Aliment Pharmacol Ther 2011; 34: 757–766
Background Hyperuricemia has been associated with metabolic disorders. In this line recent studies observed an independent link between higher uric acid serum levels and clinical diagnosis of non-alcoholic fatty liver disease (NAFLD).
Aims We aimed to assess the potential association between uric acid serum levels and histological liver damage in a homogeneous cohort of biopsy-proven NAFLD patients.
Methods Consecutive NAFLD patients (n = 166), assessed by liver biopsy (Kleiner score), anthropometric, biochemical and metabolic features, were included. Enzymatic colorimetric test was used for serum uric acid assays (Roche Diagnostics GmbH, Mannheim, Germany). Hyperuricemia was diagnosed when uric acid serum levels were >7 mg/dL in men, and >6 mg/dL in women.
Results Mean uric acid serum level was 5.75 mg/dL, and about 20% of patients had hyperuricemia, that was independently associated with younger age (OR 0.951, 95% CI 0.918–0.984, P = 0.004), lobular inflammation (OR 2.144, 95% CI 1.055–4.357, P = 0.03) and steatosis grade (OR 1.859, 95% CI 1.078–3.205, P = 0.02), by multivariate logistic regression analysis. Female gender (OR 2.656, 95% CI 1.190–5.928, P = 0.01), higher HOMA index (OR 1.219, 95% CI 1.043–1.426, P = 0.01), and hyperuricemia (OR 4.906, 95% CI 1.683–14.296, P = 0.004) were linked to NAFLD activity score (NAS) ≥ 5 by multiple logistic regression analysis. Conversely, higher HOMA index (OR 1.140, 95% CI 1.001–1.229, P = 0.04), and NAS (OR1.954, 95% CI 1.442–2.649, P < 0.001) were independently associated with significant fibrosis by logistic regression analysis.
Conclusions In NAFLD patients, hyperuricemia is independently associated with the severity of liver damage, representing, in this setting of patients, together with insulin resistance, a potential new therapeutic target in future intervention trials.
Non-alcoholic Fatty liver disease (NAFLD) is a leading cause of chronic liver disease worldwide.1 Its clinical relevance arises from the fact that a considerable proportion of subjects (20–30%) develop a condition namely non-alcoholic steatohepatitis (NASH) that is a potentially progressive hepatic disorder leading to end-stage liver disease and hepatocellular carcinoma.1, 2 In addition NAFLD is considered the hepatic manifestation of insulin resistance (IR), and is therefore strongly associated with metabolic syndrome, obesity, type 2 diabetes, dyslipidemia and hypertension,3 also representing, together with the above cited conditions, an independent cardiovascular risk factor.4
Uric acid (UA), the final oxidation product of purine metabolism involved in gouty arthritis and kidney stones genesis, has been found to be also associated with different cardiometabolic diseases, like hypertension, kidney disease, metabolic syndrome and cardiovascular disease.5, 6 In these settings, available data prompted to speculate that hyperuricemia is not only an epiphenomenon of metabolic alterations, but also a factor directly involved in the pathogenesis of the above cited disorders.
In this line, considering the link between cardiometabolic alterations and NAFLD, some studies investigated the potential relation between NAFLD and UA levels.7–15 In particular some studies showed that high UA serum levels were associated with the ultrasonographic diagnosis of NAFLD,7–11 independently of metabolic risk factors, and with ALT levels also after correction for HBsAg and anti-HCV positivity.12 In addition, prospective studies observed that baseline high UA serum levels were an independent predictor over time of the development of ultrasonographic liver steatosis.13–15 Finally a recent article also identified an independent association between higher UA serum levels and development over time of cirrhosis or death for cirrhosis in a large USA database of subjects observed between 1971 and 1975.12
All above presented data therefore suggest an association between NAFLD and UA. However, these works have been performed in patients with clinical diagnosis of NAFLD, while, in this setting, no data exist about the potential association between UA serum levels and histological severity of the liver disease.
With this in mind, in a homogeneous cohort of biopsy-proven NAFLD patients, we aimed to assess the potential association between UA serum levels and histological liver damage.
