLD is associated with a significantly higher mortality, compared to the general population, mainly due to cardiovascular and hepatic complication (Cazzo, Pareja & Chaim, 2017).
The enzymes dosage of liver parenchyma may also reflect the severity of the disease. Steatosis correlates directly with liver function tests, especially aspartate transaminase (AST) and alanine transaminase (ALT) values (Ruiz-Tovar, Alsina & Alpera, 2017). Still regarding the non-invasive methods of NAFLD evaluation, the development of scores to evaluate liver diseases are an important tool for the analysis of laboratory and clinical variables of patients (Cazzo, Pareja & Chaim, 2017).
Currently, the Non-Alcoholic Liver Disease Fibrosis Score (NAFLD-FS) is the most widely used score and can be easily calculated based on six available variables: age, body mass index (BMI), hyperglycaemia, platelet count, albumin, and AST/ALT1 ratio (Angulo et al, 2007). Nevertheless, although they do not cause morbidity and are easily evaluated by means of clinical and laboratory variables routine, the results of this score do not replace the diagnostic value of liver biopsy. The NAFLD-FS, then, has been adequate for population-based studies and for clinical screening and follow-up (Simo et al, 2014), especially in populations known to be more exposed to this disease, such as patients with MS, diabetes, dyslipidaemias, and obesity (Ruiz-Tovar, Alsina & Alpera, 2017).
Clinically, obesity reflects a generalized pro-inflammatory state with a high risk of metabolic comorbidities, such as NAFLD, which are highly influenced by the distribution of adipose tissue. Evidence suggests that visceral adipose tissue is directly associated with the development and progression of NAFLD. The most important pathological mechanisms involve increased secretion by visceral adipose tissue of pro-inflammatory cytokines, adipokines, and release of free fatty acids in the portal system and systemic circulation, causing dyslipidaemia and insulin resistance (Milić, Lulić & Štimac, 2014). Along with Research, Society andDevelopment, v. 10, n. 13, e397101321437, 2021 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v10i13.21437 insulin resistance, obesity contributes significantly to the initial accumulation of fat in the hepatocyte and to the progression of simple steatosis to non-alcoholic steatohepatitis, cirrhosis related to steatohepatitis, and hepatocellular carcinoma. From the clinical point of view, obesity increases morbidity and mortality when associated with NAFLD, due to cardiovascular and hepatic mortality (Polyzos, Kountouras & Mantzoros, 2017). Furthermore, the disease is an additional risk factor for this group of patients, especially for the possibility of progression to severe forms of hepatic fibrosis and hepatocellular carcinoma (Ruiz-Tovar, Alsina & Alpera, 2017). There is an established association between obesity and the incidence of NAFLD, as well as increased severity (Polyzos, Kountouras & Mantzoros, 2017).
In therapeutic aspects, weight loss is considered the main factor in the prevention and treatment of obesity and NAFLD (Polyzos, Kountouras & Mantzoros, 2017). In this context, bariatric surgery is recognized as an effective strategy in the induction and maintenance of weight loss in patients eligible for the procedure. Individuals with DM2, BMI between 30 Kg/m2 and 35 Kg/m2, and no response to clinical treatment may have indication for bariatric surgery. Similarly, patients with BMI higher than 35 Kg/m2, with diseases associated with obesity or those with BMI higher than 40 Kg/m2considered morbid obesityare also eligible for surgery (Abeso, 2016). This intervention plays a significant role in the natural history of NAFLD and usually leads to rapid changes in its clinical progression (Ruiz-Tovar, Alsina & Alpera, 2017). The procedure has the potential to induce great weight loss and improve MS and DM2 characteristics. Additionally, the mechanisms involved in improving obesity and DM2 after bariatric surgery probably play important roles in the resolution of metabolic and inflammatory lipid abnormalities present in the pathophysiology of NAFLD.
