To date, the only validated treatment for the management of NAFLD is weight loss, generally achieved through lifestyle changes (diet, physical activity) or bariatric surgery. Weight loss of ≥7–10% has been shown to reduce disease progression and improve steatosis, inflammation, and liver fibrosis [7,8]. However, a significant and long-lasting reduction in body weight is difficult to achieve and various dietary regimens often fail both in the long term and in highly motivated patients. Moreover, counter-regulatory mechanisms over time hinder weight loss, resulting in the recovery of body weight [30]. In addition to lifestyle changes, GLP-1 RAs have been shown to lower body weight in a range of 2–7 kg through early satiety and reduction in appetite, thus lowering caloric intake [31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50]. Furthermore, treatment with these drugs reduces cravings for food and increases the pleasure offered by eating a meal [51]. Except for SGLT2 inhibitors, all the other hypoglycaemics drugs (including insulin) either have neutral effect or, more often, cause weight gain [24].

Mechanisms involved in this process are both central and peripheral. In the CNS, serotonin represents a crucial factor in appetite regulation. Through the stimulation of its receptor in the central nervous system, GLP-1 reduces the expression of serotonin 5-HT2A receptors in the hypothalamus, resulting in decreased appetite [52]. This role is independent from serotonin secretion, which is not affected [53]. In particular, in murin models, GLP-1 RAs have demonstrated to prevent meal-initiation-inhibiting activity on arcuate nucleus and inducing meal-ending actions on lateral parabrachial nucleus [24]. It has been hypothesized that liraglutide and semaglutide act through different neurological pathways, providing the rationale for their different efficacy in inducing weight loss [54]. The reduction in appetite is also assisted by peripheral mechanisms. In fact, GLP-1 RAs are able to delay gastric emptying and to slow intestinal motility in the gastro-intestinal tract, which enhances the feeling of early satiety at a central level [45]. However, the long-term receptor stimulation by GLP-1 over time reduces the effects on gastric emptying due to the mechanisms of tachyphylaxis [55].

The main studies on the efficacy of GLP-1 RAs in inducing weight loss are summarized in Table 1. The extent of the reduction in body weight depends on which GLP-1 RAs and which dosage are used. Moreover, the dose used to treat T2DM and obesity may differ. In fact, if the glycemic control plateau is reached with relatively low dosages, weight loss effects may require higher GLP-1 RAs dosages [24]. In contrast, the variability in interindividual response on body weight loss in patients receiving GLP-1 RAs treatment is significantly greater than the variability in glycemic control response [16]. To date, no specific predictors of long-term response to weight loss have been identified.

Table 1. Main evidences on the effect of liraglutide and semaglutide on weight loss.

