Elevated serum alanine transaminase in patients with type 1 or type 2 diabetes mellitus

Q J Med 2006; 99:871 876 doi:10.1093/qjmed/hcl116 Elevated serum alanine transaminase in patients with type 1 or type 2 diabetes mellitus J. WEST 1, J. BROUSIL 2, A. GAZIS 3, L. JACKSON 2, P. MANSELL 3, A. BENNETT 4 and G.P. AITHAL 2 From the 1 Division of Epidemiology & Public Health, 2 Wolfson Digestive Diseases Centre and 4 Department of Biomedical Sciences, University of Nottingham Medical School, 3 Divisions of Diabetes & Endocrinology and Queen s Medical Centre, Nottingham, UK Received 25 April 2006 and in revised form 3 July 2006 Summary Background: The risk of chronic liver disease is higher in diabetics, and serum alanine transaminase (ALT) is a sensitive predictor of mortality from liver disease. Aim: To estimate the prevalence of elevated ALT in type 1 and type 2 diabetes, and identify possible risk factors. Methods: We identified all patients (n ¼ 2077) attending review between September 2002 and August 2003. We excluded those with no ALT measurement (n ¼ 73); those whose alcohol consumption was 414 units/week (women) (n ¼ 276) or 421 units/week (men) (n ¼ 324); and those with a diagnosis of Maturity Onset Diabetes of the Young, secondary diabetes, gestational diabetes or uncertain type of diabetes (n ¼ 51). We calculated the prevalences of elevated ALT in both type 1 and type 2 diabetes patients, and compared the Introduction Diabetes mellitus (DM) is associated with non-alcoholic fatty liver disease (NAFLD) including its severe form, non-alcoholic steatohepatitis (NASH). 1,2 Among patients with diabetes, the risk of chronic liver disease is doubled, independent of alcoholic liver disease or viral hepatitis. 3 Diabetes also increases the risk of primary liver cancers 3 5 and death from liver cirrhosis. 6,7 However, the prevalence of previously undetected liver disease demographic, microvascular risk factors and current drug use between each group using multivariate logistic regression. Results: Of 1353 patients included, 836 (61.9%) had type 2 diabetes. Elevated ALT was found in 9.5% (95%CI 7.1 12.3%) of patients with type 1 diabetes, and 12.1% (95%CI 9.9 14.5%) of those with type 2 diabetes. The risk of elevated ALT in patients with type 2 diabetes increased with increasing body mass index (p trend ¼ 0.04), and was lower in those taking insulin (OR 0.38, 95%CI 0.22 0.65). Discussion: The prevalence of elevated ALT is 3 4 times higher in patients with either type 1 or type 2 diabetes than in the general population. Further studies investigating the aetiology and mechanisms of this elevation may suggest appropriate early interventions. in a large unselected diabetic population has not been examined. 7 Alanine transaminase (ALT) is a widely available serum marker of liver disease. Even a minor elevation of ALT is a good predictor of mortality from liver disease. 8 In previous reports, including one from our centre, NAFLD was the most common diagnosis, accounting for 64 90% of cases, when detailed investigations were carried out in patients Address correspondence to Dr G.P. Aithal, Consultant Hepatobiliary Physician, D Floor, South Block, Queen s Medical Centre University Hospital, Nottingham NG7 2UH, UK. email: guru.aithal@qmc.nhs.uk! The Author 2006. Published by Oxford University Press on behalf of the Association of Physicians. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org