Bariatric surgery promotes changes in three crucial metabolic areas that influence the spectrum of NAFLD: i) improvement of glucose homeostasis; ii) improvement of lipid metabolism; and iii) reduction of inflammatory activity. Therefore, bariatric surgery can reverse pathological changes in those patients' liver (Laursen et al, 2019). In addition, the International Federation for the Surgery of Obesity and Metabolic Disorders (IFSO) states that bariatric surgery induces significant reversal or improvement of NAFLD. Considering the overall impact of bariatric surgery on obesity and related comorbidities, this option of treatment should be offered to the group of individuals who meet the current indication criteria, as a potentially effective therapy to minimize the risks of the disease. However, IFSO also concluded that, despite the solid evidence available, complementary studies to reach final conclusions about the real effect of bariatric surgery as a therapeutic option for patients with obesity and NAFLD are still required (De Luca et al, 2016).
Although the overall impact of bariatric surgery on NAFLD is positive, the ideal surgical technique still needs to be determined, as well as the long-term effects and ways to maintain the benefits achieved (Ruiz-Tovar, Alsina & Alpera, 2017).
Recently, Vertical Sleeve Gastrectomy (VSG) and Roux-en-Y Gastric Bypass (RYGB) have been the most commonly performed bariatric procedures worldwide, leading to a significant, sustainable weight loss and improvement of metabolic comorbidities (Khorgami et al, 2017). Some authors state that, although both RYGB and VGS are effective in restoring liver function and structure in obese patients with NAFLD, the RYGB approach demonstrated an advantage over VGS in the disease regression (Froylich et al, 2016). On the other hand, some authors advocate equality among the several methods and state that, in patients with NAFLD and biopsy-proven non-alcoholic steatohepatitis, preoperative abnormal liver function tests are normalized in the vast majority of VGS and RYGB patients at the end of the first year after surgery, providing evidence that both procedures are equally effective in improving liver function (Cherla et al, 2020).
Considering the growing worldwide epidemic of obesity and the fact that this population represents a risk group for NAFLD and to develop advanced stages of liver disease and cardio metabolic disease, the therapeutic approach of these patients is an important priority for health care and clinical research. Notably, bariatric surgery is recognized as an effective strategy in the induction and maintenance of weight loss. Although there is no formal indication of this procedure for the Research, Society and Development, v. 10, n. 13, e397101321437, 2021 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v10i13.21437 5 treatment of this liver disease, weight reduction is considered the main factor in the prevention and treatment of NAFLD.
Therefore, this review aims to evaluate studies on the impact of bariatric surgery on NAFLD, through secondary sources available in the scientific literature.

Methods
The integrative review is an important method for Evidence-Based Practice. The following steps were performed: elaboration of the guide question, research in the literature, data collection, critical analysis of the included studies, discussion of the results and the integrative review (Whittemore & Knafl, 2005;Souza & Silva, 2010). For the literature review, the following descriptors were considered: "fatty liver"[All Fields] OR "NAFLD"[All Fields] AND "bariatric surgery"[All Fields]; "Non-alcoholic fatty liver disease" OR "NAFLD" and "bariatric surgery". The research was conducted in the database of Latin Other papers were used for the contextualization of the present study.

Literature Review: Studies and Observed Results
We identified 31 studies trough this literature review. All of them were carefully analysed and 18 were excluded because they met at least one of the exclusion criteria, resulting in 13 studies that comprised the integrative review ( Figure 1).
Of this total, ten were prospective and three retrospective cohort studies. The main characteristics and results of the cohort studies analysed were summarized in Table 1. Altogether, they counted a sample of 507 patients, with a mean of 50.7, median of 41.5 and standard deviation of ±28.19. The study with higher number of patients was composed of a sample of 100 patients, and the lowest with 25. NAFLD evaluation was performed by invasive methods, by intraoperative biopsy (5 studies) and non-invasive methods, by enzymatic evaluation through standardized scores (5 studies). The Vertical Sleeve Gastrectomy technique was performed in 7 studies; in 3, its results were compared with those obtained in patients who underwent surgery performed with Roux-en-Y Gastric Bypass technique, which was used in isolation in 2 other studies. There is also one study that evaluated intragastric balloon and endoscopic sleeve gastroplasty techniques. All studies were classified with evidence level 2B, following the parameters established by the Oxford classification. The mean regression rate of NAFLD among the studies was 66.49%, with a standard deviation of ±22.48 and median of 68.30. The maximum value in the regression rate was 94.2% and the minimum was 12%.
