Metformin is a biguanide that has been used worldwide for more than 50 years for the treatment of T2DM although its exact mechanism of action remains incompletely understood. It decreases hepatic glucose production and improves insulin sensitivity at the level of the liver, and to a lesser extent, muscle . Metformin has been reported to reduce plasma aminotransferase levels in patients with NAFLD [104,110–112]. Early small studies suggested that metformin improved hepatic steatosis, necro-inflammation and/or fibrosis in patients with NASH [104,113,114], but not in more recent and better controlled clinical trials .
Taken together, although metformin is not recommended as a specific treatment for patients with NASH , it still has significant clinical value to control hyperglycemia and reduce cardiovascular risk in patients with prediabetes or T2DM and NASH .
Glucagon-like peptide 1 agonists
Glucagon-like peptide 1 (GLP-1) analogues have become widely used in patients with T2DM and their beneficial effects have been quite consistent across studies in this population [117,118]. Although their role in NAFLD is more uncertain, recent studies have suggested a therapeutic potential for these pharmacologic agents in NAFLD. In animal models, exendin-4 has been reported to improve both liver function tests and histology [119,120]. Moreover, in vitro studies have shown that human hepatocytes express the GLP-1 receptor and that GLP-1 analogues suppress hepatic lipogenesis and pro-inflammatory mediators . In humans with hepatic steatosis, open-label studies have shown that exenatide may improve liver enzymes
and decrease steatosis when assessed by MRS [122,123], and even improve histology . In a randomized, open-label trial by Sathyanarayana et al. , the combination of pioglitazone and exenatide improved transaminases and liver triglyceride accumulation by MRS beyond that of pioglitazone alone. In patients with T2DM treated with metformin, Jendle et al.  found that the GLP-1 analogue liraglutide added to metformin improved liver aminotransferase concentration and liver fat by computed tomography (CT) significantly more than the addition of glimepiride to the biguanide. Despite these promising results, the underlying mechanisms remain unclear and the relative contribution of direct GLP-1 agonism on the liver, versus weight loss or better glycemic control per se, demand further investigation in well-designed studies.
Thiazolidinediones (TZDs) are glucose-lowering agents that act as insulin sensitizers in humans. They are ligands for the transcription factor peroxisomal proliferator activated receptor-γ (PPAR-γ) that plays key roles in the regulation of metabolic homeostasis and inflammation. PPAR-γ is predominately expressed in adipose tissue, but is also present in the pancreas, liver, spleen, heart, and muscle . In NASH, TZDs improve adipose tissue and hepatic insulin sensitivity, reduce subclinical inflammation, and restore liver histology [115,128].
Both pioglitazone and rosiglitazone have been studied in NAFLD/NASH . Most of the early studies were heterogeneous in terms of baseline patient characteristics and showed inconsistent clinical efficacy [103,112,130–132]. However, across these studies TZDs usually led to improved insulin sensitivity, reduced liver fat content, and normalization of plasma aminotransferase levels [103,130,133].
In patients with biopsy-proven NASH, the efficacy of TZDs, and especially of pioglitazone, has been examined in several studies (Table 20.5). The first RCT was reported in patients with prediabetes or T2DM and NASH . In this population, pioglitazone (45 mg day−1) significantly reduced insulin resistance at the level of the liver, adipose tissue and skeletal muscle, and improved hepatic steatosis, inflammation and hepatocellular ballooning when compared to placebo. In 73% of patients treated with pioglitazone, the NAFLD activity score improved compared to 24% in the placebo group (p<0.001).
A larger study later extended these findings to patients with NASH but without diabetes . They randomized 247 subjects to vitamin E, pioglitazone, or placebo and found histologic improvement in liver steatosis and inflammation but not fibrosis after pioglitazone treatment. On the other hand, Ratziu et al.  evaluated the use of rosiglitazone in patients with NASH and reported more modest results, with only a ∼20% reduction in hepatic steatosis on histology but no improvement in lobular inflammation, ballooning, or fibrosis.
Confirmation about TZDs’ long-term benefit is needed because the studies have been of relative short duration (6 to 24 months). In a preliminary report in patients with T2DM and prediabetes, Cusi et al.  have reported that after 18 months of treatment more subjects had improvement with pioglitazone compared to placebo for liver steatosis (75% vs. 29%, p<0.001), ballooning (55% vs. 26%, p<0.01), and inflammation (53% vs. 26%, p=0.01), as well as in the NAFLD activity score (NAS) (83% vs. 38%, p<0.001). Moreover, pioglitazone had a modest but significant effect on fibrosis compared to placebo (p=0.03) and baseline fibrosis stage (p<0.01).