872 J. West et al. with asymptomatic elevated liver enzymes. 9 11 In a large population-based study analysing data from the Third National Health and Nutrition Examination Survey (NHANES III), ALT elevation was associated with a number of risk factors of NAFLD, such as impaired glucose metabolism, insulin resistance, central obesity, high leptin and triglycerides, even when patients with known diabetes (who are at high risk for NAFLD) were excluded. 12 Hence, raised ALT has been used as a marker of NAFLD in epidemiological studies. 13 In the current study, we have estimated the prevalence of and risk factors for elevated ALT in both Type 1 and Type 2 diabetes. Methods Queen s Medical Centre serves a population of about 350 000. Of prevalent cases, about 15% of type 2 and 95% of type 1 diabetes attend hospital diabetes services at any one time. From our Diabetic Registry, we identified all those who attended for review between September 2002 and August 2003. Data collected prospectively included age, gender, type of diabetes, height, weight, blood pressure, self-reported alcohol intake at the time of registration on the database (and updated during clinical practice), microvascular risk factors and current medications. During the study period, the routine laboratory measurements included: urea and electrolytes; HbA 1c ; lipid profile and liver profile (including ALT, -glutamyl transferase, alkaline phosphatase, bilirubin and albumin). ALT was estimated using Vitros dry-slide analysis (Johnson & Johnson, Ortho Clinical Diagnostics). Hypertension was defined as either current use of anti-hypertensive medication or blood pressure 4140/80 mmhg. Microalbuminuria was defined as either dipstick positive for protein or albumin/ creatinine ratio (in mg/mmol) of 42.5 for women or 43.5 for men. Dyslipidaemia was defined as at least one of: current use of lipid-lowering medication; serum cholesterol 45 mmol/l; LDL 43 mmol/l; triglyceride 42.3 mmol/l; HDL <1 mmol/l. 14 We categorized body mass index (BMI) into standard categories, current age into 10-year age bands, and HbA 1c into quartiles. We defined elevated ALT as 450 IU/l. Statistical analyses We calculated the prevalence and 95%CIs of elevated ALT in both type 1 and type 2 diabetes, and compared the demographic, microvascular risk factors and current drug use between each group. To assess the risk factors for elevated ALT, we used multivariable logistic regression modelling, including all covariates. In the multivariate analyses, we only included those people with no missing values. To examine linear trends, we fitted categorical variables as quantitative variables and assessed goodness-of-fit of the nested models using likelihood ratio tests. All analyses used Stata 7.0/SE. Results We identified 2077 people (of 2999 on the database at the time of the study) who attended for review during the 1-year study period. We excluded those who did not have ALT measured (n ¼ 73), those whose alcohol consumption exceeded 14 units/week for women (n ¼ 276) or 21 units/ week for men (n ¼ 324), and those with a diagnosis of Maturity Onset Diabetes of the Young, secondary diabetes, gestational diabetes or uncertain type of diabetes (n ¼ 51). Of the 1353 patients included, 836 (61.9%) had type 2 diabetes. Elevated ALT was found in 9.5% (95%CI 7.1 12.3%) of those with type 1 diabetes and 12.1% (95%CI 9.9 14.5%) of those with type 2 diabetes. Table 1 shows the characteristics of the included subjects by ALT category, and the unadjusted risk estimates for association with each characteristic. There were missing values in only the body mass index and HbA 1c variables. Table 2 shows the multivariable analyses of the risk of elevated ALT from various characteristics. The prevalence of elevated ALT decreased with increasing age in type 2 diabetes (although not statistically significant at the 5% level). In type 2 diabetes, the risk of elevated ALT increased with increasing BMI category (430 vs. 425, OR 2.76, 95%CI 0.91 8.38, p trend ¼ 0.04) and decreased with current treatment with insulin (OR 0.38, 95%CI 0.22 0.65). Discussion The prevalences of elevated ALT in our type 1 (9.5%) and type 2 (12.1%) diabetes patients were both considerably higher than the 2.7% expected in the general population, 8 and higher than 5.6% reported at baseline in clinical trials. 15 This suggests that about 10% of diabetes patients under regular review in secondary care may need further investigation for the causes of raised ALT, and has potential resource implications. However, these prevalences of raised ALT may still be underestimates, for two reasons. Firstly, a single estimation of ALT was used to define abnormality. Fluctuation in the ALT levels

