sign in sign up
<<
Home , Bilirubin, Liver function tests

FEATURE ARTICLE

Elevated Function Tests in Type 2

Elizabeth H. Harris, MD

iver function tests (LFTs) are postprandial states. Loss of insulin effect up-regulation of SREBP-1c and subse- commonly used in clinical prac- on the liver leads to glycogenolysis and quent simulation of de novo lipogenesis Ltice to screen for , an increase in hepatic produc- in the liver leads to increased intracellu- monitor the progression of known dis- tion. Abnormalities of stor- lar availability of , promot- ease, and monitor the effects of poten- age and lipolysis in insulin-sensitive tis- ing fatty liver. This also increases VLDL tially hepatotoxic drugs. sues such as the liver are an early mani- assembly and secretion.1 Thus, hyperin- The most common LFTs include festation of conditions characterized by sulinemia might directly lead to hepatic the aminotransferases, alkaline insulin resistance and are detectable ear- insulin resistance with associated fatty phosphatase, , , and lier than fasting . The pre- changes. . Aminotransferases, cise genetic, environmental, and meta- The excess in free fatty acids found such as alanine aminotransferase (ALT) bolic factors and sequence of events that in the insulin-resistant state is known and aspartate aminotransferase (AST), lead to the underlying insulin resistance, to be directly toxic to hepatocytes. measure the concentration of intracellu- however, is not fully understood.1 Putative mechanisms include cell lar hepatic that have leaked In animal models, chronic hyperin- membrane disruption at high concen- into the circulation and serve as a sulinemia is found to predispose the liver tration, mitochondrial dysfunction, tox- marker of hepatocyte injury. Alkaline to relative resistance to insulin. This is in formation, and activation and inhibi- phosphatase (AP), -glutamyl transpep- characterized by a failure of insulin to tion of key steps in the regulation of tidase (GGT), and bilirubin act as signal an increase in insulin receptor .3 Other potential explana- markers of biliary function and substrate-2. Upregulation of sterol regu- tions for elevated in . Albumin and prothrombin latory element-binding protein 1c insulin-resistant states include oxidant reflect liver synthetic function. (SREBP-1c) also occurs, leading to stress from reactive lipid peroxidation, The aminotransferases AST and ALT increased lipogenesis.2 Despite down- peroxisomal beta-oxidation, and are normally < 30–40 units/l. Elevations regulation of the insulin receptor sub- recruited inflammatory cells. The of aminotransferases greater than eight strate-2–mediated insulin signaling path- insulin-resistant state is also character- times the upper limit of normal reflect way in insulin-resistant states, the ized by an increase in proinflammatory either acute viral , ischemic hep- cytokines such as tumor fac- atitis, or drug- or toxin-induced liver IN BRIEF tor- (TNF-), which may also con- tribute to hepatocellular injury. In pre- injury. Much more common than Individuals with type 2 diabetes have liminary studies, an increased patients with acute hepatitis, however, a higher incidence of liver function frequency of specific TNF-–promoter are patients with chronic mild elevation test abnormalities than individuals polymorphism was found in nonalco- of aminotransferases, or AST and ALT who do not have diabetes. Mild holic (NASH) patients, < 250 units/l for > 6 months. chronic elevations of transaminases suggesting a possible genetic link or Chronic mild elevation of transami- often reflect underlying insulin resist- predisposition to fatty liver found in nases are frequently found in type 2 dia- ance. Elevation of transaminases insulin-resistant states.4 betic patients. This article will provide a within three times the upper limits of The above theories all attribute ele- review of the pathology, incidence, caus- normal is not a contraindication for vated transaminitis to direct hepatocyte es, and drug therapy related to type 2 starting oral antidiabetic or lipid- injury. It is also hypothesized that eleva- diabetic patients with elevated LFTs. modifying therapy. In contrast, antidi- tion in ALT, a gluconeogenic abetic agents have generally been whose gene transcription is suppressed Theories Behind LFT Elevation shown to decrease alanine amino- by insulin, could indicate an impairment in Diabetes transferase levels as tighter blood glu- in insulin signaling rather than purely The liver helps maintain normal blood cose levels are achieved. glucose concentration in the fasting and hepatocyte injury.5

