Introduction: Oxygen delivery to tissue is of major clinical interest in patients with cyanotic congenital heart disease (CHD). The use of pulse oximeter to monitor arterial oxygen saturation (SaO2) is considered accurate and reliable in the range of 90% to 100%. However with desaturation, the accuracy remains controversial below 90%. The aim of this study was to evaluate the accuracy of pulse oximetry in severe hypoxemia. METHOD: In 110 children with cyanotic CHD, pulse oximeter (N-200, Nellcor, USA) readings were compared with the direct measurement of SaO2 by blood gas analyser (Profile10, Stat, USA). All measurements were carried out after induction of anesthesia and devided into 4 groups according to saturation measured by pulse oximeter (SpO2). SpO2 in group I was higher than 90% (n=90), in group II between 80% and 89% (n=75), in group III between 70% and 79% (n=41), in group IV lower than 69% (n=18). Statistical analysis of paired SpO2 and SaO2 values was performed using correlation analysis and paired t-test. The other comparisons were perfomed with ANOVA. p<0.05 was considered statistically significant. RESULTS: Correlation coefficient of group I was 0.89 (p<0.01), group II was 0.67 (p<0.01), group III was 0.63 (p<0.01) and group IV was 0.41. The study demonstrate that SpO2 seems to have good correlation with SaO2 when SpO2 is higher than 70%. This results are contrary to other studies which show that SpO2 is not reliable when SpO2 is below 90%. However, the correlation value r seems to decrease with desaturation. CONCLUSION: The use of pulse oximeter in severe hypoxemic children with CHD is efficient in monitoring oxygenation, even though there is decrease in accuracy of the SpO2.