Management of NonalcoholicFatty Liver Disease in PatientsWith Type 2 Diabetes:A Call to ActionDiabetes Care 2017;40:419–430 | DOI: 10.2337/dc16-1787

Traditionally a disease of hepatologists, nonalcoholic fatty liver disease (NAFLD)has recently become a major concern for a broad spectrum of health careproviders. Endocrinologists and those caring for patients with type 2 diabetesmellitus (T2DM) are at center stage, as T2DM appears to worsen the course ofNAFLD and the liver disease makes diabetes management more challenging.However, the nature of this relationship remains incompletely understood.Although the increasing prevalence of NAFLD is frequently attributed to theepidemic of obesity and is often oversimplified as the “hepatic manifestation ofthe metabolic syndrome,” it is a much more complex disease process that mayalso be observed in nonobese individuals and in patients without clinical mani-festations of the metabolic syndrome. It carries both metabolic and liver-specificcomplications that make its approach unique amongmedical conditions. Diabetesappears to promote the development of nonalcoholic steatohepatitis (NASH), themore severe form of the disease, and increases the risk of cirrhosis and hepato-cellular carcinoma. Patients and physicians face many uncertainties, includingfragmented information on the natural history of the disease, challenges in thediagnosis of NASH, and few pharmacological agents with proven efficacy. How-ever, recent advances in diagnosis and treatment, combined with the risk ofserious consequences from inaction, call for health care providers to be moreproactive in the management of patients with T2DM and NASH.

Nonalcoholic fatty liver disease (NAFLD) is the most common chronic liver conditionof adults in developed countries (1,2). According to current guidelines, the diagnosisis based on the following criteria (3,4): 1) the presence of hepatic steatosis (.5% ofhepatocytes determined by histology or .5.6% determined by nuclear magneticresonance techniques); 2) no significant alcohol consumption (defined as ongoing orrecent alcohol consumption of.21 drinks/week for men and.14 drinks/week forwomen); and 3) no competing etiologies for hepatic steatosis. Histologically, itcovers a wide spectrum of liver disease ranging from isolated steatosis (withoutor with only minimal inflammation) to severe nonalcoholic steatohepatitis (NASH),characterized by inflammation, cell necrosis (ballooning), perilobular fibrosis, andeventually cirrhosis.By mechanisms that are still incompletely understood, patients with type 2 di-

abetes mellitus (T2DM) are particularly susceptible to more severe forms of NAFLD(5,6) and have a higher progression to hepatocellular carcinoma (7,8). Moreover,the coexistence of NAFLD and T2DM results in a worse metabolic profile (9) and a

1Division of Endocrinology, Diabetes andMetab-olism, University of Florida, Gainesville, Florida2Malcom Randall Veterans Affairs Medical Cen-ter, Gainesville, Florida

Corresponding author: Kenneth Cusi, kenneth.cusi@medicine.ufl.edu.

Received 18 August 2016 and accepted 8 De-cember 2016.

This article contains Supplementary Data onlineat http://care.diabetesjournals.org/lookup/suppl/doi:10.2337/dc16-1787/-/DC1.

© 2017 by the American Diabetes Association.Readers may use this article as long as the workis properly cited, the use is educational and notfor profit, and the work is not altered. More infor-mation is available at http://www.diabetesjournals.org/content/license.

Fernando Bril1 and Kenneth Cusi1,2

Diabetes Care Volume 40, March 2017 419

REV

IEW

http://crossmark.crossref.org/dialog/?doi=10.2337/dc16-1787&domain=pdf&date_stamp=2017-02-04mailto:kenneth.cusi@medicine.ufl.edumailto:kenneth.cusi@medicine.ufl.eduhttp://care.diabetesjournals.org/lookup/suppl/doi:10.2337/dc16-1787/-/DC1http://care.diabetesjournals.org/lookup/suppl/doi:10.2337/dc16-1787/-/DC1http://www.diabetesjournals.org/content/licensehttp://www.diabetesjournals.org/content/license

higher cardiovascular risk (5,10). Manyquestions remain regarding the intricaterelationship of NAFLD and T2DM, as wellas the natural history and clinical impli-cations of NAFLD in patients with pre-diabetes or T2DM. This review will focuson the appropriate management andtreatment of these complex patientsbased on current evidence.

WHY SHOULD HEALTH CAREPROVIDERS CARE ABOUT NAFLD?

Based on a recent publication assessingthe 2011–2012 National Health andNutrition Examination Survey data (11),the prevalence of prediabetes and T2DMamong U.S. adults were 38.0% and 14.3%,respectively. Moreover, among thoseadults $65 years of age, the prevalenceof diabetes was 33.0%. Even in the best-case scenario (considering a prevalence ofNAFLD among these patients of only 50%,rather than between ;65% and 70% ascurrent evidence suggests) (Fig. 1), this im-plies that 84 million people in the U.S.live with prediabetes or T2DM andNAFLD. Of these, a significant numberof patients already have NASH or are likelyto develop this complication in the absenceof any preventive intervention (12). How-ever, only a fewpatients receiveadiagnosisof NASH or are ever treated in the clinic(13). There are many reasons for this: 1)patients and clinicians are unaware ofNASH as a potentially serious medical con-dition; 2) diagnosis is missed due to a re-liance on low-sensitivity diagnostic tests(plasma aminotransferase measurementsor liver ultrasound); 3) a confirmatory di-agnosis (liver biopsy) is rarely pursued byproviders, even in patients who are at highrisk ofNASH; and4) patients andphysiciansare uninformed that weight loss and med-ical treatmentsmay reverseNASH.Anotherargument frequently heard among primarycare providers is that NAFLDmay not be ofgreat concern because cirrhosis appears tooccur infrequently in clinical practice. Onemust consider, however, that the onsetand magnitude of the obesity/T2DM epi-demic is a relatively recent phenomenonofthe past 2 decades, which, combined withthe relatively slow nature of the disease,may give a false sense of comfort to theunawarephysician and squander anoppor-tunity for early intervention. Early signs ofan impending “epidemic” of cirrhosis comefrom liver transplant surgeons. They findwith increasing frequency cryptogenic cir-rhosis, usually attributed to undiagnosed

NASH, as a cause for liver transplantation(14). It is estimated that at its currentcourse NASH will soon be the main causeof liver transplantation in the U.S. (15). Ofnote, the prevalence of T2DM among pa-tients with cryptogenic cirrhosis is muchhigher than that in patients with cirrhosisof other causes (16). Moreover, with in-creasing frequency pediatric and adulthepatology clinics are seeing referralsfrom patients in late adolescence andyoung adulthood with advanced liverdisease secondary to NASH.

But the negative consequences of hav-ing NAFLD in the setting of prediabetes orT2DM go far beyond those related to theliver. The presence of NAFLD has been

associated with a myriad of adverse met-abolic alterations in patients with T2DM(17). These patients characteristicallyshow a worse atherogenic dyslipidemiawith hypertriglyceridemia, low levels ofHDL cholesterol (HDL-C), and smallerand denser LDL particles (18,19). Insulinfailure to appropriately suppress hepaticVLDL secretion is at the core of this typicaldyslipidemia (20). Although this dyslipide-mia appears to be driven by intrahepatictriglyceride accumulation and insulin re-sistance, it appears to be independent ofthe presence of obesity or the severity ofNASH (18). Other metabolic alterationsfrequently observed in these patients in-clude higher levels of insulinemia and

Figure 1—Prevalence of NAFLD (panel A) and advanced fibrosis (panel B) in the general pop-ulation and in patientswith T2DMaccording to different diagnostic tools. Note that the presenceof T2DM significantly increases the prevalence of NAFLD and advanced fibrosis. Results wereextrapolated from the following references: plasma ALT (24,25), computed tomography (26,27),liver ultrasound (US) (28,29), CAP (30,31), liver 1H-MRS (32,35), FibroTest (46,47), NAFLD fibrosisscore (48), and vibration-controlled transient elastography (49).

420 NAFLD in T2DM Diabetes Care Volume 40, March 2017

hyperglycemia that is more difficult tocontrol. These are probably the result ofmore severe insulin resistance at the levelof the liver (9) and impaired insulin clear-ance (21). All these metabolic disarrange-ments translate intoworse cardiovasculardisease, the main cause of morbidity andmortality in this population (22). In addi-tion, diabetic microvascular complica-tions are also increased in the presenceof NAFLD, as suggested by several obser-vational studies (23).It is clear that the coexistence of NAFLD

and prediabetes or T2DM is extremelycommon and is associated with severeconsequences to the health care system.Fortunately, a better understandingof thenatural history of the disease, togetherwith access to novel diagnostic tools andrecent evidence of safe and effectivetreatment modalities have set the stagefor a major paradigm shift in the manage-ment of this disease in patients with pre-diabetesor T2DM.Weare at an importantcrossroads, and how well we incorporatethese new diagnostic and therapeutic ad-vances will likely have a large impact onthe quality of life of many patients.

