Patients with the American Society of Anesthesiologists physical status (ASA-PS) classification ≥ 3 have an increased risk of perioperative complications [1,2]. Multiple studies have reported that patients with ASA-PS ≥ 3 are associated with worse postoperative outcomes, including acute kidney injury (AKI) [3–5]. ASA-PS ≥ 3 is also associated with the intraoperative hypotension (IOH) [6]. The kidney is highly susceptible to ischemia, making it vulnerable to the effects of IOH.

Numerous studies have identified an association between IOH and postoperative AKI, showing that a longer mean arterial pressure (MAP) below 65 mmHg, as well as at 60 mmHg and 55 mmHg, is associated with an increased incidence of AKI [7–9]. However, because these studies focused on retrospective analyses of clinical data, the causality of this relationship remains unclear. Additionally, randomized controlled trials comparing a permissive hypotension group (< 60 mmHg) with a tight blood pressure management group (> 60 mmHg) failed to demonstrate that intraoperative blood pressure management can prevent postoperative AKI [10]. If IOH plays a significant mediating role, targeted intraoperative blood pressure management could help mitigate the risk of postoperative AKI. Conversely, if the ASA-PS exerts a direct effect, broader strategies focusing on preoperative optimization may improve outcomes.

In this study, we investigated the relationships among ASA-PS, IOH, and postoperative AKI. Specifically, we aimed to quantify the extent to which IOH mediates the association between ASA-PS ≥ 3 and postoperative AKI using the structural equation modeling (SEM)-based mediation analysis technique.

A total of 79 788 patients were included between January 2011 and December 2020, including 8671 patients (10.9%) with high ASA-PS scores (Fig. 2). Patients with ASA-PS ≥ 3 scores were more likely to be male, older, and have a higher BMI. They were also more likely to be on preoperative medications such as ACEi, ARB, and beta-blockers (Table 1). After applying IPTW, baseline imbalances in all measured confounders, except BMI, were effectively corrected, resulting in SMDs of < 0.1 (Table 2). Laboratory profiles in the high ASA-PS group showed lower hemoglobin levels; higher white blood cell counts; lower platelet counts; prolonged prothrombin and activated partial thromboplastin times; elevated glucose, aspartate aminotransferase, and alanine aminotransferase levels; increased sodium levels; and decreased albumin levels (Supplementary Material 3).

Postoperative AKI occurred more frequently (4.5% vs. 6.9%, P < 0.001) in patients with high ASA-PS scores (Table 3). These patients also experienced more AKI across all stages (stage 1 [3.9% vs. 5.6%, P < 0.001]; stage 2 [0.4% vs. 1.0%; P < 0.001]; stage 3 [0.1% vs. 0.3%, P < 0.001]), higher in-hospital mortality (1.2% vs. 4.6%; P < 0.001), and longer hospital stay (9 [6–13] vs. 12 [8–22] days, P < 0.001) and intensive care unit stay (1 [1–2] vs. 1 [1–4] days, P < 0.001).

SEM analysis showed that high ASA-PS scores were associated with a total effect estimate of 0.02 (95% CI [0.02–0.03]) on the log-odds of postoperative AKI (P < 0.001) (Table 4). Approximately 48% (34%–62%) of this excess risk was mediated by IOH, hypoalbuminemia, and anemia. IOH and hypoalbuminemia showed significant positive mediating effects on AKI (26% [18%–34%], P < 0.001; 26% [18%–35%], P < 0.001), whereas the mediating effect of anemia was not statistically significant (−4% [−8% to –1%], P = 0.098).

In the interaction analyses, sex significantly modified the effect of high ASA-PS scores on postoperative AKI (P = 0.013). Age of > 65 years did not significantly modify the effect of high ASA-PS scores on postoperative AKI (P = 0.053). The stratified total, direct, and mediated effects of high ASA-PS on postoperative AKI according to sex and age group are shown in Supplementary Material 4.

In the sensitivity analysis, treating the ASA-PS as a continuous variable, the results remained robust and consistent with those of the primary analysis (Supplementary Material 5). ASA-PS was associated with a total effect estimate of 0.01 (95% CI [0.00–0.01]) on the log-odds of postoperative AKI (P < 0.001). Approximately 88% (95% CI [51–124], P < 0.001) of this excess risk was mediated by IOH, hypoalbuminemia, and anemia. Although age and BMI exhibited significant nonlinear associations with AKI, sensitivity analysis incorporating RCS for these variables showed results comparable to those of the primary analysis (Supplementary Material 6).

