Objective: Several reports have documented the involvement of hyerleptinaemia in malnutrition associated with liver cirrhosis. However, the mechanisms of elevated leptin levels remains unclear. Serum concentrations of tumour necrosis factor-alpha (TNF-alpha), and two soluble TNF receptors (sTNF-RI and sTNF-RII) are increased in patients with liver cirrhosis. In rodents, administration of TNF-alpha has been shown to stimulate plasma leptin concentration, suggesting that a cytokine-leptin link may mediate anorexia and weight loss during chronic inflammation. In this study, we investigate the potential interaction of the TNF-alpha system with leptin in the development of malnutrition in liver cirrhosis.
Study design: A total of 26 male patients with liver cirrhosis and 25 healthy people were recruited at an outpatient clinic at the Veterans General Hospital in Taiwan. Serum biochemistry and anthropometric measurement by bioelectrical impedance analysis were used to assess nutrition status, and immunoassay was used to determine serum leptin, TNF-alpha sTNF-RI and sTNF-RII concentrations.
Results: In cirrhotic patients, the body fat mass (FM) and serum albumin levels were both lower than control subjects [15.8 (13.2-19.5) kg vs. 18.9 (16.2-20.1) kg; 35 (33-41) g/l vs. 43 (41-45) g/l, respectively; P < 0.05]. Serum TNF-alpha sTNF-RI and sTNF-RII were significantly elevated in cirrhotic patients compared to healthy controls [9.8 (7.2-13.5) ng/l vs. 4.3 (3.4-7.3) ng/l; 1682.1 (1344.8-2179.4) ng/l vs. 1319.6 (1037.7-1632.1) ng/l; 4462.2 (3748.5-5159.4) ng/l vs. 3559.8 (2506.9-3988.9 ng/l, respectively; P < 0.01] and correlated with disease severity (graded by Pugh-Child's scores). An inverse correlation was observed between circulating sTNF-RI and sTNF-RII to serum albumin levels (r =-0.42, r = -0.398; P < 0.05). The serum leptin levels in cirrhotic patients were significantly higher [6.0 (3.6-7.7) (g/l vs. 3.4 (2.9-4.3) (g/l; P < 0.01) and correlated with body FM (r = 0.52; P < 0.01]. Using a multiple linear regression analysis with leptin as dependent variable and FM and TNF-alpha, sTNF-R as independent variables, FM and serum sTNF-RI concentrations were found to predict independently the leptin levels in cirrhotic patients.
Conclusion: Our study demonstrated that serum levels of TNF-alpha, sTNF-RI, sTNF-RII and leptin were all elevated in cirrhotic patients. The severity of liver cirrhosis was an important factor for the activation of TNF-alpha system. The activated TNF-alpha system conjointly with hyperleptinaemia might mediate malnutrition in patients with liver cirrhosis.