DIAGNOSIS OF NAFLD AND NASH

As can be observed in Fig. 1A (24–32), theprevalence of NAFLD depends on the di-agnostic tool that is used. Although theprevalence using plasma alanine amino-transferase (ALT) concentration is rela-tively low (24,33), it actually depends onthe cutoff point selected as normal.Although a threshold of 40 IU/L is fre-quently used in clinical practice and trials,epidemiological studies have suggestedlower cutoff points to be considered asnormal (i.e., 30 IU/L for males and19 IU/L for females) in order to improvethe sensitivity of the method (34). Never-theless, there is significant evidence sug-gesting that plasma aminotransferases area poor marker of NAFLD even with lowercutoff points. Among patients with T2DMwith normal plasma aminotransferase lev-els, the prevalence of NAFLD was as highas 50% using proton magnetic resonancespectroscopy (1H-MRS), and 56% of thesepatients had a histologic confirmation ofNASH (35). When they are elevated, thelevel of ALT is usually higher than that ofaspartate aminotransferase (AST), unlessthere is advanced liver disease.The availability and low cost of liver

ultrasound have made it the techniqueof choice for routine screening (36).

Sensitivity ranges from as low as 60%(37) to as high as 94% (38). In a meta-analysis by Hernaez et al. (39), the overallsensitivity (85%) and specificity (94%)were acceptable. However, these goodresults can be deceiving as they only re-flected its performance in distinguishingbetween moderate-to-severe NAFLD andthe absence of disease, excluding an im-portant group of patients with mild he-patic steatosis, where the sensitivityof the test is significantly lower. Althoughthe performance of liver ultrasound forthe diagnosis of NAFLD is much betterthan the determination of plasma amino-transferase concentration, it still under-performs when compared with 1H-MRSor liver biopsy (24–32,40) (Fig. 1A). Theuse of semiquantitative scores based ondifferent echographic parameters maysomehow improve the outcome but stillhas a low performance when the hepatictriglyceride content is,12.5% (40).

A noninvasive algorithm based onmet-abolic and anthropometric data (BMI,waist circumference, plasma triglyceridelevels, and g-glutamyl transferase [GGT]concentration), which is known as fattyliver index (41), has also been endorsedby some associations for the diagnosis ofNAFLD because of its simplicity (4). How-ever, it should be taken into account thatmost of the published evidence comesfromcomparisons against liver ultrasoundfindings (i.e., not the gold standard andalready a technique with low sensitivity).Therefore, this is likely to overestimatethe true performance of this algorithm.When the fatty liver index was comparedagainst more accurate methods, such asliver 1H-MRS, 58% of the patients had anindeterminate classification, and only77% of the remaining 42% were correctlyclassified (42). Moreover, the presence offibrosis has also been shown to affect itsperformance (43).

More novel techniques, such as con-trolled attenuation parameter (CAP) andmagnetic resonance–based techniques(e.g., 1H-MRS and MRI-proton density fatfraction) have been shown to be moreaccurate for the diagnosis of NAFLDand have the advantage of being quan-tifiable, providing a tool to assesschanges over time (40,44,45). However,whether changes in steatosis correspondto changes in inflammation or fibrosis re-mains to be fully elucidated. As techniquesfor measuring hepatic triglyceride accu-mulation become more readily available

for clinical use, it is likely that they willbecome the first step for the screeningof hepatic steatosis in NAFLD/NASH inhigh-risk patients. As shown in Fig. 1A(24–32), patients with T2DM have agreater than twofold increase in the prev-alence of NAFLD regardless of themethodused for the diagnosis, suggesting that thispopulation may benefit from routineNAFLD/NASHscreening.Of note, theprev-alence of NAFLD is;70% in patients withT2DMwhen the best available techniquesare used (5).

Once the diagnosis of NAFLD is made,clinicians should focus their attention onassessing the risk of the patient of havingNASH or advanced fibrosis, which aremuch more common in patients withT2DM. As can be observed in Fig. 1B(46–49), cross-sectional studies demon-strate that patients with T2DM have ahigher prevalence of advanced fibrosiswhen compared with the general popula-tion, regardless of the diagnostic methodused. This is consistent with findings fromprospective studies (50,51). Although aliver biopsy remains the gold standardfor the diagnosis of NASH and for stagingthe severity of liver fibrosis, the field isslowly moving toward the use of surro-gate noninvasive techniques for the diag-nosis of this condition. Several scoreshavebeen created based on clinical variables(e.g., levels of plasma ALT, AST, platelets,and albumin, BMI, and the presence ofdiabetes) and other specific surrogatemarkers of liver inflammation and/or fi-brosis to predict the presence of NASHor advanced fibrosis (46–48,52). Many ofthese scores have been developed ina small number of patients withoutincluding a validation cohort and stillawait more rigorous testing. We includedin Table 1 themostwidely used biomarkerpanels for the prediction of advanced fi-brosis. They have been developed in co-horts of approximately $250 patientswith NAFLD and compared against the re-sults of liver histology tests. These testsinclude the FibroTest, NAFLD fibrosisscore, BARDscore, FIB-4 (Fibrosis-4) index,and NAFIC score among others (53). Al-though this is an available diagnostic op-tion to complement the medical historyand/or imaging studies, it should be notedthat with most biomarker scores patientsfrequently fall in a “gray zone” with anintermediate or undetermined risk. Forexample, in the only score specifically de-veloped for patients with T2DM (52), 44%

care.diabetesjournals.org Bril and Cusi 421

http://care.diabetesjournals.org

and 87% of the patients fell in the unde-termined group for either NASH or ad-vanced fibrosis, respectively. Plasmabiomarkers, such as plasma keratin-18(54), fibroblast growth factor 21, andothers, have also been assessed with sim-ilar disappointing results. They frequentlyfalter because of low sensitivity for mild-to-moderate NASH or fibrosis and, at thepresent moment, are of limited discrimi-natory value in diagnosing or monitoringthe disease. The development of new bio-markers using advanced technologies (i.e.,metabolomics) (55) and genetic testing(56) is being actively investigated and of-fers promise in the near future. The mostreproducible polymorphism in genome-wide association studies in NAFLD in-volves the patatin-like phospholipase3 (PNPLA3; rs738409), which has beenshown to be associated with more livertriglyceride accumulation andworse prog-nosis (57).More recently, a polymorphismof the transmembrane 6 superfamilymember 2 (TM6SP2; rs58542926) hasalso been described of relevance in termsof liver histology. However, although indi-vidual genetic predisposition to NAFLDand NASH is likely and genotyping maybe considered in selected patients, cur-rent genetic tests are not recommended

for routine clinical testing given issues re-lated to cost, further validation, and theneed for improvement in their diagnosticsensitivity/specificity.

If available, vibration-controlled tran-sient elastography (FibroScan) or mag-netic resonance elastography (MRE) canalso be used to determine the degree offibrosis (49,58). Both have been shown tocorrelate well with histologic findings andmay avoid the need for a liver biopsy in asignificant number of patients. Unlike Fi-broScan, MRE provides the elasticity ofthe entire liver and is not affected by el-evated BMI values. Although theMRE hasshown to perform better than FibroScan(59), it is more expensive and less acces-sible. Although no study to date has usedMRE to screen for advancedfibrosis in thegeneral population, Doycheva et al. (58)applied this technique in an unselectedgroup of T2DM patients (n = 100) andfound that 7.1% of patients with T2DMhad advanced fibrosis. Unfortunately,many patients cannot be properly classifiedby noninvasive techniques and may stillrequire a liver biopsy. This is the only reli-able way to rule out other chronic liver dis-eases and to distinguish isolated steatosisfrom NASH, potentially avoiding un-necessary exposure to pharmacological

treatment in patients with mild disease(i.e., isolated steatosis). Current guide-lines (3,4) state that a liver biopsy is theonly way to diagnose NASH and that non-invasive techniques have not been fullyvalidated for diagnosing this condition.However, as can be observed in Fig. 2,empiric therapy is included as an optionfor patients unwilling or unable toundergo a liver biopsy. The limitation ofthis approach is the uncertain diagnosisat baseline, the potential for unnec-essary exposure to pharmacologicaltreatment, and the ambiguity in defin-ing treatment response. However, asbetter noninvasive diagnostic tech-niques develop and safer and moreeffective treatments become available,we are likely to see a shift toward amore limited need for liver biopsies priorto treatment initiation for NASH in clinicalpractice.

