High-Flux Dialysis Lowers Plasma Leptin Concentration in Chronic Dialysis Patients Daniel W. Coyne, MD, Samuel Dagogo-Jack, MD, Samuel Klein, MD, Eddine Merabet, PhD, Jean Audrain, RN, and Michael Landt, PhD ● Leptin is a protein produced by fat cells and involved in body weight regulation. Plasma leptin is signiﬁcantly higher in some hemodialysis (HD) patients than in normal controls. We examined the inﬂuence of dialyzer membrane biocompatibility and ﬂux on elevated plasma leptin concentrations in hemodialysis patients. Employing a crossover design, leptin and tumor necrosis factor  (TNF) levels were serially determined in eight chronic dialysis patients. Patients were dialyzed sequentially on low-ﬂux cellulosic (TAF) dialyzers, low-ﬂux (F8) polysul- fone, high-ﬂux (F80B) polysulfone, then low-ﬂux polysulfone and cellulosic dialyzers again. Mean leptin concentra- tions were similar when low-ﬂux polysulfone or cellulosic dialyzers were employed (141.9  24.2 g/L versus 137.8  18.4 g/L, respectively (P  NS). In contrast, leptin fell signiﬁcantly on the high-ﬂux polysulfone dialyzer (99.4  16.2 g/L) compared with cellulosic (P F 0.005), and low-ﬂux polysulfone dialyzers (P F 0.02). Leptin clearance by the high-ﬂux polysulfone dialyzer was signiﬁcantly higher than the low-ﬂux dialyzers (50.4  21.5 v -9.6  10.3 mL/min; P  0.043), but did not account fully for the 30% decline in plasma leptin during the high-ﬂux arm of the study. Concentrations of TNF were lower when high-ﬂux polysulfone dialyzers were employed, but there was no correlation of individual TNF levels with leptin concentrations. High-ﬂux dialysis lowers plasma leptin concentrations an average of 30%, but biocompatibility does not inﬂuence leptin levels. The decrease in plasma leptin on high-ﬂux dialysis cannot be explained solely by enhanced clearance.  1998 by the National Kidney Foundation, Inc. INDEX WORDS: Hemodialysis; biocompatible membranes; leptin; high-ﬂux dialysis. L EPTIN is a 16-kd protein produced by the obese gene, which is expressed and se- creted by adipocytes. 1,2 Leptin is known to regu- late body weight and adipose tissue mass through a feedback mechanism in rodent models of obe- sity. 2,3 Administration of leptin to animals led to reduced food intake, resulting in weight loss. 4,5 Less is known about the physiological actions of leptin in humans. We have recently reported that plasma leptin concentrations are signiﬁcantly higher in hemo- dialysis (HD) patients compared with normal controls. 6 In lean adults (body mass index [BMI], 18 to 25 kg/m 2 ) leptin concentrations range up to 8 μg/L in men and 16 μg/L in women, but much higher levels occur in obesity. As in normal adults, leptin levels correlated with BMI, which reﬂected the degree of body fat in individuals. Extremely high levels (above 70 μg/L) that were out of proportion to BMI were seen in hemodialy- sis patients, particularly with a BMI above 28 kg/m 2 . There was no correlation of plasma leptin concentrations with indices of residual renal func- tion. One potential cause could be increased production of leptin in response to increased cytokines, such as tumor necrosis factor alpha (TNF), which is known to be increased in hemodialysis patients exposed to cellulosic dialy- sis membranes. 7,8 Injection of TNF increases leptin levels in rodents. 9,10 Alternatively, en- hanced clearance of leptin by high-ﬂux dialyzers might result in lower leptin levels than seen in patients dialyzed on low-ﬂux dialyzers. To deter- mine the role of dialyzer type in the elevated leptin concentrations observed in HD patients, we designed a crossover study, serially measur- ing leptin levels while patients were dialyzed on low-ﬂux cellulose dialyzers, and then low-ﬂux and high-ﬂux polysulfone dialyzers. METHODS Subjects Leptin concentrations were previously measured in 141 patients undergoing chronic hemodialysis in Chromalloy American Kidney Center. Patients who were dialyzed with bioincompatible low-ﬂux dialyzers (Terumo 175 or 220) and who had leptin levels of 70 ng/mL or greater were consid- ered for the study. Nine patients consented to participate in the crossover study. At least two baseline plasma specimens From the Departments of Internal Medicine (Renal and Endocrine Divisions) and Pediatrics, Washington University School of Medicine, St Louis, MO. Received October 13, 1997; accepted in revised form June 19, 1998. Address reprint requests to Daniel W. Coyne, MD, Depart- ment of Internal Medicine, Renal Division, Chromalloy American Kidney Center, Washington University School of Medicine, 660 South Euclid Ave, Box 8129, St Louis, MO 63110-1093. E-mail: dcoyne@imgate.wustl.edu  1998 by the National Kidney Foundation, Inc. 0272-6386/98/3206-0016$3.00/0 American Journal of Kidney Diseases, Vol 32, No 6 (December), 1998: pp 1031-1035 1031