Year	First Author	Ref.	Sample Size	Study Typology	Evaluated Drugs	Benefit	Results
2009	Astrup	[38]	564 non-diabetic obese	Prospective double-blind	Liraglutide vs orlistat vs placebo	Yes	Liraglutide resulted in greater weight loss than placebo or orlistat
2013	Wadden	[43]	422 obese/overweight	Prospective randomised	Liraglutide vs placebo	Yes	Liraglutide plus diet and exercise maintained weight loss achieved by caloric restriction and induced further weight loss
2015	Pi-Sunyer	[33]	3731 non-diabetic obese	Prospective double-blind	Liraglutide vs placebo	Yes	Liraglutide induced greater weight loss (8.4 Kg) than placebo (2.8 Kg)
2015	Davies	[47]	846 obese/overweight T2DM	Prospective double-blind	Liraglutide vs placebo	Yes	Liraglutide (3.0 mg/day), compared with placebo, resulted in greater weight loss
2016	de Boer	[41]	151 obese insulin-using T2DM	Prospective	Liraglutide or Exenatide	Yes	Liraglutide or Exenatide led to sustained weight reduction and daily insulin dose
2016	Blackman	[44]	359 non-diabetic obese	Prospective double-blind	Liraglutide vs placebo	Yes	Liraglutide (3.0 mg/day) plus lifestyle therapy led to a greater weight loss than placebo plus lifestyle therapy
2016	Armstrong	[56]	52 NASH patients	Multicentre, double-blinded	Liraglutide vs placebo	Yes	Liraglutide 1.8 mg/day led to a greater weight loss than placebo
2017	Halawi	[45]	40 obese patients	Prospective double-blind	Liraglutide vs placebo	Yes	Liraglutide delays gastric emptying and reduces body weight more than placebo
2017	Petit	[57]	68 uncontrolled T2DM patients	Prospective single-center	Liraglutide	Yes	Liraglutide 1.2 mg/day significantly reduced body weight
2018	Frøssing	[58]	72 obese/overweight with PCOS	Prospective double-blind	Liraglutide vs placebo	Yes	Liraglutide results in greater weight loss than placebo
2019	Feng	[59]	85 T2DM and NAFLD patients	Prospective, randomized	Liraglutide vs gliclazide	Yes	Liraglutide results in greater weight loss than gliclazide
2019	Yan	[50]	75 NAFLD and metformin-uncontrolled T2DM patients	RCT	Liraglutide vs sitagliptin vs insulin glargine	Yes	Combined with metformin, both liraglutide and sitagliptin, but not insulin glargine, reduced body weight
2020	Kelly	[40]	251 obese adolescents	Prospective double-blind	Liraglutide vs placebo	Yes	Liraglutide (3.0 mg/day) plus lifestyle therapy led to a greater weight loss than placebo plus lifestyle therapy in adolescent with obesity
2020	Wadden	[42]	282 obese	Prospective double-blind	Liraglutide vs placebo	Yes	Liraglutide (3.0 mg/day) amplifies weight loss due to intensive behavioral therapy
2020	Garvey	[48]	396 obese/overweight and insulin-treated T2DM patients	Prospective double-blind	Liraglutide vs placebo	Yes	Liraglutide (3.0 mg/day) led to greater weight loss than placebo
2017	Blundell	[39]	28 obese	Prospective double-blind	Semaglutide vs placebo	Yes	Semaglutide resulted in lower ad libitum energy intake and greater weight loss than placebo
2018	O’Neil	[31]	957 non-diabetic obese	Prospective double-blind	Semaglutide vs Liraglutide	Yes	Semaglutide (≥0.2 mg/day) resulted in greater weight loss than liraglutide (any dose)
2018	Pratley	[32]	1201 T2DM patients	Prospective randomised	Semaglutide vs Dulaglutide	Yes	Semaglutide 0.5 mg/week resulted in greater weight loss than dulaglutide 0.75 mg/week. Semaglutide 1 mg/week resulted in greater weight loss than dulaglutide 1.5 mg/week
2019	Matikainen	[60]	22 controlled T2DM patients	Prospective single-blind	Liraglutide vs placebo	No	Similar weight loss between 16-week liraglutide 1.8 mg/day and placebo group
2021	Wilding	[34]	1961 non-diabetic obese	Prospective double-blind	Semaglutide vs placebo	Yes	Semaglutide (2.4 mg/week) plus lifestyle intervention resulted in a greater reduction in BMI than lifestyle intervention alone
2021	Davies	[35]	1210 insulin-untreated diabetic obese/overweight patients	Prospective double-blind	Semaglutide 2.4 vs 1 mg/week vs placebo	Yes	Semaglutide 2.4 mg/week resulted in a greater weight loss than Semaglutide 1.0 mg/week or placebo
2021	Wadden	[36]	611 non-diabetic obese/overweight patients	Prospective double-blind	Semaglutide vs placebo	Yes	Semaglutide (2.4 mg/week) plus lifestyle intervention resulted in a greater reduction in BMI than lifestyle intervention alone
2021	Rubino	[37]	902 non-diabetic obese/overweight patients	Prospective double-blind	Semaglutide vs placebo	Yes	Semaglutide (2.4 mg/week) plus lifestyle intervention resulted in a greater reduction in BMI than lifestyle intervention alone
2021	Newsome	[61]	320 obese/overweight NASH patients	Prospective double-blind	Semaglutide vs placebo	Yes	Semaglutide treatment resulted in a greater reduction in BMI than placebo
2022	Rubino	[46]	338 non-diabetic obese/overweight patients	Prospective randomised	Semaglutide vs Liraglutide	Yes	Semaglutide (2.4 mg/week) resulted in greater weight loss than liraglutide (3 mg/day)
2022	Weghuber	[49]	201 obese adolescents	Prospective double-blind	Semaglutide vs placebo	Yes	Semaglutide (2.4 mg/week) plus lifestyle intervention resulted in a greater reduction in BMI than lifestyle intervention alone

Liraglutide has demonstrated to determine significant weight loss, reduce pre-diabetes rate and improve obesity-related risk factors independently from the presence of T2DM [38,40,41,50,56,57,58,59]. After a year of treatment at the daily dosage of 3 mg, it was able to reduce body weight by 8.4 kg [33]. Comparing the results from the various trials [33,38,42,43,44,47,48], the average weight loss is about 3.4–6.1% of the total body weight. In particular, among treated patients, about 2/3 show a weight loss of at least 5% and 1/3 a weight loss of more than 10% [33]. Its anorectic effect starts very early; therefore, in mice, it has significantly reduced appetite within 1 h of administration [53]. To our knowledge, only one study failed to prove significant differences in weight loss between treatment with liraglutide and placebo [60].