In Fig. 2, we have provided a sum-mary of the diagnostic approach thatwe suggest for most patients withprediabetes or T2DM who are at highrisk of NASH. However, each patientshould be assessed individually, and acareful evaluation of risks and benefitsshould be performed on a case-by-casebasis.

TREATMENT

Therapy for patients with NASH should beaimed at decreasing disease activity, de-laying the progression of fibrosis, and re-ducing the risk factors associated withtheir high cardiovascular risk (3,4). Cur-rently, there are no pharmacological treat-ments approved by regulatory agenciesfor this condition, so lifestyle interventionremains the standard of care (3,4). In thisscenario, it is not infrequent to hear pri-mary care physicians arguing that diagnos-ing NAFLD and/or NASH is pointless, aslifestyle intervention remains the onlytherapeutic option available, and all pa-tients with prediabetes or T2DM shouldreceive it regardless of their liver findings.However, this statement overlooks thefact that lifestyle intervention alone rarelyachieves complete resolution of NASH, be-ing extremely difficult to accomplish andeven more challenging to maintain overtime. Moreover, lifestyle interventionplus pharmacological treatment are likelyto offer additive benefit. Therefore, al-though all patients should be counseledand encouraged to adopt lifestylechanges, pharmacological therapy should

Table 1—Biomarker panels for the diagnosis of advanced fibrosis (stages 3 and 4)

Parameters included n PPV NPV

Patients unableto be classified(“gray zone”)

FibroTest (115) Age, sex 267 60% 98% 32%Total bilirubin

GGT a2-macroglobulinApolipoprotein A1

Haptoglobin

NAFLD fibrosisscore (116)

Age, BMI 733 82% 88% 24%Diabetes

AST/ALT ratioPlatelet, albumin

†BARD score (117) BMI 827 43% 96% N/ADiabetes

AST/ALT ratio†FIB-4 index (118) Age 541 80% 90% 30%

AST and ALTPlatelet

NAFIC score (119) Ferritin 619 36% 99% 15%Type IV collagen

Insulin†Hepascore (120) Age, sex 242 57% 92% 11%

Total bilirubinGGT a2-macroglobulin

Hyaluronic acid

N/A, not applicable; NPV, negative predictive value; PPV, positive predictive value. †Noindependent validation cohort included in the study.

422 NAFLD in T2DM Diabetes Care Volume 40, March 2017

be strongly considered early on, especiallyin patients with advanced disease orthose who are at high risk of disease pro-gression. With increasing frequency, U.S.Food and Drug Administration–approvedmedications for other conditions (e.g.,pioglitazone [60] and liraglutide [61] forT2DM; obeticholic acid [62] for primarybiliary cholangitis) are proving to be safeand effective in randomized controlledtrials for patients with NASH.Given this rapidly growing arsenal of

therapeutic options, clinicians will be in-creasingly faced with the dilemma ofchoosing the right option. In this section,we will provide clinicians with practicalrecommendations for the treatment ofthese complex patients. Managementshould be focused on the following fiveaspects (Fig. 3):

1. Lifestyle intervention2. Pharmacological treatment of liver

disease3. Treatment of hyperglycemia4. Treatment of dyslipidemia

5. Control of other cardiovascular riskfactors

Lifestyle InterventionLifestyle intervention is beneficial for pa-tients with NAFLD, improving not onlyliver disease but also hyperglycemia, ath-erogenic dyslipidemia, and blood pres-sure levels (3,4,63). However, less isknown about the long-term effects of life-style intervention on liver histology (i.e.,beyond 1 year of intervention), on theleast amount of weight loss needed toachieve such histological benefit, and onthe best strategy tomaintain it over time.

In imaging studies using the gold stan-dard 1H-MRS technique, the relative re-duction in hepatic steatosis by lifestyleintervention has beenusually in the rangeof ;40–50%, although absolute changes(perhaps the most important factor forhistological improvement) usually havebeen small and on the order of ;5%(3,10,63). Among the few well-controlledstudieswithpaired liver biopsies, Promratet al. (64) reported an improvement in

hepatic steatosis and necroinflammationin patients losing $7% of total bodyweight over 48 weeks with a moderate-intensity hypocaloric diet plus an exerciseprogram (200 min/week). Overall, thisstudy showed that improvement on his-tologywas proportional to themagnitudeof the weight loss. Using a similar ap-proach (a hypocaloric diet combinedwith walking 200min/week), a recent un-controlled study (65) in 261 patients withpaired biopsies after 12 months reportedsimilar benefits from lifestyle interven-tion. Together, it appears that a weightreduction in the magnitude range of;5–7% may clearly decrease steatosisbut that more weight loss is needed(;8–10% reduction) to reverse steatohe-patitis. Weight reductions of $10% mayalso cause a significant regression of fi-brosis (65). In line with these findings,large weight reductions obtained afterbariatric surgery showed that most pa-tients experience a decrease in steatosis(;90%), in steatohepatitis (;80%), andeven in fibrosis (;65%) (66). These re-sults were confirmed in a recent prospec-tive study (67) where ;50% of patientshad improvement in fibrosis scores. Ofnote, the magnitude of fibrosis reductiondepends on the baseline severity of liverdisease, with no improvement in fibrosisobserved 5 years after bariatric surgeryin a large cohort of patients (n = 381)with overall mild liver disease (68). How-ever, most bariatric surgery studies haveseveral limitations. These studies are usu-ally small (,100 patients) and lack stan-dardization of the preoperative very-lowcalorie diet intervention and the postop-erative dietary intervention or detailsabout how the intraoperative liver biopsysample is obtained (which may alter thebaseline histological reading). In addition,the repeat postbypass liver biopsies areusually performed at varying intervalsover time. Finally, most studies have notbeen prospective or controlled, andtherefore they were potentially at riskfor patient selection bias. Indeed, thereare no randomized controlled trials eval-uating a given bariatric surgery procedureversus lifestyle intervention, placebo(sham procedure), a given pharmacologi-cal intervention, or across surgical ap-proaches in patients with NASH. It isalso unclear whether changes in liver dis-ease are merely the result of weight re-duction or whether bariatric surgery hasan intrinsic metabolic effect on the liver.

Figure 2—Algorithm for the diagnosis of NAFLD and NASH in patients with prediabetes or T2DMin clinical practice. This suggested algorithm is based on the authors’ interpretation of availableevidence. MR, magnetic resonance; US, ultrasound. *Based on results from more sensitive testssuch as liver 1H-MRS, MRI-proton density fat fraction, or CAP. #Patatin-like phospholipasedomain-containing 3 (PNPLA3) I148M and/or transmembrane 6 superfamily member2 (TM6SF2) E167 K.

care.diabetesjournals.org Bril and Cusi 423

http://care.diabetesjournals.org

Well-designed prospective studies areneeded to determine the ideal patient,type of surgery, and long-term efficacyand safety of bariatric surgery in NAFLD/NASH. Consistent with these limitations, aCochrane review (69) concluded that it istoo early for a definitive assessment of ben-efits versus harms of bariatric surgery inNASH, and current recommendations con-sider it premature to indicate bariatric sur-gery specifically to treatNASH, although it isnot contraindicated in otherwise eligi-ble obese individuals with NAFLD or NASH(unless theyhave established cirrhosis) (3).The above evidence suggests that life-

style interventions are as good as themagnitude of weight reduction they pro-duce. Further supporting this concept,studies with weight loss medicationssuch as orlistat (70) or liraglutide (61)have reported a histological improve-ment proportional to the amount ofweight loss. This is the reason why phar-macological agents that induce weightloss should always be considered, espe-cially if lifestyle intervention is unsuccess-ful (Fig. 3). This probably implies thatthere may not be a lifestyle interventionstrategy that is better than the rest, andthat weight reduction per se should bethe primary aim. For example, resistancetraining and aerobic exercise interven-tions achieving similar weight reductionwere equally effective in reducing livertriglyceride content by;30% among pa-tients with diabetes and NAFLD (71).Comparable effects have been observed

with equally hypocaloric low-carbohydrateversus high-carbohydrate diets (72) andlow-fat versus low-carbohydrate diets(73). However, some studies reportsteatosis reduction even with minimalweight loss, indicating that other factorsmayalsoplaya (minor) role inNASH improve-ment (3,4). For example, several small (n =18–45) and short-term (4–24 weeks) studiesreported a modest reduction in intrahepatictriglyceride accumulation by 1H-MRS(;15%) after aerobic or resistance exer-cise without any significant weight loss(74). Moreover, 2- to 4-week isocaloriclow-fat diets significantly reduced intra-hepatic triglyceride accumulation as de-termined by 1H-MRS without producinganyweight losswhen comparedwith iso-caloric high-fat diets (74). Diet supple-ments, such as vitamin D or n-3 fattyacids, have also been suggested fortreatment in patients with NAFLD, buttreatment with both supplements hasfailed to show any consistent associa-tions with liver triglyceride accumulationor NASH (75,76). Clearly, more studiesare needed to completely understandthe role of lifestyle intervention in thetreatment of NASH and to establish thebest strategy to treat patients withT2DM and NASH.