Patients and methods
One-hundred and sixty-six consecutive patients with NAFLD, recruited at the Gastrointestinal & Liver Unit at the University Hospital in Palermo and fulfilling all inclusion and exclusion criteria detailed below were assessed. Patients were included if they had a histological diagnosis of NAFLD on a liver biopsy performed less than 6 months before enrolment. The diagnosis of NAFLD was based on chronically elevated ALT for at least 6 months, alcohol consumption of <20 g/day in the year before liver biopsy and evaluated by a questionnaire, and steatosis (>5% of hepatocytes) at histology with/without necroinflammation and/or fibrosis. Exclusion criteria were: (i) advanced cirrhosis; (ii) hepatocellular carcinoma; (iii) other causes of liver disease or mixed aetiologies (excessive alcohol consumption, hepatitis C, hepatitis B, autoimmune liver disease, Wilson’s disease, hemochromatosis, α1-antitrypsin deficiency); (iv) human immunodeficiency virus infection; (v) previous treatment with antiviral therapy, immunosuppressive drugs and/or regular use of steatosis-inducing drugs, evaluated by interview; (vi) therapy with medications known to affect UA metabolism or (vii) active intravenous drug addiction.
The study was performed in accordance with the principles of the Helsinki Declaration and its appendices, and with local and national laws. Approval was obtained from the hospital’s Internal Review Board and its Ethics Committee, and written informed consent was obtained from all patients and controls.
Clinical and Laboratory Assessment
Clinical and anthropometric data were collected at the time of liver biopsy. Body mass index (BMI) was calculated on the basis of weight in kilograms and height (in metres), and subjects were classified as normal weight (BMI, 18.5–24.9 kg/m2), overweight (BMI, 25–29.9), obese (BMI ≥ 30). Waist circumference was measured at the midpoint between the lower border of the rib cage and the iliac crest. The diagnosis of arterial hypertension was based on the following criteria: systolic blood pressure ≥ 135 mmHg and/or diastolic blood pressure ≥ 85 mmHg (measured three times within 30 min, in the sitting position and using a brachial sphygmomanometer), or use of blood-pressure-lowering agents. The diagnosis of type 2 diabetes was based on the revised criteria of the American Diabetes Association, using a value of fasting blood glucose ≥ 126 mg/dL on at least two occasions.16 In patients with a previous diagnosis of type 2 diabetes, current therapy with insulin or oral hypoglycaemic agents was documented. Metabolic syndrome was diagnosed according to ATPIII criteria.17
A 12-h overnight fasting blood sample was drawn at the time of biopsy to determine serum levels of ALT, GGT, total cholesterol, HDL and LDL-cholesterol, triglycerides, plasma glucose concentration, insulin and platelet count. Insulin resistance (IR) was determined with the homeostasis model assessment (HOMA) method, using the following equation18: Insulin resistance (HOMA-IR) = Fasting insulin (μU/mL) × Fasting glucose (mmol/L)/22.5. HOMA-IR has been validated in comparison with the euglycemic/hyperinsulinemic clamp technique in both diabetic and nondiabetic subjects.19
Commercially available enzymatic colorimetric test was used for serum UA assays (Roche Diagnostics GmbH, Mannheim, Germany). Hyperuricemia was diagnosed when UA serum levels were >7 mg/dL in men, and >6 mg/dL in women.20
Glomerular filtration rate (GFR) was estimated from the four-variable modification of diet in renal disease (MDRD) study equation as follows: estimated GFR (eGFR) = 175 × (serum creatinine∧1.154) × (age∧0.203) × 1.212 (if black) × 0.742 (if female).21
Serum TNFα (GE Healthcare Amersham TNF-α, Little Chalfont, Buckinghamshire, UK: Human, Biotrak Easy ELISA) and adiponectin (SPIbio – Bertin Pharma Human Adiponectin EIA Kit, Bioquote Limited, York, UK) levels were measured in duplicate in a subgroup of patients where serum samples was available.
Assessment of Histology
Slides were coded and read by one pathologist (D.C.), who was unaware of the patient’s identity and history. A minimum length of 15 mm of biopsy specimen or the presence of at least 10 complete portal tracts was required.22 Steatosis was assessed as the percentage of hepatocytes containing fat droplets (minimum 5%), and evaluated as a continuous variable. The Kleiner classification 23 was used to compute the NAS (from 0 to 8, on a scale including separate scores for steatosis, lobular inflammation, and hepatocellular ballooning) and to stage fibrosis from 0 to 4.
