Blood anion gaps and venoarterial carbon dioxide gradients as risk factors in long-term extracorporeal support.

Increases in the blood anion gap (AG) and venoarterial carbon dioxide gradients [p(V-A)CO2] may indicate changes in intracellular acid concentration not demonstrated by blood gas measurements. This prospective study examines these two physiologic markers to determine their relationship to extracorporeal membrane oxygenation (ECMO) survival and duration in 100 patients. Serum electrolytes were drawn every 6 hours and the AG calculated. Simultaneous arterial blood gases and venous blood gases were drawn every 4 hours and the p(V-A)CO2 calculated. Cumulative averages were then calculated from all the AG and p(V-A)CO2 values during each ECMO treatment. The average AG was 11 mEq/L. The average p(V-A)CO2 was 9 mm of mercury (mmHg). Patients with an AG of 11 mEq/L or less had a 12% mortality and those with a higher AG had a 43% mortality (p = 0.0005). Patients with a p(V-A)CO2 of less than 9 mmHg had a 13% mortality and those with a 9 mmHg or higher gradient had a 35% mortality (p = 0.0126). Patients with both a low AG and a low p(V-A)CO2 had a 7% mortality and survivors were on ECMO 100 (+/-37) hours. Patients with both a high AG and a high p(V-A)CO2 had a 56% mortality and survivors were on ECMO 190 (+/-105) hours. Both mortality and survivors' ECMO time increase as one or both risk factors increase. Patients with increases in both risk factors have a mortality rate 8 times greater and survivors remain on ECMO almost twice as long as those without increased risk factors. Patients may benefit from a perfusion strategy that seeks to minimize the AG and p(V-A)CO2.