Pharmacological Treatment of LiverDiseaseAs more drugs prove their safety and ef-ficacy in randomized controlled trials inpatients with NASH, and both physicians

andpatients continue to strugglewith thechallenge of achieving and maintaininga significant weight reduction, weenvision a paradigm shift in the near fu-ture toward more frequent combinationof pharmacological treatment with life-style intervention in patients with NASH.

In Fig. 4 (60–62,77–79), we have sum-marized the histologic effects of severaldrugs included in randomized controlledtrials that have reported on resolutionof NASH. Only with treatment with pio-glitazone (either 30 or 45 mg daily)(60,78,79), vitamin E (78), and liraglu-tide (61) did a significant proportion ofpatients achieve resolution of NASHwhen compared with placebo (Fig. 4A).The treatment effects were ;30% forpioglitazone and liraglutide and ;15%for vitamin E. However, cross-comparisonsbetween studies should be avoided be-cause of differences in the populationsstudied and other factors, such as differ-ences in the proportion of patients withT2DM (60,79). It should also be takeninto account that the definition of resolu-tion of NASH varied somewhat across thestudies.

Regarding individual histologic pa-rameters, as can be observed in Fig. 4(60–62,77–79), pioglitazone (either30 or 45 mg daily) (60,78,79), vitaminE (78), and obeticholic acid (62) showedthe most consistent results, with signif-icant improvements in steatosis (Fig.4C), inflammation (Fig. 4D), and bal-looning (Fig. 4E). Pioglitazone 45 mghad the highest reduction in steatosisand inflammation grades. Vitamin Eand obeticholic acid were also benefi-cial but with an improvement of asomewhat lesser magnitude (;25%and;50%, respectively, of that reportedfor pioglitazone). The overall reduc-tion in ballooning grades was rathersmall but significant for most medica-tions, except for low-dose elafibranor(77) and liraglutide (61). Discrepanciesin the liraglutide study (relatively highrates of resolution of NASH with non-significant changes in individual histo-logical scores) may be a result of thesmall size of the study, with only 45 pa-tients having paired biopsies. Certainly,larger studies with liraglutide and otherglucagon-like peptide 1 receptor agonistsare needed in order to establish theirfuture role in the treatment of NASH.Comparison of results expressed as a per-centage of patientswith improvement for

Figure 3—Algorithm for the management of patients with prediabetes or T2DM and definiteNASH. This suggested therapeutic algorithm is based on the safety and efficacy of interven-tions assessed in randomized controlled trials. BP, blood pressure; CV, cardiovascular; TG,triglyceride.

424 NAFLD in T2DM Diabetes Care Volume 40, March 2017

each histological parameter can be foundin Supplementary Fig. 1.There has been increasing interest in

finding medications that could preventfibrosis progression in NASH or thatcould even reverse established fibrosis.A mean reduction in fibrosis stage wassignificant only for pioglitazone 45 mg(60) (Fig. 4F), but other studies havealso suggested a benefit. Belfort et al.(79) reported significant changes in fibro-sis compared with baseline, and Aithal

et al. (80) also reported a reduction infibrosis in patients without diabetestreated with pioglitazone 30 mg/day.Thus, it is possible that pioglitazonemay alter the natural history of the dis-ease and its progression to cirrhosis insome patients with NASH, encouragingearly diagnosis and treatment. Othermedications with a potential for liverantifibrotic properties require confirma-tion in larger studies. For instance, afterobeticholic acid therapy more patients

showed improvement in liver fibrosiscompared with treatment with placebo(35% vs. 19%, P = 0.004), although themean fibrosis score did not change sig-nificantly. Other agents include pentox-yfilline (mean treatment effect onfibrosis [n = 46] 20.6, P = 0.038) (81)and liraglutide (treatment effect [n =45] 20.4, P = 0.11) (61). A preliminaryreport (82) has also indicated some ben-eficial effect from cenicriviroc therapy onfibrosis but not on resolution of NASH or

Figure 4—Treatment effect vs. placebo for different histological outcomes for pharmacological agents assessed in randomized controlled trialsreporting resolution of NASH as one of the outcomes. *Implies statistical significance (panels B–F). NAS, NAFLD activity score.

care.diabetesjournals.org Bril and Cusi 425

http://care.diabetesjournals.org/lookup/suppl/doi:10.2337/dc16-1787/-/DC1http://care.diabetesjournals.org

individual histological parameters after12 months of therapy.In patients with NASH and prediabetes

or T2DM, the evidence appears to showthat pioglitazone has the greatest treat-ment effect. It targets not only liver his-tology, but also the underlying metabolicdisturbances, in particular insulin resis-tance (83). Of note, histological improve-ment after pioglitazone therapy is closelycorrelated with the reversal of adiposetissue insulin resistance (84) and an in-crease in plasma adiponectin levels (85).In the long term, its metabolic and histo-logic benefits appear to persist over time(60), but they wane after treatment dis-continuation (86).Before consideration, benefits must be

balanced against potential adverse eventsof pioglitazone therapy (83). In a meta-analysis of 16,390 patients (87), althoughcongestive heart failure was slightly morefrequent with pioglitazone therapy com-pared with other treatments (hazard ratio1.41 [95% CI 1.14–1.76], P = 0.002), therewas a significant reduction in the combinedoutcomeof death,myocardial infarction, orstroke. This finding is consistent with theantiatherogenic effect of pioglitazone re-ported in other randomized controlled tri-als (88–90). Moreover, in a recent study(91) in patientswithout T2DM, pioglitazonetherapy decreased stroke and myocardialinfarction (fatal and nonfatal) by 24% inpatients with a previous ischemic strokeor transient ischemic attack. Pioglitazonetherapy also reduced the progressionfrom prediabetes to T2DM by ;50% (82)to;70% (92). Thus, it may offer a liver-specific and comprehensive metabolictreatment to patients with NASH in whomthe presence of insulin resistance and pre-diabetes are already common (17). Pioglita-zonemay causeweight gain, but it is usuallyless than commonly believed (;2–3 kg inlong-term studies of 2–4 years duration inpatients with T2DM) and recently has beenreported (60) to be 3.1 kg after 36 monthsof treatment inpatientswithprediabetesorT2DMandNASH.Anadditional concernhasbeen about whether pioglitazone may in-crease the risk of bladder cancer in males.Evidence is conflicting, with some studiesshowing association (i.e., one extra caseper 3,408 patients treated with extendedtreatment) (93), although a larger 10-yearprospective study failed to find any associ-ation (94).In summary, pioglitazone may change

the natural history of the disease in

patients with T2DM and NASH. Early ad-dition of this drug to the antidiabeticregimens of such patients should beconsidered after metformin therapy.Whether histological outcomes can beimproved in patients with T2DM andNASH by combining treatment with vita-min E and pioglitazone is being activelyexplored by our group in a long-termthree-arm study of vitamin E/pioglitazoneversus vitamin E/placebo versus pla-cebo/placebo (Clinical trial reg. no.NCT01002547, clinicaltrials.gov).

Treatment of HyperglycemiaTreatment of hyperglycemia is importantbecause NASH in patients with prediabe-tes/diabetes carries aworse prognosis (5).Paradoxically, most studies in patientswith NASH have been carried out in pa-tientswithout diabetes, so the role of con-trolling hyperglycemia per se in patientswith steatohepatitis and T2DM, indepen-dent of changes in insulin sensitivity,remains largely unknown. A common ob-servation in clinical practice is that pa-tients with uncontrolled diabetes andelevated plasma aminotransferase levelsusually normalize (or significantly im-prove) their aminotransferase levelsonce their diabetes has been better con-trolled. In the sameway, the developmentof uncontrolled hyperglycemia has beenobserved to be associated with an in-crease in plasma aminotransferase levels.Moreover, in a small proof-of-conceptstudy with paired biopsies, patients withthe greatest histological improvement af-ter 2.4 years of follow-up were the oneswith the largest A1C reduction. However,interpretation must be performed care-fully because the same group also showedthe greatest weight loss and the largestimprovement in insulin sensitivity (95).

Overall, these findings suggest that anydiabetes treatment may be of benefitfor patients with NASH if they have un-controlled hyperglycemia. Proof of thiscomes from studies that have shown areduction of steatosis after treatmentwith insulin (96) or an improvement ofplasma aminotransferase levels after so-dium–glucose cotransporter 2 inhibitortherapy (97). However, future studiesare needed to assess the effect of glyce-mic control on steatohepatitis in T2DM.

