SUBSTANTIAL CHANGES IN ARTERIAL BLOOD-GASES DURING THORACOSCOPIC SURGERY CAN BE MISSED BY CONVENTIONAL INTERMITTENT LABORATORY BLOOD-GAS ANALYSES

Substantial and clinically relevant changes in arterial blood gases ar e likely to occur during thoracoscopic surgery with one-lung ventilati on (OLV). We hypothesized that they may be missed when using the conve ntional intermittent blood gas sampling practice. Therefore, during 30 thoracoscopic procedures with OLV, the sampling intervals between con secutive intermittent laboratory blood gas analyses (BGA) were evaluat ed with respect to changes of Pao(2) Paco(2) and pHa ([H+]) using a co ntinuous intraarterial blood gas monitoring system. Frequency and timi ng of EGA were based on the clinical judgment of 16 experienced anesth esiologists who were blinded to the continuously measured values. Extr eme fluctuations of Pao(2) (37-625 mm Hg), Paco(2) (27-56 mm Hg), and pHa (7.24-7.51) were observed by continuous blood gas monitoring. Duri ng 63% of all sampling intervals, Pao(2) decreased >20% compared with the preceding EGA value, which remained undetected by intermittent ana lysis. In 10 patients with a continuously measured minimal Pao(2) valu e less than or equal to 60 mm Hg, the preceding EGA overestimated this minimal Pao(2) by >47%. Correspondingly, Pace, increases of >10% were observed in 35% of all sampling intervals, and [H+] increases of >10% were observed in 24% of all sampling intervals. Because these blood g as changes were not reliably detected by using noninvasive monitoring and their magnitude is not predictable during OLV, intermittent EGA wi th short sampling intervals is warranted. In critical cases, continuou s blood gas monitoring may be helpful. Implications: The magnitude of blood gas changes during thoracoscopic surgery with one-lung ventilati on is not predictable and not reliably detected by noninvasive monitor ing. Using a continuous intraarterial blood gas monitoring device, we demonstrated that intermittent laboratory blood gas analysis with shor t sampling intervals is warranted to detect arterial hypoxemia.