Additionally, samples collected through liver biopsy were evaluated in most studies (3) using the Non-Alcoholic Fatty Liver Disease Activity Score (NAS). Other methods were also used, such as the Brunt classification (1) and the evaluation parameters established by the American Association for the Study of Liver Disease (AASLD) and the European Association for the Study of the Liver (EASL).
Among the non-invasive evaluation methods identified, the Non-Alcoholic Liver Disease Fibrosis Score (NAFLD-FS) was used in all studies, being combined with other parameters, including ALT/AST ratio, APRI, BARD score, FLI, HSI and FIB-4. Table 2 shows the mean scans of the NAFLD FS values obtained before and after bariatric surgery in the various studies. N=Number of patients; SD= Standard Deviation; NAFLD= Non Alcoholic fatty liver disease; NAFLD Score < -1.455 = F0-F2 (with no advanced fibrosis), NAFLD Score -1.455 -0.675S = undetermined score, NAFLD Score > 0.675 = F3-F4 (with advanced fibrosis). Source: Authors.

Liver biopsy in the evaluation of NAFLD
Praveen et al. (2015) conducted a study (n = 30) in which intraoperative biopsy was performed after 6 to 8 months of bariatric surgery. NAFLD was timed by the NAS score. In the intraoperative biopsy, all patients had steatosis. In the second biopsy, 19 had resolution of steatosis. In 11 of them, there was a reduction in the degree of steatosis with significant statistical differences. Additionally, the NAS biopsy score performed 6-8 months after bariatric surgery showed a statistically significant reduction in relation to intraoperative biopsy. Regarding the surgical techniques employed, both (RYGB and VSG) showed reductions in NAS score, with no statistically significant relationship between the reduction of NAS score and the surgical technique employed.
In another study, Von Schonfer et al. (2018) conducted a retrospective cohort and assessed the evolution of NAFLD in 53 patients after bariatric surgery, comparing the two main surgical methods used (RYGB and VSG). In intraoperative biopsy, a positive and statistically significant relationship was observed between the NAS score and the presence of MS. After surgery, the patients presented a significant reduction in weight and BMI, but no significant differences were found between the techniques performed and the percentage of weight loss. There was a resolution for metabolic syndrome in 83% of patients, with a statistically significant reduction and no difference between the methods. In the histological analysis, there was a significant improvement in the NAS score, which showed an average improvement of -2. In the first biopsy, 21% of the patients presented in the most severe stages of NAFLD, with NAS between 5 and 8; after surgery, only 2% continued in these stages. This improvement was not related to the method of surgery, the sex of the patient or liver enzymes when these variables were analysed individually. among VSG patients, there was improvement in all NAS components, but only the evolution of fibrosis was statistically significant. When taking both methods into account, the mean NAS value went from 3.6 ± 1.8 to 1.2 ± 1.5. In absolute values, surgery by RYGB presented better results than VSG, but without statistical differences. Chaim et al. (2020) conducted a prospective study with 30 patients. With regard to NAFLD, in the intraoperative biopsy, a prevalence of 50% was found: 33.33% of steatohepatitis and 16.67% of steatosis. When analysing postoperative biopsy, only 16.67% of the patients had NAFLD, 10% had steatohepatitis and 6.67% had steatosis. When comparing both biopsies, a significant reduction was observed in the values of ALT, total protein, triglycerides, plasmatic albumin, total cholesterol, and gammaglutamyl transferase (GGT). In addition, BMI decreased significantly from an average of 37.9 ± 2.21 to 25.69 ± 3.79 kg/m2.
In a retrospective cohort study, Esquivel et al. (2018) analysed the biopsy of 43 patients at the time of surgery using the VSG method and a year after the procedure. Importantly, all subjects presented statistically significant reduction in the degree of NAFLD. There was also a significant decrease in liver enzymes, BMI, glycemia, LDL-c, and total cholesterol.