Continuous variables were summarised as mean ± standard deviation, and categorical variables as frequency and percentage. The t-test and chi-squared test were used as appropriate. Multiple logistic regression models were used to assess the factors independently associated with hyperuricemia, moderate-severe steatosis, moderate-severe lobular inflammation, NAS ≥ 5 and significant fibrosis. In the first model, the dependent variable was hyperuricemia coded as 0 = absent (UA ≤ 6 mg/dL in women, and ≤ 7 mg/dL in men); or 1 = present (UA > 6 mg/dL in women, and > 7 mg/dL in men). In the second model, the dependent variable was steatosis coded as 0 = mild (steatosis grade 1); or 1 = moderate-severe (steatosis grade 2–3). In the third model, the dependent variable was lobular inflammation coded as 0 = absent-mild (lobular inflammation 0–1); or 1 = moderate-severe (lobular inflammation 2–3). In the fourth model the dependent variable was NAS coded as 0 = NAS < 5; or 1 = NAS ≥ 5. In the fifth model, the dependent variable was fibrosis coded as 0 = no significant fibrosis (F0–F1); or 1 = significant fibrosis (F2–F4).
As candidate risk factors, we selected age, gender, BMI, waist circumference, baseline ALT, GGT, platelet count levels, triglycerides, total and HDL cholesterol, blood glucose, insulin, HOMA-score, diabetes, arterial hypertension, metabolic syndrome, steatosis, UA, hyperuricemia, eGFR (in hyperuricemia model only), lobular inflammation, hepatocellular ballooning, NAS and fibrosis.
To avoid the effect of colinearity, HOMA score, blood glucose levels and insulin levels, as well as UA and hyperuricemia, as well as waist circumference and BMI, or the NAS and its components, were not included in the same multivariate model. Regression analyses were performed using PROC LOGISTIC, PROC REG and subroutines in SAS.24
Patient Features and Histology
The baseline features of the 166 patients are shown in Table 1. The majority of our patients were in the overweight to obesity range, and nearly a quarter was hypertensive. Diabetes was present in about 20% of patients, and metabolic syndrome was diagnosed in 32% of patients.
|Variable||Non-alcoholic fatty liver disease (n = 166)|
|Mean age (years)||44.9 ± 13.4|
|Male gender||112 (67.5)|
|Mean body mass index (kg/m2)||29.8 ± 4.6|
|Body mass index (kg/m2)|
|Waist circumference (cm)||100.5 ± 12.0|
|Alanine aminotransferase (IU)||79.3 ± 58.8|
|Platelet count (103 mm3)||221.9 ± 61.8|
|γ Glutamyl transferase (IU)||106.5 ± 140.1|
|Cholesterol (mg/dL)||203.3 ± 47.4|
|HDL cholesterol (mg/dL)||48.1 ± 16.1|
|Triglycerides (mg/dL)||148.6 ± 76.2|
|Blood glucose (mg/dL)||96.8 ± 29.3|
|Insulin (μU/mL)||17.2 ± 10.5|
|HOMA-score||4.26 ± 3.42|
|Type 2 diabetes||33 (19.9)|
|Arterial hypertension||35 (21.1)|
|Metabolic syndrome||53 (31.9)|
|Uric acid (mg/dL)||5.75 ± 1.21|
|Estimated GFR (mL/min per 1.73 m2)||100.5 ± 20.7|
|NAFLD activity score (NAS)|
|Steatosis as continuous variable||45.4 ± 26.4|
|1 (5–33%)||63 (38.0)|
|2 (>33–66%)||48 (28.9)|
|3 (>66%)||55 (33.1)|
|Stage of fibrosis|
Mean eGFR was 100.5 ± 20.7 mL/min per 1.73 m2, and no patients had an eGFR < 60, expression of chronic kidney disease.
At liver biopsy, according to Kleiner et al. (39) 59.7% of NAFLD were classified as NASH, 10.8% as non-NASH and 29.5% were indeterminate. One in three patients had steatosis >66%, and half of the cases had fibrosis ≥2.