CLINICAL DIABETES • Volume 23, Number 3, 2005 115 FEATURE ARTICLE

Can Elevated LFTs Predict the betes. The authors concluded that higher abnormal LFTs who were awaiting liver Development of Diabetes? ALT is a risk factor for type 2 diabetes biopsy. Sixty-eight of the patients had GGT is a nonspecific marker that is and indicates a potential role of type 2 diabetes; four had type 1 diabetes. known to rise in patients with type 2 increased hepatic All of the patients had or diabetes. In epidemiological studies, it and/or inflammation in the pathogenesis abnormal LFTs. They had normal blood has a positive association with alcohol of type 2 diabetes. counts, serum , and renal intake, cigarette smoking, coronary heart function. None had decompensated heart disease, BMI, systolic blood pressure, Incidence of Elevated LFTs in failure. Only 5 gave a history of social serum triglyceride, heart rate, uric acid, Diabetes drinking; the other 67 patients were clas- and hematocrit. It has an inverse associ- Salmela et al.10 studied the prevalence of sified as abstainers. ation with physical activity level.6 abnormal LFTs and their relationship to Of the 72 patients who underwent Because GGT increases in diabetes, and clinical findings in 175 unselected dia- liver biopsy, all 4 with type 1 diabetes increases as BMI increases, it has been betic outpatients in Finland. One hun- had normal liver histology, but only 5 of proposed as another marker of insulin dred and eighteen patients were classi- the 68 with type 2 diabetes had normal resistance. fied as having type 2 diabetes and 57 as liver histology. The most commonly ele- To determine whether elevated GGT having type 1 diabetes. Of those with vated LFT in the nine patients with nor- could predict the development of type 2 type 2 diabetes, 33 patients used insulin mal histology included bilirubin and AP. diabetes, a prospective cohort study of in addition to diet and oral hypo- ALT was less frequently elevated, and 7,458 nondiabetic men aged 40–59 years glycemic drugs including sulfonylurea GGT was not elevated at all. was conducted for 12 years.7 Mean and metformin. None of the patients had Of the 63 patients with abnormal liv- serum GGT at the start of the study was known , and none er histology, 48 had fatty liver or steato- significantly higher in the 194 men who had clinically significant diabetic sis with nonspecific inflammatory developed type 2 diabetes than in the rest nephropathy. A1c (A1C) changes, whereas 14 had evidence of of the cohort who did not develop dia- averaged 11.2 ± 2.4%. fibrosis. GGT and ALT were most com- betes (20.9 vs. 15.3 units/l, P < 0.0001). LFTs measured included albumin, monly elevated. As histology worsened The association was independent of total bilirubin, AST, ALT, AP, GGT, and (steatosis to inflammation to fibrosis), serum glucose and BMI. However, when serum concentrations of cholic acid and there was no significant difference in GGT was added to a model