ALT elevation in diabetes 873 Table 1 Numbers, proportions and unadjusted ORs for abnormal ALT in patients with type 1 or type 2 diabetes, for various characteristics Diabetes... Type 1 (n ¼ 517) Type 2 (n ¼ 836) ALT... Normal (n ¼ 468) Abnormal (n ¼ 49) OR (95%CI) Normal (n ¼ 735) Abnormal (n ¼ 101) OR (95%CI) Age group (years) 15 29 (93.6%) 2 (6.5%) 0.49 (0.11 2.21) 1 (50.0%) 1 (50.0%) 5.89 (0.34 102.88) 25 112 (91.8%) 10 (8.2%) 0.63 (0.28 1.42) 9 (69.2%) 5 (35.7%) 3.27 (0.89 12.02) 35 127 (87.6%) 18 (12.4%) 1 53 (85.5%) 9 (14.5%) 1 45 96 (88.9%) 12 (11.1%) 0.88 (0.41 1.92) 100 (81.3%) 23 (18.7%) 1.35 (0.59 3.14) 55 56 (98.3%) 1 (1.8%) 0.13 (0.02 0.97) 206 (85.1%) 36 (14.9%) 1.03 (0.47 2.27) 65 34 (87.2%) 5 (12.8%) 1.04 (0.36 3.00) 242 (92.4%) 20 (7.6%) 0.49 (0.21 1.13) 75 14 (93.3%) 1 (6.7%) 0.50 (0.06 4.07) 124 (94.7%) 7 (5.3%) 0.33 (0.12 0.94) Gender Female 226 (93.4%) 16 (6.6%) 1 342 (90.7%) 35 (9.3%) 1 Male 242 (88.0%) 33 (12.0%) 1.93 (1.03 3.59) 393 (85.6%) 66 (14.4%) 1.64 (1.06 2.53) BMI 425 146 (91.8%) 13 (8.2%) 1 85 (95.5%) 4 (4.5%) 1 25 30 166 (91.7%) 15 (8.3%) 1.01 (0.47 2.20) 189 (89.2%) 23 (10.8%) 2.59 (0.87 7.71) 430 98 (94.2%) 6 (5.8%) 0.69 (0.25 1.87) 347 (87.4%) 50 (12.6%) 3.06 (1.08 8.71) Missing 58 (79.5%) 15 (20.5%) 114 (82.6%) 24 (17.4%) Microalbuminuria No 268 (92.4%) 22 (7.6%) 1 300 (89.0%) 41 (12.0%) 1 Yes 200 (88.1%) 27 (11.9%) 1.64 (0.91 2.97) 435 (88.9%) 60 (12.1%) 1.01 (0.66 1.54) Hypertension No 213 (91.4%) 20 (8.6%) 1 89 (84.8%) 16 (15.2%) 1 Yes 255 (89.8%) 29 (10.2%) 1.21 (0.67 2.20) 646 (89.4%) 85 (11.6%) 0.73 (0.41 1.30) Dyslipidaemia No 181 (93.3%) 13 (6.7%) 1 93 (88.6%) 12 (11.4%) 1 Yes 287 (88.8%) 36 (11.2%) 1.75 (0.90 3.38) 642 (87.8%) 89 (12.2%) 1.07 (0.57 2.04) Metformin No 453 (90.8%) 46 (9.2%) 1 441 (88.5%) 57 (11.5%) 1 Yes 15 (83.3%) 3 (16.7%) 1.97 (0.55 7.06) 294 (87.0%) 44 (13.0%) 1.16 (0.76 1.76) Glitazone No 468 (90.5%) 49 (9.5%) 714 (87.9%) 98 (12.1%) 1 Yes 0 (0.0%) 0 (0.0%) 21 (87.5%) 3 (12.5%) 1.04 (0.30 3.55) Insulin No 0 (0.0%) 0 (0.0%) 242 (81.5%) 55 (18.5%) 1 Yes 468 (90.5%) 49 (9.5%) 493 (92.5%) 46 (8.5%) 0.41 (0.27 0.63) HbA 1c quartile 47.67 84 (87.5%) 12 (12.5%) 1 202 (88.6%) 26 (11.4%) 1 47.67 117 (92.1%) 10 (7.9%) 0.6 (0.25 1.45) 171 (87.2%) 25 (12.8%) 1.14 (0.63 2.04) 48.57 123 (93.9%) 8 (6.1%) 0.46 (0.18 1.16) 164 (86.8%) 20 (10.6%) 1.18 (0.66 2.13) 49.55 119 (91.5%) 11 (8.5%) 0.65 (0.27 1.54) 169 (89.4%) 20 (10.6%) 0.92 (0.50 1.71) Missing 25 (75.8%) 8 (24.2%) 29 (85.3%) 5 (14.7%) Reference category. is recognized in patients with chronic liver disease, and a single measurement can underestimate disease burden. 13 Secondly, we defined ALT 450 U/l as abnormal, according to our laboratory range, which was established from 2 SD of ALT estimations in apparently healthy subjects. This concept of establishing normal ranges for ALT based on apparently healthy subjects has been challenged; recent studies have argued that cut-off values identified using normograms of ALT estimations in apparently normal subjects may miss occult chronic liver disease, and underestimate the population at risk of mortality from liver diseases. 9 We acknowledge that our pragmatic approach to defining elevated ALT may have underestimated the prevalence of NAFLD slightly.