In a subgroup of 107 patients comparable to the entire population (75 men; mean age 45 ± 13 years; mean HOMA 4.39 ± 3.65; mean UA 5.74 ± 1.22 mg/dL; 17.7% with hyperuricemia; 61.6% with NAS ≥ 5; and 45.7% with fibrosis ≥ 2), serum levels of TNFα and adiponectin were measured. Mean serum TNFα was 7.78 ± 3.72 pg/mL (range 1.20–16.60), and mean serum adiponectin was 5.55 ± 3.72 μg/mL (range 2.30–13.60).
Factors associated with hyperuricemia
Mean uric acid serum level was 5.75 mg/dL, and about 20% of patients had a condition of hyperuricemia. The univariate and multivariate comparison of variables between patients with and without hyperuricemia are reported in Table 2. Multivariate logistic regression analysis showed that the following features were independently linked to hyperuricemia: younger age (OR 0.951, 95% CI 0.918–0.984, P = 0.004), lobular inflammation (OR 2.144, 95% CI 1.055–4.357, P = 0.03) and steatosis grade (OR 1.859, 95% CI 1.078–3.205, P = 0.02). When replacing in the model lobular inflammation, steatosis grade and hepatocellular ballooning with NAS, the latter remained significantly associated with hyperuricemia (OR 1.876, 95% CI 1.340–2.628, P < 0.001).
|Variable||No hyperuricemia N = 133||Hyperuricemia n = 33||Univariate analysis P value||Multivariate analysis|
|OR (95% CI)||P value|
|Mean age (years)||46.4 ± 12.4||38.7 ± 15.5||0.003||0.951 (0.918–0.984)||0.004|
|Mean body mass index (kg/m2)||29.8 ± 4.6||29.8 ± 4.7||0.97||–|
|Waist circumference (cm)||99.7 ± 12.4||103.6 ± 10.0||0.13||–|
|Alanine aminotransferase (IU)||74.7 ± 59.0||97.8 ± 54.8||0.04||1.003 (0.996–1.010)||0.37|
|γ Glutamyl transferase (IU)||109.9 ± 150.0||92.7 ± 90.0||0.53||–|
|Platelet count (103 mm3)||218.5 ± 62.0||235.9 ± 60.3||0.15||–|
|Cholesterol (mg/dL)||202.7 ± 46.6||206.1 ± 51.2||0.71||–|
|HDL cholesterol (mg/dL)||47.9 ± 15.2||49.0 ± 19.6||0.73||–|
|Triglycerides (mg/dL)||147.2 ± 78.8||154.5 ± 65.1||0.62||–|
|Blood glucose (mg/dL)||99.1 ± 31.4||91.8 ± 16.1||0.19||–|
|Insulin (μU/mL)||16.9 ± 11.0||18.7 ± 8.2||0.38||–|
|HOMA-score||4.30 ± 3.66||4.13 ± 2.15||0.80||–|
|Estimated GFR (mL/min per 1.73 m2)||100.0 ± 20.9||102.8 ± 20.0||0.48||–|
|Lobular inflammation 0/1/2/3||12/72/44/5||0/13/20/0||0.02||2.144 (1.055–4.357)||0.03|
|Steatosis (continuous variable)||42.6 ± 26.4||56.6 ± 23.6||0.006||–|
|Steatosis grade 1/2/3||59/35/39||4/13/16||0.002||1.859 (1.078–3.205)||0.02|
|Hepatocellular ballooning 0/1/2||19/49/65||1/11/21||0.08||1.554 (0.790–3.060)||0.20|
|Stage of fibrosis 0/1/2/3/4||37/37/31/20/8||6/8/11/7/1||0.32||–|
Considering UA as a continuous variable, younger age (P = 0.006), male gender (P = 0.001) and NAS (P = 0.005) were independent factors in multiple linear regression analysis.
In the subgroup of 107 patients assessed for TNFα and adiponectin serum levels, and comparable to the entire population (men 70%, mean age 44.8 years, mean BMI 30 Kg/m2, hyperuricemia 17.7%, NAS ≥ 5 61.6%, significant fibrosis 45.8%), we observed a significant linear relation between higher UA and lower adiponectin serum levels (P < 0.001), while no association was found between UA and TNFα serum levels (P = 0.69). Accordingly, Figure 1 shows the distribution of adiponectin according to UA quartiles.