Whether the above observationsshow a direct response to changes inplasma glucose levels (decreasing gluco-toxicity) or to the alleviation of insulin

resistance remains unclear. Regardlessof the mechanism, physicians shouldpay close attention to diabetes controlin patients with NASH as this is likely,but unproven, to delay the progressionof liver disease. In addition, becauseNAFLD has been associated with worseprogression of retinopathy (98) and ne-phropathy (23), the presence of NAFLDmay identify a group of patients whocould potentially benefit from stricterglycemic control.

Treatment of DyslipidemiaPatients with NAFLD, independent of obe-sity or the histologic severity, have worseatherogenic dyslipidemia (18,19). As men-tioned earlier, it is usually characterized byhypertriglyceridemia, low HDL-C levels,and smaller and denser LDL particles. Inclinical practice, patients with NAFLD arefrequently denied a statin for a combina-tion of reasons, the most important beingthe fear of hepatotoxicity in patients withalready elevated plasma aminotransferaselevels (99).A recent study (100) showed thatonly 42.6% of patients with NAFLD receivedan appropriate statin therapy, despite theirsignificantly elevated cardiovascular risk.

Evidence from retrospective (101) andcross-sectional (102) studies has shownthat statins are safe in these patients, andthat theymayevencontribute todecreasedplasma aminotransferase levels and im-prove hepatic steatosis and histology (99).However, a randomized controlled trial(103) with simvastatin failed to show anyeffect on hepatic steatosis, and the onlysmall randomized controlled trial (104) inpatientswithNASHdid not show any effectof simvastatin on liver histology. Althoughprospective studies assessing the safetyand efficacy of statins are much neededfor confirmation, the current dogma isthat patientswith NAFLD should be treatedwith a statin because of their elevated car-diovascular risk (Fig. 3) (99). Moreover, be-cause these patients frequently exhibithypertriglyceridemia even after startingstatin therapy, they may even benefitfrom the addition of fenofibrate in combi-nationwith the statin. This has been shownto be useful in post hoc analyses of largerandomized controlled trials, such as theAction to Control Cardiovascular Risk in Di-abetes (ACCORD) (105) and the FenofibrateIntervention and Event Lowering in Diabe-tes (FIELD) (106) trials, where the subgroupof patients with high triglyceride and lowHDL-C levels (typical dyslipidemia inNAFLD)

426 NAFLD in T2DM Diabetes Care Volume 40, March 2017

had a reduction in cardiovascular out-comes after the addition of fenofibrateto treatment.Whether the addition of ezetimibe

could further contribute to a reductionin cardiovascular risk or to the progres-sion of liver disease in this populationhas been a matter of extensive debate.In the Improved Reduction of Out-comes: Vytorin Efficacy InternationalTrial (IMPROVE-IT), the addition ofezetimibe to a statin after an acutecoronary syndrome event showed asignificant reduction in the compositeprimary outcome (107), and ezetimibeis, therefore, likely to be of benefit forhigh-risk patients. However, currentguidelines still do not recommend itsroutine use for the prevention of car-diovascular disease (99). Studies assess-ing its effects on surrogate markers ofliver disease in NAFLD have shown nega-tive findings (108,109).

Control of Other Cardiovascular RisksAll other traditional cardiovascular riskfactors should also be actively ad-dressed in patients with NASH becauseof their elevated cardiovascular risk.Current data on the association of smok-ing and the progression of NAFLD isscarce. In a large multicenter cohortfrom the Nonalcoholic SteatohepatitisClinical Research Network (110), smok-ing was associated with advanced fibro-sis, at least in part through mechanismsassociated with insulin resistance. Re-gardless of the direct effects of smokingon the liver, smoking cessation shouldbe strongly encouraged in order to re-duce cardiovascular risk in this alreadyhigh-risk population. In a similar way,optimal blood pressure control shouldalso be encouraged (Fig. 3). Althoughat this time no formal recommendationcan be given regarding the best agent tomanage hypertension in NASH, smalluncontrolled clinical studies have sug-gested that ACE inhibitors (ACEIs) andangiotensin receptor blockers (ARBs)may be able to play a role in improvinginsulin signaling and preventing fibrosisin the liver (111). An open-label, uncon-trolled study (112) in humans with NASHhas shown histologic improvement after20 months of therapy with telmisartanor valsartan. With telmisartan therapy,the improvements were more impor-tant, which may be accounted for bythe activity of peroxisome proliferator–

activated receptor g. However, in themost comprehensive study to date withARBs, negative results were reported(113). Regardless of whether these med-ications contribute to the treatment ofNASH or not, either ACEIs or ARBs shouldbe prescribed early on to patients withNAFLD and T2DM (in the presence of mi-croalbuminuria and/or hypertension) inorder to delay the progression of ne-phropathy, which may progress morerapidly in the setting of NAFLD.

CONCLUSIONS

Much progress has been made in our un-derstanding of the pathogenesis, diagnosis,and treatment of NAFLD. For the first time,pharmacological treatments offer hope(i.e., pioglitazone, vitaminE, liraglutide,obe-ticholic acid, and elafibranor) of altering theprogressive nature of the disease in manypatients. However, at the present time, pa-tients are often missed as practitioners arelimited to screening with low-sensitivitytools, such asmeasurement of plasma ami-notransferase concentration and liver ultra-sound. Until recently, the field lackednoninvasive, cost-effective tests and wasat a stage equivalent to diabetic nephropa-thy before the use of microalbuminuria orosteoporosis before the availability of DEXAimaging. Both of these are indolent chronicconditions (like NASH), and today diabeticnephropathy and osteoporosis are activelypursuedand treatedearlyonwith amarkedreduction in morbidity and health carecosts. The availability of simple diagnostictests that can be widely used by practi-tioners, in combination with access tolow-cost, safe, and more effective medica-tions in the near future (114), will radicallychange disease management in the nearfuture. We predict that, invigorated bythese recent diagnostic (e.g., improved im-aging and plasma biomarkers/genetic tests)and therapeutic (e.g., pioglitazone) devel-opments, screening and early interventionforNASHwill become a standard of care forall patients with T2DM. Until then, muchwork remains to be done.

Duality of Interest. K.C. has received researchsupport from Janssen and Novartis and hasserved as a consultant for (in alphabetical order)Eli Lilly and Company, Janssen, Pfizer, andTobira Therapeutics, Inc. No other potentialconflicts of interest relevant to this article werereported.Author Contributions. F.B. and K.C. wrote,edited, and reviewed the manuscript. F.B. and

K.C. are the guarantors of this work and, as such,had full access toall thedata in thestudyand takeresponsibility for the integrityof thedataand theaccuracy of the data analysis.

References1. Loomba R, Sanyal AJ. The global NAFLD epi-demic. Nat Rev Gastroenterol Hepatol 2013;10:686–6902. Rinella ME. Nonalcoholic fatty liver disease: asystematic review. JAMA 2015;313:2263–22733. Chalasani N, Younossi Z, Lavine JE, et al.;American Gastroenterological Association;American Association for the Study of Liver Dis-eases; American College of Gastroenterology.The diagnosis and management of non-alcoholic fatty liver disease: practice guideline bythe American Gastroenterological Association,American Association for the Study of Liver Dis-eases, and American College of Gastroenterology.Gastroenterology 2012;142:1592–16094. European Association for the Study of theLiver (EASL); European Association for the Studyof Diabetes (EASD); European Association forthe Study of Obesity (EASO). EASL-EASD-EASOClinical Practice Guidelines for themanagementof non-alcoholic fatty liver disease. Diabetologia2016;59:1121–11405. Hazlehurst JM,Woods C,Marjot T, Cobbold JF,Tomlinson JW. Non-alcoholic fatty liver diseaseand diabetes. Metabolism 2016;65:1096–11086. Koliaki C, Roden M. Hepatic energy metabo-lism in human diabetes mellitus, obesity andnon-alcoholic fatty liver disease. Mol Cell Endo-crinol 2013;379:35–427. El-Serag HB, Tran T, Everhart JE. Diabetesincreases the risk of chronic liver disease andhepatocellular carcinoma. Gastroenterology2004;126:460–4688. Wang C,Wang X, Gong G, et al. Increased riskof hepatocellular carcinoma in patients with di-abetes mellitus: a systematic review and meta-analysis of cohort studies. Int J Cancer 2012;130:1639–16489. Lomonaco R, Bril F, Portillo-Sanchez P, et al.Metabolic impact of nonalcoholic steatohepati-tis in obese patients with type 2 diabetes. Di-abetes Care 2016;39:632–63810. Cusi K. Role of obesity and lipotoxicity in thedevelopment of nonalcoholic steatohepatitis:pathophysiology and clinical implications. Gas-troenterology 2012;142:711–725.e611. Menke A, Casagrande S, Geiss L, Cowie CC.Prevalence of and trends in diabetes amongadults in the United States, 1988-2012. JAMA2015;314:1021–102912. Pais R, Charlotte F, Fedchuk L, et al.; LIDOStudy Group. A systematic review of follow-upbiopsies reveals disease progression in patientswith non-alcoholic fatty liver. J Hepatol 2013;59:550–55613. Rinella ME, Lominadze Z, Loomba R, et al.Practice patterns in NAFLD and NASH: real lifediffers from published guidelines. Therap AdvGastroenterol 2016;9:4–1214. Caldwell SH, Crespo DM. The spectrum ex-panded: cryptogenic cirrhosis and the naturalhistory of non-alcoholic fatty liver disease.J Hepatol 2004;40:578–58415. Wong RJ, Aguilar M, Cheung R, et al. Non-alcoholic steatohepatitis is the second leadingetiology of liver disease among adults awaiting