874 J. West et al. Table 2 Risk of abnormal ALT in patients with type 1or type 2 diabetes, mutually adjusted for all variables (patients with missing values excluded) Diabetes... Type 1 (n ¼ 433) Type 2 (n ¼ 680) Covariate n OR (95%CI) n OR (95%CI) Age group (years) 15 19 0.61 (0.07 5.22) 0 25 100 0.92 (0.34 2.47) 12 3.63 (0.76 17.42) 35 120 1 50 1 45 94 0.76 (0.27 2.14) 100 2.07 (0.73 5.82) 55 52 0.18 (0.02 1.59) 200 1.28 (0.46 3.52) 65 35 1.14 (0.31 4.15) 223 0.56 (0.19 1.67) 75 13 0.67 (0.07 6.07) 95 0.77 (0.22 2.73) Gender Female 204 1 307 1 Male 229 1.63 (0.75 3.53) 373 1.54 (0.90 2.63) BMI 425 154 1 85 1 25 30 178 0.91 (0.40 2.07) 206 2.30 (0.74 7.15) 430 101 0.49 (0.15 1.56) 389 2.76 (0.91 8.38) Microalbuminuria No 258 1 294 1 Yes 175 1.84 (0.84 4.00) 386 0.88 (0.51 1.52) Dyslipidaemia No 175 1 91 1 Yes 258 1.29 (0.55 3.03) 589 0.98 (0.45 2.14) Hypertensive No 197 1 82 1 Yes 236 1.34 (0.57 3.16) 598 0.94 (0.43 2.03) Metformin No 419 1 403 1 Yes 14 2.01 (0.20 19.83) 277 0.81 (0.46 1.40) Glitazone No 659 1 Yes 21 0.54 (0.12 2.52) Insulin No 229 1 Yes 451 0.38 (0.22 0.65) HbA 1c quartile 47.67 83 1 185 1 47.67 114 0.79 (0.27 2.25) 169 1.67 (0.80 3.49) 48.57 120 0.61 (0.20 1.81) 163 1.95 (0.93 4.08) 49.55 116 0.83 (0.29 2.39) 163 1.54 (0.71 3.34) Reference category. As the study included patients attending their hospital review, there is likely to be some selection bias. This may be particularly true of patients with type 2 diabetes, as only 15% of prevalent cases are under hospital review. A variety of factors, including referral pattern, co-morbidity and poor control, can influence the selection of patients for regular hospital review. Although we have controlled for this to some extent by including HbA 1c in our multivariable analysis, findings of the current study may not be true of an unselected population of type 2 diabetes and will require confirmation in population-based studies. It is unlikely that selection bias would have affected the findings in type 1 diabetes as 95% of such patients are under regular hospital review. Although we have not investigated patients with ALT elevation, high alcohol consumption is unlikely to be a major confounder, as we excluded those with moderate or excess alcohol intake. The prevalence of chronic viral hepatitis in our population is likely to be very low (<1 2%), and hence is

ALT elevation in diabetes 875 unlikely to explain the high prevalence of abnormal ALT. Therefore, based on previous reports including that from our centre, we have assumed that the majority of these patients have NAFLD. 9 11 Our finding that patients with type 2 diabetes had a higher risk of elevated ALT with increasing body mass index (p trend ¼ 0.04) is consistent with the well-established association of NAFLD with metabolic syndrome, and is supportive of our underlying assumption. The spectrum of NAFLD, including its severe form NASH, has been consistently associated with insulin resistance and hyperinsulinaemia. 16 18 This has raised a concern that treatment with insulin in type 2 diabetes may aggravate liver disease. Consistent with this, in data from de Marco et al., 6 the standardized mortality ratio from cirrhosis in patients with type 2 diabetes was higher in those treated with insulin (OR 6.8 vs. diet alone) than in those on oral hypoglycaemic agents (OR 4.9 vs. diet alone). 6 However, both hyperglycaemia and hyperinsulinaemia can promote fatty infiltration of the liver. Long-standing type 2 diabetes may be associated with relative insulin deficiency, and appropriate insulin treatment would shift the balance against hepatic steatosis. Therefore, our observation that current insulin use in patients with type 2 diabetes is associated with a reduced risk of elevated ALT is reassuring, but requires confirmation in larger groups. Previous reports of raised liver enzymes in patients with type 1 diabetes have been limited to small case series. 19,20 In reports where patients with liver test abnormalities were investigated in detail, marked accumulation of glycogen and steatohepatitis were demonstrated on liver biopsy. 