Factors influencing arterial oxygenation early after bidirectional cavopulmonary shunt without additional sources of pulmonary blood flow

Objectives: Although the arterial oxygen saturation after bidirectional cav opulmonary shunting should theoretically be homogeneous if additional pulmo nary flow is obliterated, the arterial oxygen saturation has been found to vary in clinical practice. Knowledge of the preoperative and operative dete rminants of arterial oxygen saturation early after bidirectional cavopulmon ary shunting may lead to a better understanding of this unique physiology. Methods: Thirty-five patients who underwent bidirectional cavopulmonary shu nting with obliteration of additional pulmonary flow were included in this study. The arterial oxygen saturation was determined at the 5 time points o ver a 48-hour period. Multivariable regression analysis was used to identif y the independent predictors of the arterial oxygen saturation. Results: No significant interval changes occurred in the arterial oxygen sa turation during the 48 hours after bidirectional cavopulmonary shunting, wh ich ranged from 61.6% to 95.6%. There was a significant inverse correlation between the postoperative superior vena cava pressure and the arterial oxy gen saturation (P = .003). A low arterial oxygen saturation early after bid irectional cavopulmonary shunting was a predictor of mortality or exclusion from univentricular repair within 24 months (P = .012, odds ratio = 1.14). Of 11 factors identified by univariable analysis, multiple regression anal ysis indicated that age less than 8 months at the time of shunting (P < .00 01) and ventricular volume overload (P = .002) predicted a lower arterial o xygen saturation after bidirectional cavopulmonary shunting. Conclusions: Even without additional sources of pulmonary blood flow, sever al preoperative factors, including younger age and severe ventricular volum e overload, predicted a decrease in the arterial oxygen saturation early af ter bidirectional cavopulmonary shunting. This, in rum, predicted poor outc ome during 2 years of follow-up.