for predict- chenodeoxycholic acid. Fifty-seven per- mean values of ALT and GGT. There- ing the development of type 2 diabetes, it cent of the 175 diabetic outpatients (100 fore, although abnormal LFT results are did not improve the power of BMI and subjects) had at least one abnormal LFT; common in diabetes, especially in over- glucose for predicting the development 27% (48 subjects) had at least two abnor- weight type 2 diabetic patients, they are of type 2 diabetes. mal tests. The type 2 diabetic patients not reliable in predicting histological Ohlson et al.8 found elevated ALT in more frequently had elevated ALT (22.9 changes in the liver. nondiabetic Swedish men to be a risk vs. 5.3%) and GGT (23.7 vs. 10.5%) lev- In a larger study, Erbey et al.11 ana- factor for type 2 diabetes, independent of els than those with type 1 diabetes. On lyzed 18,825 noninstitutionalized obesity, body distribution, plasma the other hand, patients with type 1 dia- patients within the United States with an glucose, lipid, AST, bilirubin concentra- betes more frequently had elevated oversampling of African Americans and tions, and family history of diabetes. bilirubin levels (21.1 vs. 10.2%). Howev- Mexican Americans. Of the total sample, With similar results, Vozaroza et al.9 fol- er, increases in LFTs were rarely more 4.1% had elevated ALT, and 6.7% had lowed 451 nondiabetic Pima Indians for than twice the upper limit of normal. type 2 diabetes. Of those with type 2 dia- an average of 6.9 years to determine Multivariate analysis showed BMI betes, the prevalence of elevated ALT whether hepatic enzyme elevations could > 25 kg/m2 and poor diabetic control was 7.8%, compared to a 3.8% preva- be linked to the development of type 2 (fasting blood glucose > 216 mg/dl) lence in those without diabetes. The diabetes. At baseline, ALT, AST, and were the most significant clinical vari- prevalence of ALT elevation greater than GGT were related to percent body fat. ables associated with elevated ALT and three times normal was not significantly After adjustment for age, sex, body fat, GGT. Elevated ALT was also associated different between the nondiabetic and whole body insulin sensitivity, and acute with onset of diabetes within the past 4 diabetic patients (0.4 vs. 0.7%). Those insulin response, only elevated ALT at years, mature onset of diabetes (35–51 who were overweight (BMI 25–30 baseline was associated with an increase years), and use of diet or sulfonylurea. kg/m2) and obese (BMI > 30 kg/m2) in hepatic glucose output. Prospectively, To investigate the reliability of LFTs were more likely to have elevated ALT. increasing ALT concentrations were in assessing histological changes, There was a 10.6% prevalence in obese associated with a decline in hepatic Salmela et al.10 looked at 72 consecutive diabetic patients versus a 6.6% preva- insulin sensitivity and risk of type 2 dia- diabetic inpatients with hepatomegaly or lence in obese nondiabetic patients.