With pioglitazone being the only TZD available in the United States and most countries, it must be kept in mind that while an effective drug for T2DM and NASH, caution must be exerted about its off-target effects (Figure 20.1) . Thiazolidinediones have the potential to exacerbate congestive heart failure and promote water retention. Moreover, its use has been associated with a greater risk of osteoporosis and bladder cancer. In a meta-analysis of 19 trials in 16,390 patients with T2DM , the rate of congestive heart failure was noted to be slightly higher in pioglitazone users versus control patients using other oral agents for diabetes (2.3% vs. 1.8%, HR 1.41; 95% CI 1.14–1.76; p=0.002). However, there was also a significant reduction in the combined outcome of death, myocardial infarction, or stroke (p=0.005). Pioglitazone has been associated with bladder cancer, and currently the FDA recommends avoiding it if active bladder cancer is present, and use with caution if there is a prior history of the disease. However, a recent 8-year interim analysis (from January 1, 1997 to December 31, 2010) reported a trend for the risk of bladder cancer to be less over time, with hazard ratios (HR) that were nonsignificant in fully adjusted models that included key variables such as tobacco use, duration of pioglitazone therapy, or cumulative pioglitazone dose (Clinical Trial NCT01637935). For example, duration of therapy with pioglitazone for more than 2 years (HR=1.4; 95% CI 1.03–2.0) or 4 years (HR=1.62; 95% CI 0.96–2.74) in the 5-year analysis was higher than the 8-year analysis (HR=1.30; CI 0.91–1.86, NS). Similarly, cumulative doses of pioglitazone >28,000mg in the 5-year analysis (HR=1.43; CI 0.96–2.12) did not continue to increase the risk of bladder cancer in the 8-year analysis (HR=1.25; CI 0.91–1.74, NS). Recent guidelines from the American Association for the Study of Liver Diseases, American College of Gastroenterology, and the American Gastroenterological Association  recommend the use of pioglitazone as a valid alternative for the management of patients with biopsy-proven NASH in patients with diabetes. However, there is clearly a need for more studies to evaluate its long-term efficacy and safety. Careful patient selection is crucial in order to maximize benefits while reducing the risks associated with pioglitazone use, as recently reviewed elsewhere .
Obeticholic acid (OCA)
Obeticholic acid, a derivative of the primary human bile acid chenodeoxycholic acid, is an agonist of the farnesoid X receptor (FXR: a key nuclear receptor that regulates hepatic glucose and lipid metabolism). In a small study , OCA was found to only marginally ameliorate insulin resistance. Improvements in insulin sensitivity and liver aminotransferases only occurred at doses of 25mg, but not at higher doses (50mg). The same was true for plasma markers of inflammation and fibrosis. However, in early 2014, a RCT of OCA (25mg day−1) versus placebo in patients with biopsy-proven NASH was stopped early during an interim analysis because of significant positive results on liver histology. In data from about half of the 283 randomized patients, it was concluded that OCA led to a highly significant improvement in histology according to an intention-to-treat analysis. However, long-term data is still needed to fully assess its safety and efficacy.
Increased awareness among healthcare providers will be needed for the early diagnosis and optimal management of the patient with NAFLD/NASH. Patient care will entail going beyond liver-specific complications to embrace a comprehensive strategy that ameliorates obesity-induced lipotoxicity while addressing associated comorbidities, such as T2DM or CVD. Reminding ourselves that the condition is really that of a chronic “lipotoxic liver disease” helps highlight the crosstalk between dysregulated adipose tissue and the liver, be more aware of its chronic nature and the need for a long-term management plan. It is likely that diagnosis accuracy will improve in the near future with a combination of more specific plasma biomarkers, advances in genetic testing and better metabolic profiling (i.e., metabolomics approaches).
Earlier and more accurate disease staging will allow targeting therapy earlier and direct it only at patients with a greater risk of disease progression.Therapies that relieve the liver from systemic lipotoxicity by targeting adipose tissue insulin resistance and inflammation (e.g. significant weight loss, exercise, and/or pioglitazone) hold reasonable chances of long-term success. However, new and more effective approaches will be needed. One can envision the need for combining agents that target different defects, much like what we do today for T2DM or hypertension. Multiple agents are undergoing testing in RCTs for the treatment of NASH (see https://clinicaltrials.gov/). The different approaches proposed include “hepatoprotective” agents such as silymarin (milk thistle) or the antibiotic rifaximin. They are believed to improve the deleterious gut flora present in obesity and NAFLD and mediate positive changes in insulin sensitivity.
Novel insulin-sensitizers such as resveratrol (an antioxidant), fenretinide (antagonize insulin action), or the new adenosine monophosphate-activated protein kinase (AMPK) activators (oltipraz) that improve insulin sensitivity and fatty acid synthesis, are undergoing investigation. Other approaches include the antioxidant S-adenosyl-L-methionine (SAMe), aimed at reducing lipid peroxidation and secondary cellular injury.
Until we find the optimal long-term lifestyle and pharmacologic approach for NAFLD in T2DM, based on the evidence reviewed, the clinician must remain alert and offer an integrated management plan to the increasing number of patients with this chronic and relentless liver condition.