care.diabetesjournals.org Bril and Cusi 427

http://care.diabetesjournals.org

liver transplantation in the United States. Gas-troenterology 2015;148:547–55516. Bugianesi E, Vanni E,Marchesini G.NASH andthe risk of cirrhosis and hepatocellular carcinomain type 2 diabetes. Curr Diab Rep 2007;7:175–18017. Ortiz-Lopez C, Lomonaco R, Orsak B, et al.Prevalence of prediabetes and diabetes andmetabolic profile of patients with nonalcoholicfatty liver disease (NAFLD). Diabetes Care 2012;35:873–87818. Bril F, Sninsky JJ, Baca AM, et al. Hepatic stea-tosis and insulin resistance, but not steatohepati-tis, promote atherogenic dyslipidemia in NAFLD.J Clin Endocrinol Metab 2016;101:644–65219. Siddiqui MS, Fuchs M, Idowu MO, et al. Se-verity of nonalcoholic fatty liver disease andprogression to cirrhosis are associated with ath-erogenic lipoprotein profile. Clin GastroenterolHepatol 2015;13:1000–1008.e320. Fabbrini E, Mohammed BS, Magkos F,Korenblat KM, Patterson BW, Klein S. Alter-ations in adipose tissue and hepatic lipid kinet-ics in obese men and women with nonalcoholicfatty liver disease. Gastroenterology 2008;134:424–43121. Bril F, Lomonaco R, Orsak B, et al. Relation-ship between disease severity, hyperinsuline-mia, and impaired insulin clearance in patientswith nonalcoholic steatohepatitis. Hepatology2014;59:2178–218722. Targher G, Bertolini L, Rodella S, et al. Non-alcoholic fatty liver disease is independentlyassociated with an increased incidence of car-diovascular events in type 2 diabetic patients.Diabetes Care 2007;30:2119–212123. Musso G, Gambino R, Tabibian JH, et al.Association of non-alcoholic fatty liver diseasewith chronic kidney disease: a systematic re-view and meta-analysis. PLoS Med 2014;11:e100168024. IoannouGN, Boyko EJ, Lee SP. Theprevalenceand predictors of elevated serum aminotransfer-ase activity in the United States in 1999-2002. AmJ Gastroenterol 2006;101:76–8225. Forlani G, Giorda C, Manti R, et al. The bur-den of NAFLD and its characteristics in a nation-wide population with type 2 diabetes. J DiabetesRes 2016;2016:293198526. Tota-Maharaj R, Blaha MJ, Zeb I, et al. Eth-nic and sex differences in fatty liver on cardiaccomputed tomography: the multi-ethnic studyof atherosclerosis. Mayo Clin Proc 2014;89:493–50327. McKimmie RL, Daniel KR, Carr JJ, et al. He-patic steatosis and subclinical cardiovasculardisease in a cohort enriched for type 2 diabetes:the Diabetes Heart Study. Am J Gastroenterol2008;103:3029–303528. Lazo M, Hernaez R, Eberhardt MS, et al.Prevalence of nonalcoholic fatty liver diseasein the United States: the Third National Healthand Nutrition Examination Survey, 1988-1994.Am J Epidemiol 2013;178:38–4529. Williamson RM, Price JF, Glancy S, et al.; Ed-inburgh Type 2 Diabetes Study Investigators.Prevalence of and risk factors for hepatic steato-sis and nonalcoholic fatty liver disease in peoplewith type 2 diabetes: the Edinburgh Type 2 Di-abetes Study. Diabetes Care 2011;34:1139–114430. Carvalhana S, Leitão J, Alves AC, BourbonM, Cortez-Pinto H. How good is controlled at-tenuation parameter and fatty liver index for

assessing liver steatosis in general population:correlation with ultrasound. Liver Int 2014;34:e111–e117

31. Kwok R, Choi KC, Wong GL, et al. Screeningdiabetic patients for non-alcoholic fatty liverdisease with controlled attenuation parameterand liver stiffness measurements: a prospectivecohort study. Gut 2016;65:1359–1368

32. Browning JD, Szczepaniak LS, Dobbins R,et al. Prevalence of hepatic steatosis in an urbanpopulation in the United States: impact of eth-nicity. Hepatology 2004;40:1387–1395

33. Ong JP, Pitts A, Younossi ZM. Increasedoverall mortality and liver-related mortality innon-alcoholic fatty liver disease. J Hepatol 2008;49:608–612

34. Prati D, Taioli E, Zanella A, et al. Updateddefinitions of healthy ranges for serum alanineaminotransferase levels. Ann Intern Med 2002;137:1–10

35. Portillo-Sanchez P, Bril F, Maximos M, et al.High prevalence of nonalcoholic fatty liver dis-ease in patients with type 2 diabetes mellitusand normal plasma aminotransferase levels.J Clin Endocrinol Metab 2015;100:2231–2238

36. Williams CD, Stengel J, Asike MI, et al. Prev-alence of nonalcoholic fatty liver disease andnonalcoholic steatohepatitis among a largelymiddle-aged population utilizing ultrasoundand liver biopsy: a prospective study. Gastroen-terology 2011;140:124–131

37. Hepburn MJ, Vos JA, Fillman EP, Lawitz EJ.The accuracy of the report of hepatic steatosison ultrasonography in patients infected withhepatitis C in a clinical setting: a retrospectiveobservational study. BMC Gastroenterol 2005;5:14

38. Kim SH, Lee JM, Kim JH, et al. Appropriate-ness of a donor liver with respect to macrostea-tosis: application of artificial neural networks toUS imagesdinitial experience. Radiology 2005;234:793–803

39. Hernaez R, Lazo M, Bonekamp S, et al. Diag-nostic accuracy and reliability of ultrasonographyfor the detection of fatty liver: a meta-analysis.Hepatology 2011;54:1082–1090

40. Bril F, Ortiz-Lopez C, Lomonaco R, et al. Clin-ical value of liver ultrasound for the diagnosis ofnonalcoholic fatty liver disease in overweightand obese patients. Liver Int 2015;35:2139–2146

41. Bedogni G, Bellentani S, Miglioli L, et al. TheFatty Liver Index: a simple and accurate predic-tor of hepatic steatosis in the general popula-tion. BMC Gastroenterol 2006;6:33

42. Cuthbertson DJ, Weickert MO, Lythgoe D,et al. External validation of the fatty liver indexand lipid accumulation product indices, using1H-magnetic resonance spectroscopy, to iden-tify hepatic steatosis in healthy controls andobese, insulin-resistant individuals. Eur J Endo-crinol 2014;171:561–569

43. Fedchuk L, Nascimbeni F, Pais R, CharlotteF, Housset C, Ratziu V; LIDO Study Group. Per-formance and limitations of steatosis bio-markers in patients with nonalcoholic fattyliver disease. Aliment Pharmacol Ther 2014;40:1209–1222