However, no significant increase in HDL-c was found. Kalinowski et al. (2017) conducted a randomized clinical trial with 66 patients, analysing the evolution of liver status through NAS and liver function tests. In this study, 53% of the patients had steatohepatitis at the time of surgery, and 100% had some degree of NAFLD. A significant relationship was found between NAS value and age, female gender, and serum concentrations of AST and ferritin. Conversely, in the evaluation after one year, there was a significant decrease in weight.
ALT, AST, and GGT decreased significantly only in patients who underwent VGS. Albumin, though, reduced significantly only in RYGB patients.

Non-invasive evaluation of NAFLD
In a retrospective cohort study, Ledoux et al. (2019) evaluated NAFLD in 554 patients before and after one year of bariatric surgery, through metabolic and ultrasound parameters. Interestingly, weight loss was significantly higher in patients who were operated using the RYBG method in relation to the VGS method. RYBG surgery showed better evolution in relation to fasting glycemia, LDL-c, and C-reactive protein inflammation marker (CRP). On the other hand, in liver function tests, such as alkaline phosphatase and transaminases, there was a greater reduction in patients operated by VGS. The gamma-glutamyl transferase values showed a reduction in both VSG and RYBG, with no significant differences between the methods.
Abdominal US showed a reduction in the number of patients with hepatic steatosis, but without significant differences between the VSG and RYBG methods.
In another study, Cazzo, Jimenez, Pareja e Chaim (2015) evaluated NAFLD at the time and one year after bariatric surgery, in a sample of 63 patients. To determine the efficiency of non-invasive methods for diagnosing NAFLD, the researchers performed an intraoperative biopsy and compared it to the NAFLD fibrosis score. They observed an overall accuracy of 87.3% for the non-invasive method as a diagnosis, with sensitivity of 97% and specificity of 75.9%. Compared to the evaluation after one year of the procedure, there was a significant decrease (p-value <0.05) of the NAFLD fibrosis score, which decreased from an average of 1,142 ± to 1,261 to 0.006 ± 1,027. In addition, surgery led to a resolution of NAFLD in 55% of patients. This result had a significant association with female gender, percentage of weight lost, post-surgical BMI, platelets after surgery, and resolution of DM2.
Ruiz-tovar e Zubiaga (2019)  HIS, and NAFLD-Fibrosis Score) showed a significant reduction (p-value <0.05), while FIB-4 did not present relevant alterations. There was also a significant reduction (p-value <0.05) in glucose and insulin values, but not in Hb1Ac. It was also important to highlight that 50% of patients who had DM2 were able to stop the medication after one year of the procedure.
Finally, Nascimento et al. (2015) conducted a study in which they compared the NAFLD score for fibrosis preoperatively and postoperatively, in six months and one year after bariatric surgery, in patients operated in the private and public health system. They also compared the results between the health systems. The authors observed that 95% of the patients presented a decrease in the absolute value of the evaluated score. Among those in the public system, the mean value increased from -0.6845 ± 2.17 to -1.6898 ± 1.91, with statistical difference (p-value <0.05), one year after surgery. In the private system, the evolution went from -2.1212 ± 1.65 to -3.04 ± 0.93, but with no statistical significance. The variation in BMI and weight before and after one year of surgery was significant for both groups, while there were no differences in changes in AST, ALT and albumin values.

Conclusions
The current literature presents data suggestive of the association between bariatric surgery and the regression of NAFLD and corroborating the potential of this surgical procedure as a formal indication for the control of this pathological process. Nonetheless, the studies present herein showed large variations in methodological applications and samples with heterogeneous profiles, which makes it difficult to generalize the results obtained.
Importantly, studies that quantify the reduction in NAFLD or its resolution after bariatric surgery, using invasive and non-invasive evaluation criteria, are scarce in the scientific literature. Therefore, this points to the need for further clinical studies, so that the evaluation and documentation of the action of different types of bariatric surgery on the regression rate of NAFLD, with greater evidence and reproducibility, might be possible.