Fatty liver is independently associated with increased cardiovascular risk in the presence of metabolic dysfunction

Background While metabolic dysfunction-associated fatty liver disease (MAFLD) has recently been proposed as positive diagnosis to substitute (non)-alcoholic fatty liver disease applying novel and broad criteria for metabolic dysfunction, it is yet unclear if this definition is more accurate in identifying MAFLD patients at increased cardiovascular risk compared to those with metabolic dysfunction without fatty liver (FL). Methods Asymptomatic adults from 2 prospective cohort studies were included: 1401 subjects from the SAPHIR-study aged 40-70years and 4286 subjects from the SAKKOPI-study aged 45-80 years undergoing screening colonoscopy. FL was diagnosed using FLI≥60 (SAPHIR) and abdominal ultrasound (SAKKOPI). MAFLD was diagnosed according to the recent expert-consensus if FL was accompanied by metabolic dysfunction: diabetes mellitus type 2 (T2DM), BMI≥25kg/m² or BMI<25kg/m² with≥2 metabolic abnormalities. Insulin resistance (IR) was assessed using HOMA-IR≥2.5, cardiovascular risk was assessed using SCORE2. Results 415 patients (30%) and 1903 (44%) were diagnosed with fatty liver in the SAPHIR/SAKKOPI study, of which 413 (99.5%) and 1814 (95.3%) also met the MAFLD definition. At the same time, 665 (47%)/1644 (38%) subjects without FL also fulfilled criteria for metabolic dysfunction. Most prominent difference in patients with metabolic dysfunction with and without FL were IR and (central) obesity resulting in higher subclinical inflammation (ferritin) and cardiovascular risk (SCORE2, [ Table 1 ]). Forward stepwise regression of all available biometric and laboratory data to explain presence of FL confirmed the relevance of waist circumference, IR and triglycerides, among others ([ Table 1 ]). Tab. 1 Patient characteristics SAPHIR SAKKOPI Metabolic dysfunction+ FL (MAFLD) n=413 (29%) Metabolic dysfunction w/o FL n=665 (47%) Metabolic dysfunction+ FL (MAFLD) n=1814 (42%) Metabolic dysfunction w/o FL n=1644 (38%) Age 51.6±5.7 52.4±6.1 60.2±8.3 60.2±8.8 Female sex 80 (19%) 284 (43%) 661 (36%) 867 (53%) MetS 284 (69%) 123 (19%) 1020 (56%) 414 (25%) T2DM 38 (9%) 13 (2%) 282 (15%) 114 (7%) Prediabetes 214 (52%) 287 (43%) 1144 (63%) 852 (52%) HOMA-IR≥2.5 188 (46%) 56 (8%) 893 (49%) 283 (17%) Art. hypertension 338 (82%) 450 (68%) 1471 (81%) 1221 (74%) BMI 30.9±3.9 26.3±2.4 29.7±4.4 26.4±3.3 Obesity 405 (98%) 486 (73%) 758 (42%) 202 (12%) WHR 0.95±0.06 0.89±0.07 0.97±0.08 0.93±0.07 WC≥102/88 329 (80%) 227 (34%) 1357 (75%) 842 (51%) Hypertrigyleridemia 231 (56%) 99 (15%) 680 (37%) 328 (20%) HypoHDL 83 (20%) 52 (8%) 425 (23%) 209 (13%) hsCRP>0.2 230 (55%) 278 (42%) – – CRP>0.5 122 (30%) 132 (20%) 515 (28%) 342 (21%) Ferritin 242 (136-362) 129 (70-218) 157 (93-264) 108 (62-174) SCORE2 6.0 (4.2-8.5) 4.1 (2.9-5.8) 7.9 (5.0-12.0) 6.2 (3.8-10.0) MAFLD-group T2DM 38 (9%) 13 (2%) 282 (16%) 114 (7%) MAFLD-group BMI 368 (89%) 477 (72%) 1365 (75%) 1050 (64%) MAFLD-group Lean 7 (2%) 175 (26%) 167 (9%) 480 (29%) Alcohol–tea totalers 93 (23%) 204 (31%) 588 (35%) 642 (41%) Alcohol –<1 drink/day 226 (55%) 337 (56%) 666 (39%) 625 (40%) Alcohol –<2/3 drinks/day 69 (17%) 111 (17%) 364 (22%) 263 (17%) Alcohol–abusers 25 (6%) 13 (2%) 71 (4%) 20 (1%) MetS–metabolic syndrome; T2DM–type 2 diabetes mellitus; HOMA-IR–homeostasis model assessment of insulin resistance; WC–waist circumference; WHR–waist hip ratio; FLI–fatty liver index SAKKOPI + Dependent variable: FL Relevance for explaining presence of FL Adjusted OR 95%CI P value WC, per cm 1 1.031 1.017-1.046 <0.001 HOMA-IR*, per log 2 2.192 1.806-2.669 <0.001 ALT, per U/L 3 1.026 1.016-1.038 <0.001 Triglycerides*, per log 4 1.491 1.207-1.844 <0.001 Alcohol –<1 drink/day 5 1.117 0.909-1.373 0.293 Alcohol –<2/3 drinks/day 1.330 1.024-1.728 0.032 Alcohol–abusers 3.765 1.945-7.689 <0.001 BMI, per kg/m² 6 1.118 1.076-1.162 <0.001 Ferritin*, per log 7 1.197 1.065-1.346 0.003 OGTT, per mg/dL 8 1.004 1.002-1.007 <0.001 GGT*, per log 9 1.254 1.050-1.498 0.012 Uric acid, per mg/dL 10 1.094 1.017-1.177 0.015 Systolic BP, per mmHg 11 1.006 1.001-1.012 0.019 AST, per U/L 12 0.986 0.972-1.000 0.046 *these parameters were log-transformed for regression analyses; BP–blood pressure; FL–fatty liver; OGTT–oral glucose tolerance test after 2h; OR–odds ratio; WC -waist circumference; SAKKOPI + Dependent variable: SCORE2 Adjusted B 95%CI P value Age, per year 0.454 0.439-0.469 <0.001 However, further adjusting for age, sex, and 3 most relevant metabolic-dysfunction-components, FL was not anymore associated with cardiovascular risk (SAKKOPI: B=0.978, 95%CI: -0.188-0.384, p=0.503), [ Table 2 ]). Tab. 2 Female sex -3.171 -3.44-(-2.902) <0.001 WC, per cm 0.019 0.006-0.032 0.005 HOMA-IR*, per log 1.046 0.736-1.356 <0.001 Triglycerides*, per log 1.813 1.536-2.090 <0.001 FL 0.978 -0.188-0.384 0.503 *these parameters were log-transformed for regression analyses; FL–fatty liver; WC -waist circumference; + Analyses were not performed in SAPHIR study since definition of FL was based on FLI incorporating BMI, WC and triglycerides Conclusions Metabolic dysfunction (central obesity, IR) per se rather than FL (i.e., MAFLD) explains cardiovascular risk in subjects with metabolic dysfunction. Publication History Article published online: 26 August 2022 © 2022. Thieme. All rights reserved. Georg Thieme Verlag Rüdigerstraße 14, 70469 Stuttgart, Germany