44. Noureddin M, Lam J, Peterson MR, et al.Utility of magnetic resonance imaging versushistology for quantifying changes in liver fat in

nonalcoholic fatty liver disease trials. Hepatol-ogy 2013;58:1930–194045. Lin SC, Heba E,Wolfson T, et al. Noninvasivediagnosis of nonalcoholic fatty liver disease andquantification of liver fat using a new quantita-tive ultrasound technique. Clin GastroenterolHepatol 2015;13:1337–1345.e646. Poynard T, Lebray P, Ingiliz P, et al. Preva-lence of liver fibrosis and risk factors in a gen-eral population using non-invasive biomarkers(FibroTest). BMC Gastroenterol 2010;10:4047. Perazzo H, Munteanu M, Ngo Y, et al.; FLIPConsortium. Prognostic value of liver fibrosisand steatosis biomarkers in type-2 diabetesand dyslipidaemia. Aliment Pharmacol Ther2014;40:1081–109348. Long MT, Pedley A, Massaro JM, HoffmannU, Fox CS. The association between non-invasivehepatic fibrosis markers and cardiometabolicrisk factors in the Framingham Heart Study.PLoS One 2016;11:e015751749. Koehler EM, Plompen EP, Schouten JN,et al. Presence of diabetes mellitus and steato-sis is associated with liver stiffness in a generalpopulation: the Rotterdam study. Hepatology2016;63:138–14750. Hossain N, Afendy A, Stepanova M, et al.Independent predictors of fibrosis in patientswith nonalcoholic fatty liver disease. Clin Gas-troenterol Hepatol 2009;7:1224–1229.e1-2.51. Adams LA, Sanderson S, Lindor KD, AnguloP. The histological course of nonalcoholic fattyliver disease: a longitudinal study of 103patientswith sequential liver biopsies. J Hepatol 2005;42:132–13852. Bazick J, Donithan M, Neuschwander-TetriBA, et al. Clinical model for NASH and advancedfibrosis in adult patients with diabetes andNAFLD: guidelines for referral in NAFLD. Diabe-tes Care 2015;38:1347–135553. Castera L. Noninvasive evaluation of non-alcoholic fatty liver disease. Semin Liver Dis2015;35:291–30354. Cusi K, Chang Z, Harrison S, et al. Limitedvalue of plasma cytokeratin-18 as a biomarkerfor NASH and fibrosis in patients with non-alcoholic fatty liver disease. J Hepatol 2014;60:167–17455. Dumas ME, Kinross J, Nicholson JK.Metabolic phenotyping and systems biologyapproaches to understanding metabolic syn-drome and fatty liver disease. Gastroenterology2014;146:46–6256. Luukkonen PK, Zhou Y, Sädevirta S, et al.Hepatic ceramides dissociate steatosis and in-sulin resistance in patients with non-alcoholicfatty liver disease. J Hepatol 2016;64:1167–117557. Yki-Järvinen H. Diagnosis of non-alcoholicfatty liver disease (NAFLD). Diabetologia 2016;59:1104–111158. Doycheva I, Cui J, Nguyen P, et al. Non-invasive screening of diabetics in primary carefor NAFLD and advanced fibrosis by MRI andMRE. Aliment Pharmacol Ther 2016;43:83–9559. Imajo K, Kessoku T, Honda Y, et al. Magneticresonance imaging more accurately classifiessteatosis and fibrosis in patients with nonalco-holic fatty liver disease than transient elastog-raphy. Gastroenterology 2016;150:626–637.e760. Cusi K, Orsak B, Bril F, et al. Long-term piogli-tazone treatment for patients with nonalcoholic

428 NAFLD in T2DM Diabetes Care Volume 40, March 2017

steatohepatitis and prediabetes or type 2 diabe-tes mellitus: a randomized trial. Ann Intern Med2016;165:305–31561. Armstrong MJ, Gaunt P, Aithal GP, et al.;LEAN trial team. Liraglutide safety and efficacyin patients with non-alcoholic steatohepatitis(LEAN): a multicentre, double-blind, randomi-sed, placebo-controlled phase 2 study. Lancet2016;387:679–69062. Neuschwander-Tetri BA, Loomba R, SanyalAJ, et al.; NASH Clinical Research Network. Farne-soid X nuclear receptor ligand obeticholic acidfor non-cirrhotic, non-alcoholic steatohepatitis(FLINT): a multicentre, randomised, placebo-controlled trial. Lancet 2015;385:956–96563. Hannah WN Jr, Harrison SA. Lifestyle anddietary interventions in the management ofnonalcoholic fatty liver disease. Dig Dis Sci2016;61:1365–137464. Promrat K, Kleiner DE, Niemeier HM, et al.Randomized controlled trial testing the effectsof weight loss on nonalcoholic steatohepatitis.Hepatology 2010;51:121–12965. Vilar-Gomez E, Martinez-Perez Y, Calzadilla-Bertot L, et al. Weight loss through lifestylemodification significantly reduces features ofnonalcoholic steatohepatitis. Gastroenterology2015;149:367–378.e566. Mummadi RR, Kasturi KS, ChennareddygariS, Sood GK. Effect of bariatric surgery on non-alcoholic fatty liver disease: systematic reviewand meta-analysis. Clin Gastroenterol Hepatol2008;6:1396–140267. Lassailly G, Caiazzo R, Buob D, et al. Bariatricsurgery reduces features of nonalcoholic steato-hepatitis inmorbidly obese patients. Gastroenter-ology 2015;149:379–38868. Mathurin P, Hollebecque A, Arnalsteen L,et al. Prospective study of the long-term effectsof bariatric surgery on liver injury in patientswithout advanced disease. Gastroenterology2009;137:532–54069. Chavez-Tapia NC, Tellez-Avila FI, Barrientos-Gutierrez T, Mendez-Sanchez N, Lizardi-Cervera J,Uribe M. Bariatric surgery for non-alcoholicsteatohepatitis in obese patients. Cochrane Da-tabase Syst Rev 2010 (1):CD00734070. Harrison SA, Fecht W, Brunt EM,Neuschwander-Tetri BA. Orlistat for overweightsubjects with nonalcoholic steatohepatitis: arandomized, prospective trial. Hepatology2009;49:80–8671. Bacchi E, Negri C, Targher G, et al. Both re-sistance training and aerobic training reduce he-patic fat content in type 2 diabetic subjects withnonalcoholic fatty liver disease (the RAED2 Ran-domized Trial). Hepatology 2013;58:1287–129572. Kirk E, Reeds DN, Finck BN, Mitra MS,Patterson BW, Klein S. Dietary fat and carbohy-drates differentially alter insulin sensitivity dur-ing caloric restriction. Gastroenterology 2009;136:1552–156073. Haufe S, Engeli S, Kast P, et al. Randomizedcomparison of reduced fat and reduced carbo-hydrate hypocaloric diets on intrahepatic fat inoverweight and obese human subjects. Hepa-tology 2011;53:1504–151474. Bril F, Ntim K, Lomonaco R, Cusi K. Treat-ment of nonalcoholic fatty liver disease (NAFLD)and nonalcoholic steatohepatitis (NASH). In In-ternational Textbook of Diabetes Mellitus. 4thed. DeFronzo RA, Ferrannini E, Alberti KGMM,

Zimmet P, Eds. Chichester, U.K., John Wiley &Sons Ltd., 2015, p. 292–30575. Bril F, Maximos M, Portillo-Sanchez P, et al.Relationship of vitamin D with insulin resistanceand disease severity in non-alcoholic steatohe-patitis. J Hepatol 2015;62:405–41176. Argo CK, Patrie JT, Lackner C, et al. Effectsof n-3 fish oil on metabolic and histological pa-rameters in NASH: a double-blind, randomized,placebo-controlled trial. J Hepatol 2015;62:190–19777. Ratziu V, Harrison SA, Francque S, et al.Elafibranor, an agonist of the peroxisome pro-liferator-activated receptor-alpha and -delta,induces resolution of nonalcoholic steatohepa-titis without fibrosis worsening. Gastroenterol-ogy 2016;150:1147–1159.e578. Sanyal AJ, Chalasani N, Kowdley KV, et al.;NASH CRN. Pioglitazone, vitamin E, or placebofor nonalcoholic steatohepatitis. N Engl J Med2010;362:1675–168579. Belfort R, Harrison SA, Brown K, et al. Aplacebo-controlled trial of pioglitazone in sub-jects with nonalcoholic steatohepatitis. N Engl JMed 2006;355:2297–230780. Aithal GP, Thomas JA, Kaye PV, et al. Ran-domized, placebo-controlled trial of pioglita-zone in nondiabetic subjects with nonalcoholicsteatohepatitis. Gastroenterology 2008;135:1176–118481. Zein CO, Yerian LM, Gogate P, et al. Pentox-ifylline improves nonalcoholic steatohepatitis: arandomized placebo-controlled trial. Hepatol-ogy 2011;54:1610–161982. Sanyal AJ, Ratziu V, Harrison S, et al.Cenicriviroc placebo for the treatment of non-alcoholic steatohepatitis with liver fibrosis: re-sults from the Year 1 primary analysis of thePhase 2b CENTAUR study (Abstract). Hepatol-ogy 2016;64:1118A83. Yau H, Rivera K, Lomonaco R, Cusi K. Thefuture of thiazolidinedione therapy in the man-agement of type 2 diabetes mellitus. Curr DiabRep 2013;13:329–34184. Gastaldelli A, Harrison SA, Belfort-Aguilar R,et al. Importance of changes in adipose tissueinsulin resistance to histological response dur-ing thiazolidinedione treatment of patients withnonalcoholic steatohepatitis. Hepatology 2009;50:1087–109385. Gastaldelli A, Harrison S, Belfort-Aguiar R,et al. Pioglitazone in the treatment of NASH: therole of adiponectin. Aliment Pharmacol Ther2010;32:769–77586. Lutchman G, Modi A, Kleiner DE, et al. Theeffects of discontinuing pioglitazone in patientswith nonalcoholic steatohepatitis. Hepatology2007;46:424–42987. Lincoff AM, Wolski K, Nicholls SJ, Nissen SE.Pioglitazone and risk of cardiovascular events inpatients with type 2 diabetes mellitus: a meta-analysis of randomized trials. JAMA 2007;298:1180–118888. Dormandy JA, Charbonnel B, Eckland DJ,et al.; PROactive Investigators. Secondary pre-vention of macrovascular events in patientswith type 2 diabetes in the PROactive Study(PROspective pioglitAzone Clinical Trial In mac-roVascular Events): a randomised controlled tri-al. Lancet 2005;366:1279–128989. Mazzone T, Meyer PM, Feinstein SB, et al. Ef-fect of pioglitazone compared with glimepiride on