116 Volume 23, Number 3, 2005 • CLINICAL DIABETES FEATURE ARTICLE

Nonalcoholic al antibody, anti–smooth muscle anti- ease, including individuals with dia- The most common cause of elevated body), or hereditary causes of liver dis- betes, the rates of elevated ALT level LFTs in type 2 diabetic patients is non- ease (-1 antitrypsin, , iron, above twice the upper limit of normal alcoholic fatty liver disease (NAFLD). iron-binding capacity, or ).13 They were 1.8% in the simvastatin group and NAFLD is a clinicopathological condi- had no history of alcohol or hepatotoxic 1.6% in the placebo group. This was not tion representing a spectrum of histolog- drug use nor signs of chronic liver disease. a significant difference.18 In the ical findings from hepatic steatosis or fat Patients also had no evidence of sarcoid Pravastatin in Elderly Individuals at accumulation in hepatocytes without on chest X-ray. Of note, there is no men- Risk of Vascular Disease (PROSPER) inflammation, to hepatic steatosis with a tion in the article of evaluating unlikely trial,19 only one patient in the pravastatin necroinflammatory component that may but potential causes of transaminitis, such group and one patient in the placebo or may not have fibrosis, or NASH. as muscle disorders, adrenal insufficiency, group had ALT or AST level above three NAFLD is defined by the absence of and celiac disease. Of the 81 “marker neg- times the upper limits of normal. None or minimal alcohol consumption, liver ative” patients with no identified etiology of the patients had rhabdomyolysis. biopsy showing macrovesicular steatosis of elevated liver enzymes, 73 had abnor- Thirty-six patients in the pravastatin with or without necro-inflammatory mal histology, all with some degree of group had myalgias, compared to 32 activity, and exclusion of other forms of steatosis. In the patients without clear eti- patients in the placebo group. liver disease. Although the pathogenesis ology of liver disease, the prevalence rate High-dose therapy is associat- is still unclear, it is characterized by of steatosis and steatohepatitis was 50.6 ed with more frequent abnormalities of accumulation of triglycerides within the and 32%, respectively. LFTs, although they are generally still hepatocytes. Insulin resistance is thought In a similar study, 354 patients, both relatively infrequent. In the Treating to to play an important role in the triglyc- with and without diabetes, underwent New Targets (TNT) trial,20 patients with eride accumulation. Excess intracellular liver biopsy to investigate abnormal clinical (CVD) fatty acids, oxidant stress, ATP deple- LFTs. When patients with clinical or were randomized to 10 or 80 mg of ator- tion, and mitochondrial dysfunction all serological evidence of a specific diag- vastatin. The incidence of persistent ele- contribute to hepatocyte injury and nosis were excluded, 66% of the patients vation in ALT, AST, or both (defined as inflammation followed by fibrosis.3 had evidence of steatosis and steatohep- two consecutive measurements obtained Not surprisingly, the most common atitis on biopsy.14 4–10 days apart that were more than laboratory abnormality in patients with three times the upper limit of the normal NAFLD is mild to moderate elevation of and Type 2 Diabetes range) was 0.2 and 1.2%, respectively serum aminotransferases. As in the histo- Hepatitis C virus (HCV), the leading (P < 0.001).20 logical study of diabetic patients with cause of liver disease in the United Because of large trials such as these, abnormal LFTs by Salmela et al.,10 level States, is a known independent predictor current recommendations from the of elevation in NAFLD does of type 2 diabetes, the most common American College of Physicians suggest not predict severity of liver histology.12 endocrine disease even in patients with- that type 2 diabetic patients with other out .15,16 HCV is known to have cardiovascular risk factors should take a NAFLD in Nondiabetic Patients a higher prevalence within diabetic statin for primary prevention of NAFLD is replacing alcohol and viral patients. When comparing 176 diabetic macrovascular complications. These hepatitis as the most common etiology patients to 6,172 blood donors matched patients do not need routine monitoring of chronically elevated LFTs in the for recognized risk factors of acquiring of LFTs while on unless they United States in both diabetic and non- HCV, there was a higher prevalence of have baseline abnormalities in LFTs, diabetic individuals.3 Of those patients HCV infection within the diabetic myopathy, or are taking other drugs that with NAFLD, 60–95% are obese, patients (11.5 vs. 2.5%, P < 0.001).17 Of could increase their risk of adverse 28–55% have type 2 diabetes, and the diabetic patients with HCV, 72.3% events.21 20–92% have hyperlipidemia. had abnormally elevated LFTs, compared For diabetic patients with baseline In a prospective study of 1,124 adults to 27.7% of diabetic patients without evi- transaminases less than three times the who were referred for evaluation of dence of HCV (P < 0.001). This would upper limit of normal, it is not con- chronically elevated LFTs, 81 were deter- suggest that any diabetic patient with ele- traindicated to initiate, continue, or mined to have unknown etiology based on vated LFTs needs screening for HCV. advance statin therapy as long as patients the absence of serum markers for infec- are carefully monitored.22 The frequency tious (including hepatitis B and C), meta- Statins in Type 2 Diabetic Patients of required monitoring in these patients bolic (thyroid-stimulating hormone), With is under debate. There is also a debate as autoimmune (serum protein electrophore- In the Heart Protection Study of 20,536 to whether transaminase elevation in sis, antinuclear antibody, antimitochondri- high-risk individuals with vascular dis- statin therapy even constitutes true hepa-

