EXTERNAL VALIDATION OF THE SPARK-AKI CLASSIFICATION SYSTEM IN PREDICTING POSTOPERATIVE ACUTE KIDNEY INJURY AMONG FILIPINO ADULT PATIENTS UNDERGOING NON-CARDIAC SURGERY IN A TERTIARY HOSPITAL

Postoperative acute kidney injury (AKI) of any cause bears significant patient complications, but most literature on postoperative AKI have focused on cardiac surgeries due to their inherent high risk nature. As such, validated AKI risk stratification systems for cardiac surgeries are already in use in clinical practice but no such scoring system is currently used for non-cardiac ones. To address this gap in clinical practice, Park, et al (2019) first developed the SPARK-AKI scoring system in South Korea, a non-patented AKI predictive risk stratification model for non-cardiac surgeries. This study aims to validate the SPARK-AKI classification system in predicting the occurrence of post-operative AKI among adult Filipino patients undergoing non-cardiac surgery.

This is a single-center, retrospective cohort study of adult Filipino patients aged 19 years old and above who were admitted for and completed any non-cardiac surgery in a tertiary hospital from January 1, 2017 to December 31, 2022. We excluded the following patients from data analysis:

·         Patients 18 y/o and below

·         Duration of surgery less than 1 hour

·         Nephrectomy or kidney transplant surgeries

·         Any form of renal replacement therapy prior to surgery

·         Incomplete peri-operative variables based on the SPARK-AKI risk scoring

·         No creatinine determination at least 48 hours post-operatively

·         Patients already with acute kidney injury on admission

A total of 1,161 charts were included in the final data analysis, where they were scored based on their documented variables and were classified into SPARK-AKI categories accordingly.

A total of 243 patients (20.93%) developed AKI post-operatively, with note of increasing incidence of both low-risk and critical AKI as the SPARK-AKI classification increases. The highest incidence of postoperative AKI was seen among the elderly (aged ≥ 80), those with advanced kidney disease (eGFR ≤ 29), orthopedic procedures, and emergency procedures. Focusing on patients with chronic kidney disease (n = 160), 89 (55.62%) developed acute kidney injury post-operatively. Forty-five patients from the entire study population were mortalities (3.8%), 47 required renal replacement therapy (4.13%), and 66 were admitted to the ICU post-operatively (5.69%). However, the resulting sensitivity, specificity, positive predictive value, and area under the curve values of the SPARK-AKI were below acceptable ranges when applied to this population.

Table 1. Diagnostic characteristics of the SPARK-AKI classification.

The SPARK-AKI scoring system has low predictive value when applied to the Filipino population. This suggests that the biggest reason for non-uniform results in this scoring system, despite its previously good discriminating power in a different cohort, are population-specific characteristics that may not have been considered in this tool.

This abstract was also submitted for the Singapore Society of Nephrology Annual Scientific Meeting 2023 (September 30-October 1, 2023).