Arterial versus mixed venous lactate levels in 1579 cardiac intensive care patients

Abstract Background/Introduction Timely diagnosis and treatment of cardiogenic shock (CS) are critical, with baseline arterial lactate being one of the earliest prognostic biomarkers 2. A threshold of ≥2 mmol/L is considered clinically significant 3. However, arterial sampling is often difficult in critical settings such as ongoing resuscitation, severe vasoconstriction, or during ECMO cannulation. This raises the clinical question of whether venous lactate can reliably substitute arterial lactate for diagnosis and monitoring in CS patients. Purpose To investigate the correlation and agreement between arterial and mixed venous lactate levels in critically ill cardiac patients, and to assess whether venous lactate can be used interchangeably with arterial lactate in this population. Methods A retrospective study was performed in 1,579 patients admitted to a quaternary cardiac intensive care unit (86% surgical, 14% non-surgical), all monitored with pulmonary artery catheters. Simultaneous radial arterial and mixed venous blood gas lactate levels were obtained, resulting in 4,801 paired samples (median interval 8 min, IQR 3–18). Data were analyzed using regression, Bland-Altman analysis, and mixed-effects modeling. Results Patients were predominantly male (67%), mean age 59.6 ± 16.3 years, with high illness severity (mean VIS 19.8 ± 13.6, ICU mortality 31.4%). Arterial and venous lactate values were strongly correlated across the entire cohort (p 0.0001, R² = 0.99), with a mean difference of 0.02 ± 0.32 mmol/L. In samples with arterial lactate 2–5 mmol/L 3 (n = 1,250), venous and arterial lactates remained highly correlated (R² = 0.95; mean difference 0.04 ± 0.30 mmol/L). In samples with arterial lactate 5 mmol/L 3 (n = 591), correlation was preserved (R² = 0.97; mean difference 0.05 ± 0.71 mmol/L). Mixed-effects modeling confirmed no significant difference between sampling sites (p = 0.653), nor any significant interaction over time (p = 0.219). Conclusion(s) This is the first large-scale study in cardiac ICU patients (1,579 patients; 4,801 paired samples) to demonstrate strong and consistent correlation and agreement between arterial and mixed venous lactate levels across a wide range of concentrations. The interchangeability of these measures supports faster decision-making when arterial access is difficult, reduces redundant blood sampling, and has important implications during ECMO cannulation and cardiopulmonary resuscitation. Moreover, as persistent arterial–venous correlation has previously only been shown in septic shock 4,5, this study extends these findings to CS, highlighting mechanistic differences between shock types while validating clinical flexibility. Finally, the results provide rationale for integrating continuous lactate monitoring into pulmonary artery catheter systems, advancing precision care in circulatory failure.Paired art. vs ven. lactate