Semaglutide is a 2nd generation GLP-1 agonist available both in oral (daily administration) and subcutaneous (weekly administration) formulation [23,24]. The treatment with semaglutide resulted in a lower energy intake and greater weight loss than placebo both in adults [31,34,35,36,37,39,61] and adolescents [49]. In particular, at a dosage of 2.4 mg per week, it has been shown to induce an average weight loss between 9.6% and 17.4%, as well as a reduction in waist circumference and arterial hypertension and an improvement of lipid profile [34,35,36,37,39]. The dose used for the treatment of obesity (2.4 mg/week) is significantly more effective in determining weight loss than that used for the treatment of T2DM (1.0 mg/week) [35]. Furthermore, the ability to induce weight loss is more pronounced for semaglutide than other GLP-1 RAs [32]. For doses equal to or greater than 0.2 mg/day (1.4 mg/week), semaglutide is also superior to liraglutide (at any dose) in causing weight loss [31]. Notably, in the recent STEP 8 randomized clinical trial [46], semaglutide (2.4 mg/week) achieved a mean weight loss of 15.8% after 68 weeks of treatment compared with 6.4% obtained with liraglutide. A weight loss greater than 10% (therapeutic target in NAFLD) was achieved in 70.9% of semaglutide-treated patients and 25.6% of liraglutide-treated patients, with comparable rates of gastrointestinal adverse events [46]. Semaglutide-related weight loss continued until weeks 28 to 44 of treatment and was sustained thereafter [61].

For these reasons, beyond T2DM, liraglutide and semaglutide have also been approved for the treatment of obesity in diabetic and non-diabetic patients [16,17]. The maximum approved dose of liraglutide for the treatment of obesity is 3 mg/day, whereas the maximum dose for the treatment of T2DM is 1.8 mg/day. The maximum dose of semaglutide used for the treatment of obesity is 2.8 mg/week (0.4 mg/day), whereas the maximum dose for the treatment of T2DM is 1 mg/week.

Although not highly sensitive [62], hepatic cytolysis markers (aspartate aminotransferase—AST, and, above all, alanine aminotransferase—ALT) represent the epiphenomenon of the potential hepatic inflammation in NAFLD patients. Indeed, patients with NAFLD and elevation of liver cytolysis enzymes have an increased risk of NASH and progression to cirrhosis [63]. Meanwhile, treatments that reduce the risk of NAFLD progression are all associated with a significant reduction in hepatic cytolysis [7,63,64,65]. In overweight or obese patients with NAFLD, reduction in body weight ≥7–10% improves liver enzymes together with the histological grade of disease [66,67]. Despite normalization of liver injury indices is an independent predictor of fibrosis improvement in patients with NASH [65], for several drugs this improvement is not associated with a significant histological and prognostic benefit [7,68]. Growing evidence supports the positive impact of GLP-1 RAs in reducing hepatic cytolysis enzymes [56,61,69,70,71,72,73,74,75,76,77,78,79,80,81]. Buse [69] and Klonoff [70] et al. first showed that long-term treatment with exenatide in patients with T2DM was associated with a significant improvement in biomarkers of liver injury, as well as a reduction in body weight, blood pressure and glycosylated hemoglobin levels. Notably, approximately 40% of patients with baseline liver enzyme elevations achieved normal ALT after treatment. Subsequently, Fan et al. [72] compared treatment with exenatide and metformin in diabetic patients with NAFLD. The authors showed that a 12-week treatment with exenatide was associated with a significant reduction in liver enzymes compared to treatment with metformin. Similarly, Sathyanarayana et al. [71] have also demonstrated that treatment with exenatide, as well as pioglitazone, improves liver injury rates in patients with T2DM. However, the reduction in ALT levels was significantly greater following combined treatment with pioglitazone and exenatide. A similar impact on biomarkers of liver damage has also been demonstrated for lixisenatide [75], dulaglutide [78,79], liraglutide [73,76,77] and semaglutide [61,80]. In patients with T2DM and NAFLD, dulaglutide has been shown to significantly reduce levels of liver injury biomarkers compared to placebo at 6 months of treatment [79]. As concerns liraglutide, Armstrong et al. [73] performed a meta-analysis of 4442 patients with T2DM. At a dosage of 1.8 mg/day, liraglutide has been proven to significantly lower ALT levels in patients with elevated baseline values, with a dose-dependent effect. In fact, no significant differences in the impact on biomarkers of liver injury were observed at the dosage of 0.6 and 1.2 mg/day compared to placebo. Changes in liver fat content would be directly related to the reduction in body weight, serum ALT and triglyceride levels [77]. Finally, in a post hoc analysis of two large trials on the effect of semaglutide on body weight and cardiovascular outcomes, this drug was shown to significantly reduce ALT levels in patients with T2DM and/or obesity [80]. The degree of improvement in the indices of liver damage was directly related to the dosage of semaglutide and greater for high doses. Maximal reduction in ALT levels was achieved approximately 28–30 weeks after initiation of treatment, with subsequent stabilization. These effects have been shown to be closely related to drug-induced weight loss. Recently, Newsome et al. [61] confirmed the efficacy of semaglutide in determining an improvement in hepatic cytolysis indices with a dose-dependent effect.

