Alterations in hemoglobin oxygen affinity can be detected by exposing blood to different PO(2) and recording oxygen saturation, a method termed tonometry. It is the gold standard to measure the PO(2) associated with 50 % oxygen saturation, the index used to quantify oxygen affinity (P(50)Tono). P(50)Tono is used in the evaluation of patients with erythrocytosis suspected to have hemoglobin with abnormal oxygen affinity. Since tonometry is labor intensive and not generally available, we investigated whether accurate estimates of P(50) could also be obtained by venous blood gas analysis, co-oximetry, and standard equations (P(50)Ven). In 50 patients referred for evaluation of erythrocytosis, pH, PO(2), and oxygen saturation were measured in venous blood to estimate P(50)Ven; P(50)Tono was measured for comparison. Agreement among P(50)Ven and P(50)Tono was evaluated (Bland-Altman analysis). Mean P(50)Tono was 25.8 (range 17.4-34.1) mmHg. The mean difference (bias) of P(50)Tono-P(50)Ven was 0.5 mmHg; limits of agreement (95 % confidence limits) were -5.2 to +6.1 mmHg. The sensitivity and specificity of P(50)Ven to identify the 25 patients with P(50)Tono outside the normal range of 22.9-26.8 mmHg were 5 and 77 %, respectively. We conclude that estimates of P(50) based on venous blood gas analysis and standard equations have a low bias compared to tonometry. However, the precision of P(50)Ven is not sufficiently high to replace P(50)Tono in the evaluation of individual patients with suspected disturbances of hemoglobin oxygen affinity.