Existing risk stratification methods in cardiothoracic surgery do not use pre-operative bicarbonate levels. We investigate if these levels carry predictive value when compared with demographic, hematologic, and other renal function indicators. We used the MIMIC-IV database to identify 11,261 patients on the cardiothoracic surgery unit who underwent ICD-9 or ICD-10 coded cardiac procedures. 30-day mortality was then modeled with respect to these indicators. Minimum pre-operative bicarbonate level is an independent risk factor for 30-day mortality post-cardiothoracic surgery (). The lowest quartile with respect to bicarbonate had increased mortality compared to all other quartiles, suggesting that patients may be classified into high- or low-risk based on bicarbonate lesser or greater than 20 mmol/L. Minimum pre-operative bicarbonate levels carry predictive information with respect to post-operative mortality. Bicarbonate could be considered alongside existing risk stratification methods in addition to other renal function indicators to improve risk stratification.