Another target in the treatment of NAFLD is represented by the reduction in triglycerides’ accumulation in the liver. In this regard, GLP-1 RAs have been proven to significantly improve hepatic steatosis. Preclinical studies showed that exendin-4, a GLP-1R agonist derived from the saliva of the Gila monster, was able to significantly reduce the hepatic accumulation of triglycerides both in mouse models [82] and in human hepatocytes [83] by activating the signaling cascade downstream of the insulin receptor (see below). Furthermore, the addition of exenatide to pioglitazone (already known for its benefits on NAFLD) was associated with a greater reduction in hepatic fat content compared to pioglitazone monotherapy [71].

Although inconsistent data is available [84,85], most studies subsequently confirmed the in vivo benefit of GLP-1 RAs on fatty liver disease [50,56,57,58,61,71,74,76,77,81,84,85,86,87,88,89,90]. Compared with other hypoglycemic agents, exenatide significantly reduces the hepatic triglyceride content and epicardial adipose tissue related to weight loss [88]. Braslov et al. [74] underlined that the improvement of hepatic fat accumulation (assessed using the fatty liver index—FLI) induced by exenatide is not detectable for the other oral hypoglycemic agents. In addition to GLP-1 RAs, a similar impact on steatosis (assessed by FLI) was also showed by SGLT2i [81].

Vanderheiden [89], Petit [57] and Yan [50] et al. demonstrated through magnetic resonance imaging a significant reduction in hepatic and subcutaneous fat in patients treated with liraglutide. Treatment with liraglutide would reduce liver fat content by 44% [58]. In addition to the effect on liver fat, liraglutide treatment is also associated with a greater reduction in total fat mass than treatment with sulfonylurea [59,77]. In this setting, the study by Armstrong et al. is very interesting [56]. Through a multicenter, double-blind, placebo-controlled phase 2 trial, the authors evaluated the impact of liraglutide on liver histology in patients with NASH. Treatment with liraglutide (1.8 mg/day) resulted in a greater improvement in steatosis and hepatocyte ballooning compared to placebo. Particularly, liraglutide has been shown to achieve histological resolution of steatohepatitis in 39% of cases after 48 weeks of treatment, compared to 9% of cases in the placebo group. Patients treated with liraglutide had a relative risk of 4.5 (95% CI: 1.1–18.9; p = 0.017) of achieving resolution of NASH compared with placebo, regardless of the presence of T2DM. However, NAFLD activity score (NAS) was not significantly different between the two groups, although there may be additional benefits associated with liraglutide dosages >1.8 mg/day, not evaluated in this trial. No significant differences in weight loss and improvement in glycosylated hemoglobin levels were found between liraglutide responders and non-responders. A similar effect of liraglutide in patients with NASH was also highlighted by Eguchi et al. [76]. Recently, the effects of semaglutide treatment on patients with biopsy-proven NASH have been evaluated [61]. After 72 weeks, treatment with semaglutide at a dose of 0.4 mg/day resulted in the histological resolution of NASH without progression of fibrosis in 59% of patients, compared with 17% in the placebo group. Furthermore, an improvement in NAS was observed in 83% of patients treated with semaglutide 0.4 mg/day, 80% of those treated with 0.2 mg/day, 71% of those treated with 0.1 mg/day and only 44% of those treated with placebo.