19,20 We did not identify any statistically significant risk factors in type 1 diabetes, although elevated ALT was slightly more common in men, and appeared to show some association with microalbuminuria and dyslipidaemia. Risk factors for elevated liver enzymes in Type 1 diabetes should be investigated further in future studies. In summary, the prevalence of elevated ALT is 3 4 times higher in patients with either type 1 or type 2 diabetes than in the general population. Further studies investigating the aetiology and mechanisms of this elevation may suggest appropriate early interventions. References 1. Marchesini G, Brizi M, Morselli-Labate A, Blanchi G, Bugianesi E, McCullogh A, Forlani G, Melachionda N. Association of non-alcoholic fatty liver disease and insulin resistance. Am J Med 1999; 107:450 5. 2. Matteoni C, Younossi Z, Gramlich T, Bopari N, Liu Y, McCullough N. Nonalcoholic fatty liver disease: a spectrum of clinical and pathological severity. Gastroenterology 1999; 116:1413 19. 3. El-Serag HB, Tran T, Everhart JE. Diabetes increases the risk of chronic liver disease and hepatocellular carcinoma. Gastroenterology 2004; 126:460 8. 4. Davila JA, Morgan RO, Shaib Y, McGlynn KA, El-Serag HB. Diabetes increases the risk of hepatocellular carcinoma in the United States: a population based case control study. Gut 2005; 54:533 9. 5. Shahib YH, El-Serag HB, Davila JA, Morgan R, McGlynn KA. Risk factors of intrahepatic cholangiocarcinoma in the United States: a case-control study. Gastroenterology 2005; 128:620 6. 6. De Marco R, Locatelli F, Zoppini G. Cause-specific mortality in type 2 diabetes. Diabetes Care 1999; 22:756 61. 7. Trombetta M, Spiazzi G, Zoppini G, Muggeo M. Review article: type 2 diabetes and chronic liver disease in the Verona diabetes study. Aliment Pharmacol Ther 2005; 22:24 7. 8. Kim HC, Nam CM, Jee SH, Han KH, Oh DK, Suh II. Normal serum aminotransferase concentration and risk of mortality from liver diseases: prospective cohort study. Br Med J 2004; 328:983 7. 9. Hultcrantz R, Glaumann H, Lindberg G, Nilsson LH. Liver investigation in 149 asymptomatic patients with moderately elevated activities of serum aminotransferases. Scand J Gastroenterol 1986; 21:109 13. 10. Skelly MM, James PD, Ryder SD. Findings on liver biopsy to investigate abnormal liver function tests in the absence of diagnostic serology. J Hepatol 2001; 35:195 9. 11. Daniel S, Ben-Menachem T, Vasudevan G, Ma CK, Blumenkehl M. Prospective evaluation of unexplained chronic liver transaminase abnormalities in asymptomatic and symptomatic patients. Am J Gastroenterol 1999; 94:3010 14. 12. Ruhl CE, Everhart JE. Determinants of the association of overweight with elevated serum aminotransferase activity in the United States. Gastroenterology 2003; 124:71 9. 13. Diehl AM, Clark JM. Defining non-alcoholic fatty liver disease: implications for epidemiological studies. Gastroenterology 2003; 124:248 50. 14. NICE 2002/055. Type 2 diabetes management of blood pressure and blood lipids, 2002. [http://www.nice.org.uk] 15. Lebovitz HE, Kreider M, Freed MI. Evaluation of liver function in type 2 diabetic patients during clinical trials. Diabetes Care 2002; 25:815 21. 16. Sanyal AJ, Campbell-Sargent C, Mirshahi F, Rizzo WB, Contos MJ, Sterling RK, Luketic VA, Shiffman ML, Clore JN. Nonalcoholic steatohepatitis: association of insulin resistance and mitochondrial abnormalities. Gastroenterology 2001; 120:1183 92. 17. Pagano G, Pacini G, Musso G, Gambino R, Mecca F, Depetris N, Cassader M, David E, Cavallo-Perin P, Rizzetto M. Nonalcoholic steatohepatitis, insulin resistance, and metabolic syndrome: further evidence for an etiologic association. Hepatology 2002; 35:367 72. 18. Chitturi S, Abeygunasekera S, Farrell GC, Holmes-Walker J, Hui JM, Fung C, Karim R, Lin R, Samarasinghe D, Liddle C,

876 J. West et al. Weltman M, George J. NASH and insulin resistance: insulin hypersecretion and specific association with the insulin resistance syndrome. Hepatology 2002; 35:373 9. 19. Olsson R, Wesslan C, William-Olsson T, Zettergren L. Elevated aminotransferases and alkaline phosphatases in unstable diabetes mellitus without ketoacidosis or hypoglycaemia. J Clin Gastroenterol 1989; 11:541 5. 20. Munns CFJ, McCrossin RB, Thomsett MJ, Batch J. Hepatic glycogenosis: reversible hepatomegaly in type 1 diabetes. J Paediatr Child Health 2000; 36:449 52.