Factors Associated with Histological Features
High BMI, high waist circumference, high blood glucose and insulin levels, high HOMA, high UA levels, hyperuricemia, lobular inflammation and hepatocellular ballooning were associated with moderate-severe steatosis (grade 2–3) by univariate analysis (P < 0.10), even if only high HOMA (OR 1.182, 95% CI 1.014–1.379, P = 0.03), hyperuricemia (OR 4.600, 95% CI 1.476–14.377, P = 0.009), and lobular inflammation (OR 1.826, 95% CI 1.018–3.276, P = 0.04) remained significantly associated at multivariate logistic regression analysis. When replacing hyperuricemia (categorical variable) with UA (continuous variable) in the multivariate analysis, we obtained similar results (OR 1.509, 95% CI 1.104–2.062, P = 0.01 for UA).
Similarly female gender, older age, high waist circumference, high insulin levels, high HOMA, high UA levels, hyperuricemia, hepatocellular ballooning and steatosis were associated with moderate-severe lobular inflammation (grade 2–3) by univariate analysis (P < 0.10), even if only older age (OR 1.032, 95% CI 1.003–1.061, P = 0.03), hyperuricemia (OR 2.696, 95% CI 1.076–6.754, P = 0.03), and hepatocellular ballooning (OR 1.715, 95% CI 1.026–2.869, P = 0.04) remained significantly associated at multivariate logistic regression analysis. When replacing hyperuricemia with UA in the multivariate analysis, we obtained similar results (OR 1.432, 95% CI 1.044–1.965, P = 0.02 for UA)
The univariate and multivariate comparison of variables between patients with NAS < 5, and those with NAS ≥ 5 are reported in Table 3. Multivariate logistic regression analysis showed that the following features were independently linked to NAS ≥ 5: female gender (OR 2.656, 95% CI 1.190–5.928, P = 0.01), higher HOMA index (OR 1.219, 95% CI 1.043–1.426, P = 0.01) and hyperuricemia (OR 4.906, 95% CI 1.683–14.296, P = 0.004). When replacing hyperuricemia with UA in the multivariate analysis, we obtained similar results (OR 1.558, 95% CI 1.123–2.163, P = 0.008 for UA). Considering the independent link between female gender and NAS ≥ 5, interestingly we observed a higher prevalence of metabolic syndrome in women compared with men (27/54 vs. 26/112, P = 0.01), as well as a trend for higher mean HOMA values (4.94 ± 3.32 vs. 3.93 ± 3.43, P = 0.07).
|Variable||NAS < 5 n = 67||NAS ≥ 5 n = 99||Univariate analysis P value||Multivariate analysis|
|OR (95% CI)||P value|
|Mean age (years)||45.0 ± 12.0||44.8 ± 14.3||0.98||–|
|Mean body mass index (kg/m2)||29.0 ± 4.8||20.3 ± 4.5||0.06||–|
|Waist circumference (cm)||98.0 ± 12.7||102.1 ± 11.4||0.05||1.012 (0.931–1.099)||0.78|
|Alanine aminotransferase (IU)||67.6 ± 62.7||87.2 ± 54.9||0.03||1.006 (0.999–1.013)||0.09|
|γ Glutamyl transferase (IU)||127.3 ± 150.1||92.5 ± 131.8||0.11||–|
|Platelet count (103 mm3||223.6 ± 62.3||220.8 ± 61.8||0.77||–|
|Cholesterol (mg/dL)||211.0 ± 45.3||198.3 ± 48.3||0.10||–|
|HDL cholesterol (mg/dL)||49.2 ± 15.4||47.4 ± 16.6||0.47||–|
|Triglycerides (mg/dL)||143.6 ± 74.2||152.0 ± 77.8||0.49||–|
|Blood glucose (mg/dL)||93.3 ± 20.2||99.2 ± 34.0||0.20||–|
|Insulin (μU/mL)||13.3 ± 9.3||19.9 ± 10.5||<0.001||–|
|HOMA-score||3.20 ± 2.89||5.00 ± 3.57||<0.001||1.219 (1.043–1.426)||0.01|
|Uric acid (mg/dL)||5.52 ± 1.15||5.91 ± 1.22||0.03||–|
The univariate and multivariate comparison of variables between patients with and without significant fibrosis are reported in Table 4. Multivariate logistic regression analysis showed that the following features were independently linked to significant fibrosis: higher HOMA index (OR 1.140, 95% CI 1.001–1.229, P = 0.04), and NAS (OR 1.954, 95% CI 1.442–2.649, P < 0.001).