CLINICAL DIABETES • Volume 23, Number 3, 2005 117 FEATURE ARTICLE

toxicity.22 For diabetic patients over the people with type 2 diabetes on pioglita- effects were observed in the 5,006 age of 40 years, and certainly in the set- zone versus other oral antidiabetic patients who took rosiglitazone. The per- ting of multiple cardiovascular risk fac- agents. Patients were classified into centage of patients who developed ALT tors or known CVD, the potential risk of treatment groups based on the antidia- greater than three times the upper limit statin therapy from the perspective of betic therapy received on the date of of normal were 0.32% of the rosiglita- is far outweighed by the their first antidiabetic pharmacy claim. zone group, 0.17% of the placebo group, proven benefit from CVD risk reduction. Patients in the pioglitazone group were and 0.40% for the group taking either then matched to patients within the sulfonylurea, metformin, or insulin. The Oral Antidiabetic Agents in Type 2 rosiglitazone, sulfonylurea, and met- respective incidence rates of 0.29, 0.59, Diabetic Patients With Elevated formin groups. Matched groups were and 0.64/100 person-years show no dif- Transaminases similar in demographic and clinical char- ference between treatment of rosiglita- The introduction of the insulin sensitiz- acteristics. Patients evaluated included zone, placebo, and other antihyper- er troglitazone and subsequent cases of 4,458 matched pairs of pioglitazone- glycemic agents and the development of hepatotoxicity led Jick et al.23 to inves- versus rosiglitazone-treated patients, ALT levels greater than three times the tigate the baseline risk of liver disease 1,474 pairs of pioglitazone- versus sul- upper limit of normal.26 Furthermore, of in type 2 diabetic patients on oral fonylurea-treated patients, and 1,137 the 5.6% of the patients whose serum agents other than .23 pairs of pioglitazone- versus metformin- ALT values were between one and two Researchers identified 40,190 type 2 treated patients. Over a 2-year period, and a half times the upper limit of nor- diabetic individuals within the U.K.- patients on pioglitazone had no mal at baseline, 66% of those treated based General Practice Research increased risk of liver failure or hepatitis with antihyperglycemic medicines nor- Database who were treated with oral beyond that of patients on the other malized their ALTs, whereas only 38.7% diabetic agents, including sulfony- antidiabetic agents. of those treated with placebo normalized lureas, metformin, guar gum, and acar- In placebo-controlled, double-blind ALT levels.26 This supports the impor- bose, between 1989 and 1996 and who clinical trials with pioglitazone, the inci- tant link among glycemic control, insulin had no known liver disease at the time dence of elevated ALT values greater resistance, and hepatic function and sug- oral therapy was begun. Of the total than three times the upper limit of nor- gests that improved glycemic control sample, 1.5% (or 605) were given a mal was virtually identical between and improvement in insulin resistance new diagnosis of liver disorder during patients on pioglitazone and those on can reduce mild chronic elevation of the study period. Of those 605 cases, placebo (0.26 vs. 0.25%). Enzyme eleva- transaminitis often found in diabetic 249 (41.2%) were attributed to a pre- tions were reversible in all patients who patients. disposing condition, 186 (31%) were developed elevated ALT on The decrease in LFTs demonstrated mild asymptomatic liver enzyme pioglitazone.25 with rosiglitazone and pioglitazone ther- abnormalities that were not considered Lebovitz et al.26 studied more than apy in diabetic patients has also been clinically relevant, and 113 (18.7%) 6,000 patients with type 2 diabetes in a shown in pilot studies using thiadolidine- had a specific nondrug etiology listed. double-blind clinical trial using various diones to treat NASH, a surrogate for The remaining 57 (8.7%) with clinical- doses of rosiglitazone, placebo, and insulin resistance. One study27 placed 18 ly relevant liver disease and no identi- either glyburide, metformin, or insulin. nondiabetic patients with NASH on fied predisposing condition or cause Mean A1C levels at the start of the study pioglitazone, 30 mg daily, for 48 weeks. were attributed to other drugs, fatty were similar across all groups (8.5–9%). By the end of the study, serum ALT lev- liver, and unknown. Oral antidiabetic Measurement of liver enzymes occurred els decreased in all patients and normal- agents could not be ruled out in two at screening, baseline, then every 4 ized in 72% of them. Serum ALT levels cases, giving an incidence of 0.002/100 weeks for the first 3 months of treatment fell from an average of 99 units/L at person-years. and at 6- to 12-week intervals thereafter. baseline to 40 units/L at 48 weeks. The idiosyncratic hepatic reaction Patients were excluded from the study if Another study28 used rosiglitazone, 4 leading to hepatic failure and death in they had ALT, AST, or ALP greater than mg twice daily, for 48 weeks in the treat- some patients on troglitazone before its two and a half times the upper limit of ment of 30 patients with NASH, 50% of removal from the market is not likely normal at screening. This is consistent whom had either impaired glucose toler- to be a class effect of the peroxisomal with current recommendations of when ance or diabetes. Of the 25 patients who proliferator-activated receptor-gamma not to use rosiglitazone or pioglitazone. completed the study, all had significant agonists. Of those on rosiglitazone, ~ 3,800 improvements in mean serum ALT lev- Rajagopalan et al.24 performed a ret- were monitored for at least 6 months, els, changing from a baseline of 104 rospective analysis of claims data com- 2,800 for at least 1 year, and 1,000 for at units/l to 42 units/l at 48 weeks. At 72- paring the incidence of liver failure in least 2 years. No evidence of hepatotoxic week follow-up, after 24 weeks off

