Changes in Serum Bicarbonate Level as a Predictor of Mortality in Surgical Intensive Care Patients
Author(s):
Maria Barahona; Andrew Tran; Rubayet Hossain; Esther Tseng; Nimitt Patel; Laura Brown; Vanessa Ho
Background:
The surgical and trauma intensive care units (ICU) treat a variety of time-sensitive life-threatening disorders. To optimize patient outcomes, early identification of patients who are at risk for deterioration is paramount. Serum bicarbonate (HCO3) level is a standard test which can provide useful prognostic value in the prediction of mortality. Change in HCO3 (ΔHCO3) may be a proxy for the quality and adequacy of early resuscitation in a patient with acidosis. It is unknown whether ΔHCO3 would have prognostic value.
Hypothesis:
ΔHCO3, defined as (HCO3 at ICU admission)-(HCO3 in the Emergency Department), would be associated with mortality.
Methods:
We performed a retrospective chart review of admitted patients to the surgical and trauma ICUs between February 2014 and March 2022 at our hospital–a large, urban, tertiary care academic center. We examined variables including age at admission, gender, race, basic metabolic panel components at hospital arrival and ICU admission. Our main outcome of interest was in-hospital mortality; we also examined 30-day and 90-day mortality. The ΔHCO3 was calculated from serum chemistry results. We performed logistic regressions with ΔHCO3, treated as a continuous variable, as the main predictor variable and in-hospital, 30-day, and 90-day mortality as the response variables; adjusting for age, gender, and race. We calculated odds ratios (OR) and 95% confidence intervals (95% CI).
Results:
A total of 433 patients were identified. Patient demographics include mean age 56.1 ± 21.2 years; proportion male 68.1%. The race distribution in our study was White 73.0%; proportion Black 21.2%; and proportion Asian 1.2%. In our regression model, ΔHCO3 was a statistically significant predictor of reduced in-hospital mortality (OR: 0.916; 95% CI: [0.852, 0.982]; p < 0.05), 30-day mortality (OR: 0.921; 95% CI: [0.858, 0.987]; p < 0.05), and 90-day mortality (OR: 0.908; 95% CI: [0.845, 0.973]; p < 0.05).
Conclusions:
An increase in ΔHCO3 which occurred between the Emergency Department and ICU admission was associated with better in-hospital survival and survival up to 90 days. The improvement in mortality with correction of acidosis likely corresponds to better resuscitation, especially in sepsis. Our finding potentially highlights the use of ΔHCO3 as a cheap, minimally invasive, and early predictor of patient outcomes.