Pre-analytic errors in the determination of arterial O-2-partial pressure under hyperoxic conditions

Objective: A variety of influences reduce the validity of the measured oxyg en partial pressure (paO(2)). Most errors occur when obtaining the blood sa mple and preparing it for analysis. Unfortunately, there is great controver sy concerning the relevance and extent of these pre-analytic errors. Apart from this, the exact estimation of influencing factors under hyperoxic cond itions has been neglected. Therefore,the objective of this study was to ass ess pre-analytic measuring errors for paO(2) under the condition of hyperox ia as completely as possible and to work out solutions to eliminate these e rrors. Methods: paO(2) analysis was performed on more than 2000 blood samples. Err ors analyzed were the technique of sample taking (direct puncture or from a n indwelling catheter), aspirated air bubbles (0.05-0.35 ml), time and temp erature of sample storage, and the material, size and manufacturer of the a nalyzing syringe. Results: The paO(2) was on average 41 mmHg lower in samples taken from the indwelling catheter than by direct puncture. An air bubble size of 0.1-0.25 ml caused a decrease of 17-41 mmHg. Storage time of 2 min accounted for an paO(2) reduction of 6-67 mmHg depending on the type of syringe used. Glass syringes turned out to be more accurate than plastic syringes. The best re sults were obtained not from commercial "blood gas syringes" but from a sim ple plastic injection syringe. For all pre-analytic errors correction facto rs were established. Conclusion: All pre-analytic errors investigated caused a significant paO(2 ) decrease. Even an ideal procedure (almost no air bubble, short storage on ice) contributes a significant error. Only the appropriate correction fact ors as calculated from this study for routine use lead to the correct resul ts. If they are not taken into account the paO(2) values will be falsely lo w, potentially reading to misinterpretation and misjudgement of a patient's condition.