Although not fully clarified, the mechanism of action of GLP-1 RAs on fatty liver disease may be multifactorial. If the reduction in body weight and the improvement of glycemic control are crucial factors in determining the effects of GLP-1 on NAFLD, direct pharmacological mechanisms could also be involved. Indeed, in liraglutide-treated mice, the reduction in hepatic steatosis and endoplasmic reticulum oxidative stress occurred regardless of body weight loss (see below) [91,92].

Liver fibrosis is the main determinant of clinical outcomes in patients with NASH [93,94]. In particular, clinically relevant fibrosis is detectable in about 15% of diabetic patients [95]. A treatment for NAFLD must be able to impact liver fibrosis, aiming to reduce the rates of mortality, liver transplantation and liver-related events.

To date, few studies that evaluated the efficacy of GLP-1 RAs on NAFLD-related liver fibrosis are available [56,61,81,96,97,98]. Moreover, most of the data are obtained from indirect markers of fibrosis, such as the APRI (AST-to-platelet count ratio index), the NFS (NAFLD fibrosis score), the FIB-4 (fibrosis index based on 4 factors) or from the assessment of liver stiffness (LSM). In this regard, Ohki et al. [97] observed a significant reduction in APRI in patients treated with liraglutide, similar to those treated with pioglitazone. Recently, Colosimo et al. [81] confirmed a significant reduction in FIB-4 in diabetic patients treated with GLP-1 RAs, similar to those treated with SGLT2i. Finally, Tan et al. [98] found that diabetic patients treated with liraglutide showed a significant reduction in NFS, FIB-4 and LSM after a 12-month follow-up compared to patients treated with other hypoglycemic agents.

Few studies evaluated the impact of GLP-1 RAs on histologically detected liver fibrosis [56,61,96]. Significant data emerged in this regard from the previously discussed trial by Armstrong et al. [56]. In addition to the reduction in body weight, improvement in liver injury rates and resolution of steatohepatitis in 39% of cases, treatment with liraglutide was associated with lower progression of liver fibrosis compared to placebo. Indeed, despite the short duration of the trial (48 weeks), progression of fibrosis was detected in 9% of patients treated with liraglutide and 36% of patients treated with placebo (p = 0.04). However, patients with more severe fibrosis at baseline appear to have a lower response to liraglutide. Thus, early treatment of NASH with GLP-1 RAs may be more effective. Recently, Newsome et al. [61] carried out a 72-week trial on the effects of semaglutide on NASH assessed by liver biopsy before and after treatment. If semaglutide at a dosage of 0.4 mg/day leads to resolution of steatohepatitis in 59% of cases, no significant difference emerged in the improvement of fibrosis compared to placebo. However, the treatment was shown to be effective in reducing the progression of fibrosis (5% vs. 19% of the placebo group), in accordance with what previously demonstrated [56].

The protective role on fibrosis progression demonstrated by GLP-1 RAs could be explained by the effect in reducing the expression of three out of five collagen genes [99]. In fact, this effect demonstrated on mouse models exposed to semaglutide exclusively impacts the de novo production of collagen and does not affect the pre-treatment liver fibrosis. In this regard, in our opinion, it can be assumed that longer-lasting treatments could also show a significant effect on the regression of pre-existing NASH-related fibrosis. In fact, if the drug-induced reduction in hepatic fat content and the consequent lipotoxicity can determine a resolution of the steatohepatitis in the medium-term, avoiding the progression of liver damage, longer times are probably needed for the clearance of the pre-existing fibrosis.

Finally, preclinical data suggest that treatment with GLP-1 RAs could result in lower incidence of NASH-related hepatocellular carcinoma (HCC) [100]. In mouse models, liraglutide has been shown to completely suppress hepatocarcinogenesis. The protective effect on the development of HCC could be mediated both by the improvement of steatohepatitis (steatosis, inflammation, and hepatocyte ballooning) and the better glycemic compensation and weight loss, with reduced systemic inflammatory state and oxidative stress [101,102]. Moreover, in conditions of insulin resistance (such as during NAFLD), a significant imbalance in the levels of adipokines is determined, with lower levels of adiponectin and an increase in those of leptin [103]. It has been demonstrated that this imbalance favors hepatocarcinogenesis [104]. In particular, leptin promotes the proliferation and migration of neoplastic cells. GLP-1 RAs are able to increase adiponectin levels and restore the balance among adipokines [82,103,105]. This pathway could represent one of the protective mechanisms against HCC development exerted by GLP-1 RAs.