Metabolic acidosis with an increased anion gap (AG) is frequently seen among patients with end-stage renal failure that is corrected to a variable degree by chronic hemodialysis. The degree of acidosis is generally interpreted from the concentration of total carbon dioxide (tCO(2)) in blood drawn from the vascular access used for dialysis. As with many dialysis units in the United States, our laboratory studies for outpatients are performed in a central laboratory several hundred miles away and must be shipped there by air freight. We observed a consistent and clinically important difference between the tCO(2) content of samples reported from the central laboratory compared with results reported from a local university hospital chemistry laboratory. The central laboratory readings were always lower, resulting in an increase in the AG. Delays in centrifugation of the blood to separate the clot from the serum and in the initiation of analysis led to an increase in the lactate content of the samples. That increase, however, was insufficient to explain the difference in tCO(2) levels. These data suggest that something happens to the samples in transit to cause an artifactual reduction of the tCO(2) level. For many dialysis patients, the severity of their acidosis may be falsely represented by the tCO(2) content of blood samples reported from central laboratories.