Arterial blood gases, CBCs, and BMPs from a major hospital center

Background: Spurious hypoxemia—artefactual reduction in measured arterial oxygen pressure (paO₂) from cellular oxygen consumption in blood specimens—has been attributed to leukocytosis (“leukocyte larceny”) or thrombocytosis (“platelet larceny”). However, the relative contribution of elevated cell counts and temperature-dependent metabolism remains unclear. Methods: We retrospectively analyzed 7,431 arterial blood gas (ABG) specimens from 2,643 patients at UC Davis Medical Centre (collection-to-analysis time ≤80 minutes). Spurious hypoxemia was defined paO₂ <60 mmHg (reported at the patient's body core temperature) with oxygen saturation ≥90% by co-oximetry. Kaplan–Meier analysis, Cox proportional hazards regression, and mixed-effects logistic regression assessed predictors. The institutional review board approved the study with waiver of consent (IRB ID:2065699-1). Results: Extreme leukocytosis (WBC ≥50×10⁹/L) occurred in 31 specimens (0.4%; median 65.4×10⁹/L) and thrombocytosis (platelets ≥500×10⁹/L) in 126 specimens (1.7%; median 576×10⁹/L). Spurious hypoxemia occurred in 176 specimens (2.4%). Each 1°C decrease in body temperature increased odds of spurious hypoxemia by 85% (OR 1.85, 95% CI 1.28 to 2.70, p=0.001). Each 1 mg/dL increase in serum creatinine increased the odds of spurious hypoxemia by 31% (OR 1.31, 95%CI 1.02 to 1.68, p=0.037); and each 0.01 unit increase in pH increased the odds of spurious hypoxemia by 12% (OR 1.12, 95% CI 1.07 to 1.16, p < 0.0001. Neither leukocytosis nor thrombocytosis predicted spurious hypoxemia when added to temperature-based models. Adding cell counts did not improve model discrimination (ΔAUC ≤0.001). Conclusions: Patient temperature at specimen collection, not extreme leukocytosis or thrombocytosis, is the primary determinant of spurious hypoxemia. These findings challenge the traditional “leukocyte larceny” paradigm and have implications for blood gas specimen handling in hypothermic patients.