Comparison of lactate values obtained from different sites and their clinical significance in patients with severe sepsis

context and objective: the ideal site for lactate collection has not been clearly established. this study aimed to evaluate associations between lactate levels in arterial blood (lart), peripheral venous blood (lper) and central venous blood (lcen) in patients with severe sepsis or septic shock. design and setting: cross-sectional analytical study in an tertiary university hospital. method: samples from patients with a central venous catheter and from healthy volunteers (control group) were collected. blood was drawn simultaneously for measurements of lart, lper and lcen, and the first sample was collected less than 24 hours after the onset of organ dysfunction. the results were analyzed using pearson correlation, bland-altman and mcnemar tests. results: a total of 238 samples were collected from 32 patients. the correlation results were r = 0.79 (p < 0.0001) for lart/lper and r = 0.84 (p < 0.0001) for lart/lcen. bland-altman showed large limits of agreement: -3.2 ± 4.9 (-12.8 to 6.4) and -0.8 ± 5.9 (-12.5 to 10.8), for lper and lcen respectively. in the control group, there was greater correlation (r = 0.9009, p = 0.0004) and agreement: -0.7 ± 1.2 (-3.1 to 1.7). regarding clinical intervention, there was good agreement between lart/lcen (96.3%; three disagreements), with worst results for lart/lper (87.0%) with 10 cases of disagreement (p = 0.04). in eight patients (80.0%) lper was higher than lart. conclusion: lcen, and not lper, can replace lart with good correlation and clinical agreement. lper tends to overestimate lart, thus leading to unnecessary therapeutic interventions