118 Volume 23, Number 3, 2005 • CLINICAL DIABETES FEATURE ARTICLE

rosiglitazone, liver enzyme levels had to mixed insulin resistance and sensitivity in liv- 18Heart Protection Study Collaborative ers of lipodystrophic and ab/ab mice. Mol Cell Group: MRC/BHF Heart Protection Study of increased to near pretreatment levels. 6:77–86, 2000 lowering with simvastatin in 20,536

3 high-risk individuals: a randomized placebo-con- Neuschwander-Tetri BA, Caldwell S: Nonal- trolled trial. Lancet 360:7–22, 2002 Conclusions coholic steatohepatitis: summary of AASLD sin- Individuals with type 2 diabetes have a gle topic conference. Hepatology 37:1202–1219, 19Shepherd J, Blauw GJ, Murphy MB, Bollen 2003 EL, Buckley BM, Cobbe SM, Ford I, Gaw A, higher incidence of LFT abnormalities Hyland M, Jukema JW, Kamper AM, Macfarlane 4Grove J, Daly AK, Bassendine MF, Day CP: than individuals who do not have dia- Association of a tumor necrosis factor promoter PW, Meinders AE, Norrie J, Packard CJ, Perry IJ, betes. The most common abnormality is polymorphism with susceptibility to alcoholic Stott DJ, Sweeney BJ, Twomey C, Westendorp steatohepatitis. Hepatology 26:143–146, 1997 RG, the PROSPER study group: Pravastatin in elevated ALT. Any diabetic patient elderly individuals at risk of vascular disease found to have a mild chronic elevation 5O’Brien RM, Granner DK: Regulation of (PROSPER): a randomized controlled trial. gene expression by insulin. Biochem J Lancet 360:1623–1630, 2002 of ALT, or elevation of ALT ≤ 250 278:609–619, 1991 20LaRosa JC, Grundy SM, Waters DD, Shear units/l for > 6 months should have 6Wannamethee G, Ebrahim S, Shaper AG: C, Barter P, Fruchart JC, Gotto AM, Greten H, screening for treatable causes of chronic Gamma-glutamyltransferase: determinants and Kastelein JJ, Shepherd J, Wenger NK, the Treat- liver disease, particularly hepatitis B, associations with mortality from ischaemic heart ing to New Targets (TNT) Investigators: Intensive disease and all causes. Am J Epidemiol lipid lowering with in patients with hepatitis C, and hemochromatosis, 42:699–708, 1995 stable coronary disease. N Engl J Med which are found with increased inci- 7Perry IJ, Wannamethee SG, Shaper AG: 352:1425–1435, 2005 dence in type 2 diabetes. In patients for Prospective study of serum gamma-glutamyl- 21 transferase and risk of NIDDM. Diabetes Care Snow V, Aronson M, Hornbake R, Mottur- whom a directed medical history and 21:732–737, 1998 Pilson C, Weiss K: Lipid control in the manage- ment of type 2 diabetes mellitus: a clinical prac- physical examination do not raise suspi- 8Ohlson LO, Larsson B, Bjorntorp P, Erksson tice guideline from the American College of cion of other causes of elevated LFTs, H, Svardsudd K, Welin L, Tibblin G, Wilhelmsen Physicians. Ann Intern Med 140:644–650, 2004 L: Risk factors for type 2 diabetes mellitus: thir- such as , alcohol, autoimmu- 22 teen and one-half years of follow-up of the partic- Pasternak RC, Smith SC Jr, Bairey-Merz nity, metabolic etiology, or hereditary CN, Grundy SM, Cleeman JI, Lenfant C, for the ipants in a study of Swedish men born in 1913. American College of Cardiology, American Heart etiology, and for those who have no evi- Diabetologia 31:798–305, 1988 Association and National Heart, Lung and Blood dence of more serious liver disease, such 9Vozarova B, Stefan N, Lindsay RS, Saremi Institute: ACC/AHA/NHLBI clinical advisory on the use and safety of statins. Stroke elevations in bilirubin or prothrombin A, Pratley RE, Bogardus C, Tataranni PA: High alanine aminotransferases is associated with 33:2337–2341, 2002 time or decreases in albumin, further decreased hepatic insulin sensitivity and predicts 23Jick SS, Stender M, Myers M: Frequency of diagnostic workup is probably not the development of type 2 diabetes. Diabetes liver disease in type 2 diabetic patients treated 51:1889–1895, 2002 with oral antidiabetic agents. Diabetes Care required. 