Cardiac Surgery Associated Acute Kidney Injury (CSA-AKI) in patients with chronic kidney disease (CKD): A Report from the STS Database

D. T.. Engelman¹, N. Parsons², D. Oleske³, C. Solinsky⁴, P. Winterberg⁵, K. W.. Lobdell^6
1Heart & Vascular Program, Baystate Health, University of Massachusetts Medical School, Baytate, Springfield, Massachusetts ²STS, Royse City, Texas ³Alexion, AstraZeneca Rare Disease, Boston, Illinois ⁴Alexion, AstraZeneca Rare Disease, Medford, Massachusetts ⁵Alexion, AstraZeneca Rare Disease, Milton, Massachusetts ⁶Sanger Heart & Vascular Institute, Charlotte, North Carolina

Purpose: AKI occurs in ~50% patients with CKD undergoing cardiac surgery1 compared with ~22% patients without CKD2 and is associated with increased costs, complications, and mortality.3 The primary aim of this investigation was to assess CSA-AKI incidence and other outcomes for patients with variable degrees of preoperative CKD.

Methods: We studied patients with CKD who underwent coronary artery bypass grafting (CABG), valve, and CABG + valve procedures between 1/1/2015 and 12/31/2021 with cardiopulmonary bypass (CPB) in the US Society of Thoracic Surgeons Adult Cardiac Surgery Database (STS ACSD). CKD was defined by estimated preoperative glomerular filtration rate (eGFR) between 20 and < 60 L/min/1.73m2 using the CKD-EPI GFR estimating formula. Exclusion criteria included emergency cases, history of cancer, off-pump procedures, and preoperative dialysis or eGFR of < 20 or ≥60 mL/min/1.73m2.
CKD was stratified by Kidney Disease: Improving Global Outcomes (KDIGO) GFR staging G3a, G3b, and G4 (Table 1). Postoperative AKI was defined by KDIGO staging 1 to 3 using peak postoperative creatinine.4
The expected risk of postoperative renal failure (KDIGO stage 3 AKI) and operative mortality were estimated using the STS procedure-specific risk models.

Results: Final study population included 287,359 (of 2,065,752 evaluated) patients from 1,132 US hospitals (Table 1). Median age of final study population (all CKD stages) was 72 years (interquartile range: 65, 77); 36.6% female; 6.6% African-American. Isolated CABG was the most common procedure type (62.9% overall). The procedure types and proportions, aortic cross clamp, and CPB times were similar across all CKD stages.
The observed incidence of stage 3 KDIGO AKI postoperatively increased with higher preoperative CKD stage (Figure 1a), as did requirement for postoperative dialysis (G3a 2.5%, G3b 5.7%, G4 14.4%) and operative mortality (G3a 3.3%, G3b 5.1%, G4 7.4%). The proportion of patients receiving red blood cell transfusions also increased with worsening CKD stage (G3a 46.4%, G3b 61.2%, G4 75.0%).
Expected estimates of postoperative renal failure (approximating KDIGO stage 3 AKI) using the STS risk model underestimated the observed rates by 7%, 17%, 32% in CKD G3a, G3b, and G4 categories, respectively (Figure 1a). Observed operative mortality was higher than baseline mortality risk estimates by 19%, 19%, 25% in CKD G3a, G3b, and G4 categories, respectively (Figure 1b).

Conclusion: Worsening preoperative CKD was associated with increasing CSA-AKI rates, dialysis and transfusion need, and mortality at 30 days. The STS ACSD procedure-specific risk models tend to underestimate AKI and mortality risk in this subgroup of patients. Additional analyses are needed to inform risk mitigation strategies in this high-risk population.

Identify the source of the funding for this research project: This analysis was funded by Alexion, AstraZeneca Rare Disease, Boston, MA.