|Variable||Fibrosis F0–F1 n = 88||Fibrosis F2–F4 n = 78||Univariate analysis P value||Multivariate analysis|
|OR (95% CI)||P value|
|Mean age (years)||42.3 ± 11.7||47.8 ± 14.6||0.007||1.029 (0.999–1.060)||0.06|
|Mean body mass index (kg/m2)||29.1 ± 4.9||30.5 ± 4.3||0.05||–|
|Waist circumference (cm)||97.6 ± 12.8||103.5 ± 10.5||0.005||0.979 (0.896–1.070)||0.64|
|Alanine aminotransferase (IU)||78.3 ± 60.9||80.4 ± 56.6||0.81|
|γ Glutamyl transferase (IU)||98.3 ± 113.8||115.9 ± 165.0||0.42||–|
|Platelet count (103 mm3)||226.1 ± 50.6||217.3 ± 72.3||0.36||–|
|Cholesterol (mg/dL)||202.1 ± 42.2||204.9 ± 52.9||0.70||–|
|HDL cholesterol (mg/dL)||50.0 ± 16.8||46.0 ± 15.1||0.11||–|
|Triglycerides (mg/dL)||133.6 ± 67.1||145.3 ± 82.5||0.31||–|
|Blood glucose (mg/dL)||93.4 ± 24.0||100.6 ± 34.1||0.11||–|
|Insulin (μU/mL)||13.9 ± 7.9||20.9 ± 11.8||<0.001||–|
|HOMA-score||3.32 ± 2.79||5.33 ± 3.75||<0.001||1.140 (1.001–1.299)||0.04|
|Uric acid (mg/dL)||5.74 ± 1.15||5.77 ± 1.28||0.90||–|
|NAS 0–2/3–4/5–8||17/35/66||1/14/63||<0.001||1.954 (1.442–2.649)||<0.001|
|Lobular inflammation 0/1/2/3||12/57/19/0||0/28/45/5||<0.001||–|
|Steatosis (continuous variable)||39.4 ± 27.1||52.1 ± 24.1||0.002||–|
|Steatosis grade 1/2/3||41/25/22||22/23/33||0.006||–|
|Hepatocellular ballooning 0/1/2||16/37/35||4/23/51||<0.001||–|
In this study of 166 compensated patients with histological diagnosis of NAFLD, mostly overweight or obese and at high prevalence of metabolic syndrome, we found that hyperuricemia was associated with histological features of the liver disease, representing an independent risk factor for higher steatosis, lobular inflammation and NAS.
Literature data showed that higher UA serum levels are associated with metabolic syndrome and its components.5, 6 Due to the strong association between these features and NAFLD other works also highlighted a link between UA serum levels and NAFLD.7–15 However, all these studies found a relationship between UA serum levels and clinically diagnosed NAFLD (ultrasound evidence or laboratory surrogates of steatosis), while data on patients with biopsy-proven NAFLD are lacking.