22:2067–2071, 1999 Routine monitoring of LFTs in 10Salmela PI, Sotaniemi EA, Niemi M, Maen- tausta O: in diabetic patients. 24Rajagopalan R, Iyer S, Perez A: Compari- patients with type 2 diabetes should Diabetes Care 7:248–254, 1984 son of pioglitazone with other antidiabetic drugs for associated incidence of liver failure: no evi- occur at the start of drug therapy and if 11 Erbey JR, Silberman C, Lydick E: Preva- dence of increased risk of liver failure with patients develop symptoms raising con- lence of abnormal serum alanine aminotrans- pioglitazone. Diabetes Obes Metab 7:161–169, ferase levels in obese patients and patients with 2005 cern about hepatic impairment. Beyond type 2 diabetes. Am J Med 109:588–590, 2000 25 that, periodic screening will have to be 12 ACTOS Prescribing Information [article Alba LM, Lindor K: Review Article: Non- online]. Lincolnshire, Ill., Takeda Pharmaceuti- based on clinical judgment, keeping in alcoholic fatty liver disease. Aliment Pharmacol cals North America, 2003. Available from mind that elevation of transaminases Ther 17:977–986, 2003 http://www.actos.com/pi.pdf does not always correlate with histologi- 13Daniel S, Ben-Menachem T, Vasudevan G, 26Lebovitz H, Kreider M, Freed M: Evalua- cal changes in the liver. Elevation of Ma D, Blumenkehl M: Prospective evaluation of tion of liver function in type 2 diabetic patients unexplained chronic liver transaminase abnormal- during clinical trials: evidence that rosiglitazone ALT within three times the upper limit ities in asymptomatic and symptomatic patients. does not cause hepatic dysfunction. Diabetes of normal is not a contraindication for Am J Gastroenterol 94:3010–3014, 1999 Care 25:815–821, 2002 27 starting any oral antidiabetic or lipid- 14Skelly MM, James PD, Ryder SD: Findings Promrat K, Lutchman G, Uwaifo GI, Freed- modifying therapy. In contrast, antidia- on liver biopsy to investigate abnormal liver func- man RJ, Soza A, Heller T, Doo E, Ghany M, tion tests in the absence of diagnostic . J Premkumar A, Park Y, Liang TJ, Yanovski J, betic agents have generally been shown Hepatol 35:195–199, 2001 Kleiner D, Hoofnagle JH: A pilot study of piogli- to decrease ALT levels as tighter blood tazone treatment for nonalcoholic steatohepatitis. 15Baig NA, Herrine SK, Rubin R: Liver dis- Hepatology 39:188–195, 2003 glucose levels are achieved. ease and diabetes mellitus. Clin Lab Med 28 21:193–207, 2001 Neushwander-Tetri BA, Brunt EM, Wehmeier KR, Oliver D, Bacon BR: Improved REFRENCES 16Knobler H, Schihmanter R, Zifroni A, nonalcoholic steatohepatitis after 48 weeks of Finakel G, Schattner A: Increased risk of type 2 treatment with the PPAR gamma ligand rosiglita- 1Lewis GF, Carpentier A, Khosrow A, Giacca diabetes in noncirrhotic patients with chronic zone. Hepatology 38:1008–1016, 2004 A: Disordered fat storage and mobilization in the hepatitis C virus infection. Mayo Clin Proc pathogenesis of insulin resistance and type 2 dia- 75:355–359, 2000 betes. Endocr Rev 23:201–229, 2002 17Simo R, Hernandez C, Genesca J, Jardi R, Elizabeth H. Harris, MD, is a fellow at 2Shimomura I, Matsuda M, Hammer RE, Mesa J: High prevalence of hepatitis C virus Bashmakov Y, MS, Goldstein JL: infection in diabetic patients. Diabetes Care the University of North Carolina, Chapel Decreased IRS-2 and increased SREBP-1c lead 19:998–1000, 1996 Hill.

CLINICAL DIABETES • Volume 23, Number 3, 2005 119