In this work we present a graphite furnace atomic absorption spectrometric method for blood lead using palladium as a chemical modifier. Whole blood was diluted 10- fold with a 0.1% v/v triton X-100 solution; 10-mu-L of this solution and 10-mu-L of the palladium-based modifier (2 mg Pd/L, 2% w/v citric acid and 0.01 M nitric acid) were injected onto the L'vov platform by using the alternate volume mode. The following furnace operating parameters were used: (a) drying steps, 120-degrees-C for 10s and 250-degrees-C for 30s; (b) pyrolysis steps, 800-degrees-C for 45s (with oxygen) and 1100-degrees-C for 25s; (c) atomization, 1600-degrees-C for 3s; (d) clean out, 2700-degrees-C for 4s. Accuracy was tested by using (i) a NIST standard (SRM-909) and the Behring Control Blood for Metal 1 (OSSD 21) with lead concentrations of 23.7 +/- 2.1-mu-g/L (found: 21.2 +/- 0.7-mu-g/L) and 413 +/- 51-mu-g/L (found: 407 +/- 6-mu-g/L), respectively; (ii) recovery studies (ca. 100 +/- 1%), and (iii) a reported method (mean relative error: 5.1%). Approximate standard deviations of 0.3 (within-run) and 0.7 (between-runs) mu-g Pb/L were found in the precision study. The detection limit (3-sigma) and the characteristic mass (for a 10- mu-L injection volume) were 0.1-mu-g Pb/L and 15 pg/0.0044 A.s, respectively. The proposed method was used to establish the lead levels of patients with renal insufficiency; a mean concentration (+/- SD) of 59 +/- 39-mu-g Pb/L (range: 12 - 160-mu-g Pb/L) was found. The method was interference-free reliable and reproducible.