In a cross-sectional study of more than 30% of French dialysis patients (N = 7,123), we evaluated the relationships between predialysis plasma bicarbonate concentration and nutritional markers. Data including age, gender, cause of end-stage renal disease (ESRD), time on dialysis, body mass index (BMI), blood levels of midweek predialysis albumin, prealbumin, and bicarbonate were collected. Normalized protein catabolic rate (nPCR), dialysis adequacy parameters, and estimation of lean body mass (LBM) were computed from pre- and postbicarbonate-dialysis urea and creatinine levels according to the classical formulas of Garred. Average values (+/- 1 SD) were age 61 +/- 16 years, BMI 23.3 +/- 4.6 kg/m2, dialysis time 12.4 +/- 2.7 h/week, HCO3 22.8 +/- 3.5 mmol/L, albumin 38.7 +/- 5.3 g/L, prealbumin 340 +/- 90 mg/L, Kt/V 1.36 +/- 0.36, nPCR 1.13 +/- 0.32 g/kg BW/day, and LBM 0.86 +/- 0.21% of ideal LBM. A highly significant negative correlation was observed between predialysis bicarbonate levels (within a range of 16-30 mmol/L, 95% of this population) and nPCR confirmed by analysis of variance using bicarbonate classes (p < 0.0001). Bicarbonate was also negatively correlated with albumin, prealbumin, BMI, and LBM. No relationship was noted between bicarbonate and Kt/V despite a positive correlation between Kt/V and nPCR. It is likely that a persistent acidosis observed despite standard bicarbonate dialysis was caused by a high dietary protein intake which results in an increased acid load, but also overcomes the usual catabolic effects of acidosis.