Angeliki M Angelidi, Athanasia Papazafiropoulou, Eleni-Margarita Tzouganatou, Konstantina Anagnostopoulou, Vasilis Velissaris, Vangelis Markakis, Ekaterini Vagena, Αndreas Melidonis
Abstract
Introduction-Aim: Non-alcoholic fatty liver disease (NAFLD) is the commonest cause of abnormal liver function tests and liver disease in the western countries. NAFLD is usually asymptomatic. As a result some non-invasive diagnostic models have been proposed for the diagnosis and staging of NAFLD. The objective of this study was to evaluate whether some of the most common and easily assessed models can also be used to screen for the presence of NAFLD in patients with type 2 diabetes (T2D) in clinical practice.
Patients and Methods: The study population included 110 patients with T2D (28 men) [mean age (±SD) 60.1±9.5 years, HbA1c 6.4±1.0%, body-mass index 28.6±4.8 Kg/m2, duration of diabetes 8.5±4.0 years] attending the outpatient diabetic clinic of our hospital. Anthropometric, clinical, and laboratory data were analyzed during regular health checkups. NAFLD was diagnosed using ultrasound. NAFLD liver fat score, HAIR (Hypertension, ALT, Insulin Resistance), BARD, APRI (AST to Platelet Ratio Index), FIB-4 and LAP (Lipid Accumulation Product) scores were estimated. Discrimination capability was assessed based on the area under the receiver operating characteristic curve (AUC), sensitivity and specificity, positive (PPV) and negative (NPV) predictive values were calculated.
Results: NAFLD, using ultrasound, was diagnosed in 77 patients (70%). Receiver operating characteristic analysis showed that for the NAFLD liver fat score a cut off of ≥-1.44 had a sensitivity of 93%, a specificity of 72%, with an AUC of 0.95 and a PPV of 89% and a NPV of 82%. For the HAIR score a cut off of ≥0.50 had a sensitivity of 62%, a specificity of 44%, with an AUC of 0.58 and a PPV of 72% and a NPV of 33%. For the BARD score a cut off of ≥2.50 had a sensitivity of 51%, a specificity of 64%, with an AUC of 0.59 and a PPV of 77% and a NPV of 36%. For the APRI a cut off of ≥0.23 had a sensitivity of 49%, a specificity of 64%, with an AUC of 0.55 and a PPV of 76% and a NPV of 35%. For the FIB-4 score a cut off of ≥1.00 had a sensitivity of 53%, a specificity of 54%, with an AUC of 0.52 and a PPV of 73% and a NPV of 33%. For the LAP score a cut off of ≥30.93 had a sensitivity of 94%, a specificity of 82%, with an AUC of 0.89 and a PPV of 92% and a NPV of 85%.
Conclusions: The results of the present study showed that NAFLD liver fat and LAP scores showed good sensitivity and specificity for the presence of NAFLD in patients with T2D. Both scores are simple, accurate and non-invasive tools to predict NAFLD. In contrast, HAIR, BARD, APRI and FIB-4 scores showed poor sensitivity and specificity.
Keywords: diabetes mellitus, non-alcoholic fatty liver disease, predicting scores