carotid intima-media thickness in type 2 diabetes:a randomized trial. JAMA 2006;296:2572–258190. Nissen SE, Nicholls SJ, Wolski K, et al.;PERISCOPE Investigators. Comparison of piogli-tazone vs glimepiride on progression of coro-nary atherosclerosis in patients with type 2diabetes: the PERISCOPE randomized controlledtrial. JAMA 2008;299:1561–157391. Kernan WN, Viscoli CM, Furie KL, et al.; IRISTrial Investigators. Pioglitazone after ischemicstroke or transient ischemic attack. N Engl JMed 2016;374:1321–133192. DeFronzo RA, Tripathy D, Schwenke DC,et al.; ACT NOWStudy. Pioglitazone for diabetesprevention in impaired glucose tolerance.N Engl J Med 2011;364:1104–111593. Tuccori M, Filion KB, Yin H, Yu OH, Platt RW,Azoulay L. Pioglitazone use and risk of bladdercancer: population based cohort study. BMJ2016;352:i154194. Lewis JD, Habel LA, Quesenberry CP, et al.Pioglitazone use and risk of bladder cancer andother common cancers in persons with diabe-tes. JAMA 2015;314:265–27795. Hamaguchi E, Takamura T, Sakurai M, et al.Histological course of nonalcoholic fatty liverdisease in Japanese patients: tight glycemic con-trol, rather than weight reduction, amelioratesliver fibrosis. Diabetes Care 2010;33:284–28696. Juurinen L, Tiikkainen M, Häkkinen AM,Hakkarainen A, Yki-Järvinen H. Effects of insulintherapy on liver fat content and hepatic insulinsensitivity in patients with type 2 diabetes. Am JPhysiol Endocrinol Metab 2007;292:E829–E83597. Leiter LA, Forst T, Polidori D, Balis DA, Xie J,Sha S. Effect of canagliflozin on liver functiontests in patients with type 2 diabetes. DiabetesMetab 2016;42:25–3298. Targher G, Bertolini L, Chonchol M, et al.Non-alcoholic fatty liver disease is indepen-dently associated with an increased prevalenceof chronic kidney disease and retinopathy intype 1 diabetic patients. Diabetologia 2010;53:1341–134899. Bril F, Lomonaco R, Cusi K. The challenge ofmanaging dyslipidemia in patients with nonal-coholic fatty liver disease. Clin Lipidol 2012;7:471–481100. Blais P, Lin M, Kramer JR, El-Serag HB,Kanwal F. Statins are underutilized in patientswith nonalcoholic fatty liver disease and dysli-pidemia. Dig Dis Sci 2016;61:1714–1720101. Chalasani N, Aljadhey H, Kesterson J,MurrayMD, Hall SD. Patients with elevated liverenzymes are not at higher risk for statin hepa-totoxicity. Gastroenterology 2004;126:1287–1292102. Dongiovanni P, Petta S, Mannisto V, et al.Statin use and non-alcoholic steatohepatitis inat risk individuals. J Hepatol 2015;63:705–712103. Szendroedi J, Anderwald C, Krssak M, et al.Effects of high-dose simvastatin therapy on glu-cose metabolism and ectopic lipid deposition innonobese type 2 diabetic patients. DiabetesCare 2009;32:209–214104. Nelson A, Torres DM,Morgan AE, Fincke C,Harrison SA. A pilot study using simvastatin inthe treatment of nonalcoholic steatohepatitis: arandomized placebo-controlled trial. J Clin Gas-troenterol 2009;43:990–994105. Ginsberg HN, Elam MB, Lovato LC, et al.;ACCORD Study Group. Effects of combination

care.diabetesjournals.org Bril and Cusi 429

http://care.diabetesjournals.org

lipid therapy in type 2 diabetes mellitus. N Engl JMed 2010;362:1563–1574106. Keech A, Simes RJ, Barter P, et al.; FIELDstudy investigators. Effects of long-term fenofi-brate therapy on cardiovascular events in9795 people with type 2 diabetes mellitus (theFIELD study): randomised controlled trial. Lan-cet 2005;366:1849–1861107. Cannon CP, Blazing MA, Giugliano RP,et al.; IMPROVE-IT Investigators. Ezetimibeadded to statin therapy after acute coronarysyndromes. N Engl J Med 2015;372:2387–2397108. Takeshita Y, Takamura T, Honda M, et al.The effects of ezetimibe on non-alcoholic fattyliver disease and glucose metabolism: a randomi-sed controlled trial. Diabetologia 2014;57:878–890109. Loomba R, Sirlin CB, Ang B, et al.; San DiegoIntegratedNAFLDResearchConsortium(SINC). Eze-timibe for the treatment of nonalcoholic steatohe-patitis: assessment by novel magnetic resonanceimaging and magnetic resonance elastographyin a randomized trial (MOZART trial). Hepatology2015;61:1239–1250110. Zein CO, Unalp A, Colvin R, Liu YC,McCullough AJ; Nonalcoholic Steatohepatitis Clin-ical Research Network. Smoking and severity of

hepatic fibrosis in nonalcoholic fatty liver disease.J Hepatol 2011;54:753–759111. GohGB,PagadalaMR,Dasarathy J, et al. Renin-angiotensin system and fibrosis in non-alcoholicfatty liver disease. Liver Int 2015;35:979–985112. Georgescu EF, Ionescu R, Niculescu M,Mogoanta L, Vancica L. Angiotensin-receptorblockers as therapy for mild-to-moderate hy-pertension-associated non-alcoholic steatohe-patitis. World J Gastroenterol 2009;15:942–954113. Torres DM, Jones FJ, Shaw JC,Williams CD,Ward JA, Harrison SA. Rosiglitazone versus rosi-glitazone and metformin versus rosiglitazoneand losartan in the treatment of nonalcoholicsteatohepatitis in humans: a 12-month random-ized, prospective, open-label trial. Hepatology2011;54:1631–1639114. Cusi K. Treatment of patients with type 2diabetes and non-alcoholic fatty liver disease:current approaches and future directions. Dia-betologia 2016;59:1112–1120115. Ratziu V, Massard J, Charlotte F, et al.;LIDO Study Group; CYTOL study group. Diagnos-tic value of biochemical markers (FibroTest-FibroSURE) for the prediction of liver fibrosisin patients with non-alcoholic fatty liver dis-ease. BMC Gastroenterol 2006;6:6

116. Angulo P, Hui JM, Marchesini G, et al. TheNAFLD fibrosis score: a noninvasive system thatidentifies liver fibrosis in patients with NAFLD.Hepatology 2007;45:846–854117. Harrison SA, Oliver D, Arnold HL, Gogia S,Neuschwander-Tetri BA. Development andvalidation of a simple NAFLD clinical scoring sys-tem for identifying patients without advanceddisease. Gut 2008;57:1441–1447118. Shah AG, Lydecker A, Murray K, Tetri BN,Contos MJ, Sanyal AJ; Nash Clinical ResearchNetwork. Comparison of noninvasive markersof fibrosis in patients with nonalcoholic fattyliver disease. Clin Gastroenterol Hepatol 2009;7:1104–1112119. Sumida Y, Yoneda M, Hyogo H, et al.;Japan Study Group of Nonalcoholic Fatty LiverDisease (JSG-NAFLD). A simple clinical scoringsystem using ferritin, fasting insulin, andtype IV collagen 7S for predicting steatohepati-tis in nonalcoholic fatty liver disease. J Gastro-enterol 2011;46:257–268120. Adams LA, George J, Bugianesi E, et al. Com-plex non-invasive fibrosis models are more accu-rate than simple models in non-alcoholic fattyliver disease. J Gastroenterol Hepatol 2011;26:1536–1543

430 NAFLD in T2DM Diabetes Care Volume 40, March 2017