In our study we found that the severity of steatosis was independently associated not only with lobular inflammation, but also with hyperuricemia. This finding is in line with different studies in Asiatic and Italian populations demonstrating an independent link between UA and clinically diagnosed NAFLD,7–11 or hypertransaminasaemia.12
To the best of our knowledge, this is the first study that assessed the potential association between hyperuricemia and severity of liver damage in NAFLD patients. In this line the novel finding of this article lies in the identification of an independent association between the presence of hyperuricemia and the severity of both lobular inflammation and NAS in a cohort of histologically diagnosed NAFLD patients. In particular we found that lobular inflammation was independently associated with both older age and UA levels. Similarly, we found that a NAS ≥ 5, expression of liver inflammatory alterations suggestive of a diagnosis of NASH, was independently associated with hyperuricemia, after correction for and together with IR, a well-known factor linked to the pathogenesis of NASH. Our data therefore are in line with recent studies that identified in higher UA levels not only an independent predictor of NAFLD occurrence among large cohorts of Asiatic subjects,13–15 but also a predictor of cirrhosis development and cirrhosis-related death in a large cohort of American subjects.12 These epidemiological evidences, together with our data on a well characterised biopsy-proven group of NAFLD patients therefore suggest a pathogenic role of UA on the severity of liver damage. In this line recent studies strongly suggested to consider hyperuricemia a direct factor in the pathogenesis of metabolic disorders like hypertension or impaired glucose tolerance, and not an epiphenomenon of the same metabolic alterations.5, 25, 26
Although this study was not designed to clarify the pathogenetic link between UA and severity of liver disease in NAFLD patients, a few hypotheses may be put forward according to the literature. Experimental data showed that UA is able to induce endothelial dysfunction, IR, oxidative stress and systemic inflammation,5, 6 all factors involved in NAFLD pathogenesis.1 In particular a recent mouse model showed that hyperuricemia might be partially responsible for the pro-inflammatory endocrine imbalance in the adipose tissue (increased monocyte chemotactic protein-1 and decreased adiponectin production), which is an underlying mechanism of the low-grade inflammation and IR in subjects with NAFLD.27 In line with these experimental data we found a relation between higher UA serum levels and lower adiponectin serum levels in our NAFLD patients, suggesting, together with the above cited study, a possible interference of UA on adiponectin expression, such as one of the possible mechanisms by which UA could act in the pathogenesis of NAFLD.
Interestingly we found that a high NAS was also independently associated with female gender. We have no explanation for this issue, and further large scale studies should be designed to explore what is the pathogenic stem of this association. However, consistent with our data, we can hypothesise that the higher prevalence of metabolic syndrome and the higher HOMA values found in our population among women compared with men, could partially explain our results.
Finally, we found no association between significant fibrosis and UA, while we confirmed IR and NAS, well-known risk factors for fibrosis,28–30 as independently associated with significant fibrosis. Our data are not in contrast to the finding of Afzali and colleagues,12 that identified higher UA baseline serum levels as a predictor of cirrhosis development in a large cohort of patients without histological characterisation of liver damage. In fact, according to our data (direct relation between NAS and UA) and to those of Afzali et al.,12 it is possible to speculate that UA might participate in the progression of liver disease and finally in the cirrhosis development, by promoting and amplifying liver inflammation.
Our data therefore identified in hyperuricemia, a new potential pathogenetic trigger for NAFLD patients, suggesting that the pharmacological correction of this disorder, together with lifestyle counselling, could impact on disease severity and disease progression.
The main limitation of this study lies in its cross-sectional nature, making it impossible to dissect the temporal relation between hyperuricemia and histological severity of NAFLD. A further methodological question is the potentially limited external validity of the results for different populations and settings. Our study included a cohort of Italian subjects enrolled at a tertiary care centre, who may be different from the majority of prevalent cases of NAFLD in the general population. Lack of data on the prevalence of hyperuricemia in a matched control population, and on serum levels of other adipocytokines, and on liver expression of pro-inflammatory cytokines and adipocytokines, might also affect interpretation of the results.
In conclusion, this study, on a cohort of patients with histological diagnosis of NAFLD, showed an independent link between hyperuricemia and the severity of liver damage. In particular we found that hyperuricemia was independently associated with the severity of steatosis, lobular inflammation and NAS, representing, together with IR, a potential relevant factor in the pathogenesis of NAFLD. These results, needing further validation in large scale studies, could also suggest testing hyperuricemia as a new therapeutic target in future intervention trials on NASH patients.
S. Petta designed the study, contributed to data acquisition, was responsible for writing the manuscript, and participated in statistical analysis. C. Camma’, V. Di Marco, and A. Craxì (Director of the GI & Liver Unit) were responsible for the project and writing of the manuscript. D. Cabibi participated in patient management and data collection. All authors have seen and approved the final version of the manuscript.
Declaration of personal and funding interests: None.