In this large retrospective cohort study of patients undergoing non-cardiac surgery under general anesthesia, a substantial proportion of the increased risk of postoperative AKI in patients with high preoperative comorbidity burden (ASA-PS ≥ 3) was mediated through modifiable factors, particularly IOH, as well as hypoalbuminemia and anemia. Hypoalbuminemia has a significant effect on postoperative AKI. Albumin improves renal perfusion and glomerular filtration, acts as an antioxidant, scavenges harmful reactive oxygen species, inhibits renal tubular cell apoptosis, and binds to nephrotoxic substances [16]. Anemia also increases the risk of AKI by impairing oxygen delivery and exacerbating ischemic injury. This lack of oxygen impairs the function of highly metabolic tubular cells, triggering cellular injury and a rapid decline in kidney function [17]. IOH and hypoalbuminemia showed significant positive mediating effects on AKI, whereas anemia exhibited a non-significant mediating effect. However, the accuracy of this finding is limited because the measured hemoglobin levels did not capture all intraoperative values, suggesting the need for further studies incorporating real-time data for precise assessment. Furthermore, the low coefficient of hemoglobin suggests that the observed effect of anemia may lack true clinical significance, a result potentially attributable to adjustment for other mediators and confounders.

A substantial (51.8%) direct effect independent of IOH suggests the presence of additional pathways through which comorbidities predispose patients to AKI. These include baseline chronic kidney disease, inflammatory states, nephrotoxic drugs, and metabolic derangements that directly affect renal function perioperatively [9]. In previous studies, patients with ASA-PS scores ≥ 3 had more frequent preoperative use of ACEi, ARB, and beta-blocker, contrastive media, and a higher prevalence of laboratory abnormalities such as elevated inflammatory biomarkers, glucose, and liver enzymes that are all known risk factors for renal dysfunction [13,18].

Significant sex-related differences were observed in the relationship between the ASA-PS and postoperative AKI. One prior study has reported that young women receiving non‐cardiac surgery had a lower risk of postoperative AKI than men, whereas another study has reported no significant differences between sexes [18,19]. Our results showed that the odds of postoperative AKI associated with high ASA-PS scores were higher in women than in men. In male patients, the effect of high ASA-PS scores on AKI was primarily mediated by IOH, whereas hypoalbuminemia showed a predominant effect in female patients. Experimental data suggest that testosterone may increase the renal susceptibility to ischemia, whereas estrogen may confer protection [20,21].

Unlike sex, old age did not show a significant interaction with high ASA-PS scores regarding postoperative AKI (P = 0.053), indicating no substantial difference in the direct effect of ASA-PS on AKI between older and younger patients. In patients aged < 65 years, the effect of high ASA-PS scores on AKI was primarily mediated by IOH, whereas hypoalbuminemia showed a predominant effect in patients aged > 65 years. Synthesizing the findings across sex and age, older male patients may require significant attention because of the greater mediating effect of IOH on postoperative AKI occurrence [22,23].

This study has several important clinical implications. Although previous randomized trials have yielded mixed results, targeted intraoperative blood pressure management remains important for mitigating the downstream effects of comorbidities on renal outcomes. In high-risk patients, the simultaneous correction of hypoalbuminemia and prevention of IOH may yield meaningful differences in reducing perioperative renal injury. Additionally, the significant direct effect of comorbidities on AKI suggests that comprehensive perioperative care, including the optimization of chronic disease control and avoidance of nephrotoxic exposure, remains essential.

This study has some limitations. First, because this study was conducted at a single tertiary care center and was retrospective in design, selection and institutional biases may have influenced our findings. Furthermore, the retrospective nature of our study precludes strict adherence to the assumption of sequential ignorability required for formal causal mediation. Consequently, the potential for residual confounding remains, necessitating a cautious interpretation of statistical mediation pathways. Second, because IOH definitions vary widely across studies, the mediation results based on our selected absolute intraoperative MAP threshold of 60 mmHg may differ if alternative thresholds or relative definitions are applied [8,10]. To further explore the impact of intraoperative MAP, we applied different thresholds for mediation analyses; their findings are detailed in Supplementary Material 7. Third, although cardiac surgery was excluded, the remaining surgical types were diversely distributed between the high and low ASA-PS groups that requires caution when interpreting the study results. Fourth, intraoperative vasopressor use might have affected renal perfusion that in turn could have influenced the relationship between IOH and AKI. However, this aspect could not be analyzed in the present study owing to insufficient data. Fifth, although the preoperative use of ACEi and ARB may have a direct effect on postoperative AKI, we only considered their mediating effect on IOH. Finally, the exclusion of approximately 50% of the patients because of missing postoperative creatinine values likely introduced a significant selection bias. Stable patients from less complex procedures may have been omitted based on clinical judgment, requiring caution when interpreting the generalizability of the results.

In conclusion, a high ASA-PS score is associated with an increased risk of postoperative AKI that is partially mediated by IOH, hypoalbuminemia, and anemia during non-cardiac surgery. Targeted blood pressure management, correction of albumin and hemoglobin levels, and comprehensive optimization of comorbidities may reduce kidney injury and enhance recovery in this population. Further research is needed to explore strategies for optimizing preoperative conditions and managing IOH to prevent postoperative AKI and optimize patient outcomes.