Mild Increases in Plasma Creatinine after High-Risk Abdominal Surgery Are Associated with Long-Term Renal Injury: A Retrospective Cohort Study

Background: The impact of mild acute kidney injury (AKI) observed in the immediate postoperative period after major surgery on long term renal function remains poorly dened. According to the “Kidney Disease: Improving Global Outcomes” (KDIGO) classication, a mild injury corresponds to a KIDIGO stage 1, characterized by an increase in creatinine of at least 0.3 mg/dl within a 48-hour window or 1.5 to 1.9 times the baseline level within the rst week post-surgery. We tested the hypothesis that patients who underwent moderate-to high-risk abdominal surgery and developed mild AKI in the following days would be at an increased risk of long-term renal injury compared to patients with no postoperative AKI. Methods: In this single centre retrospective study, all consecutive adult patients with a plasma creatinine value ≤ 1.5 mg/dl who underwent high-risk abdominal surgery between 2014-2019 and who had at least three recorded creatinine measurements (before surgery, during the rst seven postoperative days, and at long-term follow up [6 months-2 years]) were included. AKI was dened using a “modied” (without urine output criteria) KDIGO classication as mild (stage 1 characterised by an increase in creatinine of > 0.3 mg/dl within 48-hours or 1.5-1.9 times baseline) or moderate-to-severe (stage 2-3 characterised by increase in creatinine 2 to 3 times baseline or to ≥ 4.0 mg/dl). Development of long-term renal injury was compared in patients with and without postoperative AKI Results: Among the 815 patients included, 109 (13.4%) had postoperative AKI (81 mild [KDIGO 1] and 28 moderate-to-severe [KDIGO 2-3]). The median long-term follow-up was 360, 354 and 353 days for the three groups respectively (P=0.190). Patients who developed mild AKI had a higher risk of long-term renal injury than those who did not (odds ratio 3.1 [95%CI 1.7-5.5]; P<0.001). In multivariable analysis, mild postoperative AKI was independently associated with an increased risk of developing long-term renal injury (adjusted odds ratio 4.5 [95%CI 1.8-11.4]; P=0.002). Conclusions: Mild AKI after high-risk abdominal surgery is associated with a higher risk of long-term renal one year after surgery.

In these study populations, AKI has consistently been reported to be associated with increased lengths of hospital stay, higher readmission rates and greater healthcare costs [12][13][14][15]. Development of AKI in these patients has also been associated with altered short and long-term clinical outcomes, including death [10,16].
One of the most common systems used to diagnose AKI is the "Kidney Disease: Improving Global Outcomes" (KDIGO) classi cation, in which kidney dysfunction is based on changes in serum creatinine and urine output [17]. However, as urine output is rarely accurately measured in the perioperative setting, postoperative AKI is frequently assessed based on an increase in serum creatinine alone. Mild kidney injury is more frequent than moderate or severe injury after a major surgical procedure. Unfortunately, these three stages are generally combined into a global composite of "AKI", ignoring the obvious differences in incidence and severity across the stages. Moreover, many clinicians (surgeons, intensivists and anaesthesiologists), routinely under-recognise the importance of mild AKI, [18] as they often consider postoperative AKI to be a transient phenomenon without short or long-term consequences [19]. However, Turan et al. recently reported that mild postoperative AKI could affect long-term renal function in patients who had had various non-cardiac surgical procedures [20]. Nevertheless, there is limited speci c literature regarding the long-term renal consequences of mild AKI after major abdominal surgery [19][20][21].
We therefore conducted a retrospective cohort study to analyse the association between postoperative AKI and long-term renal injury after high-risk abdominal surgery. We hypothesised that patients with a slight increase in their postoperative plasma creatinine, corresponding to mild AKI, would be at higher risk of long-term renal injury compared to patients without postoperative creatinine increase. .

Methods
The Ethics Committee of Erasme, Brussels, Belgium on February 10th, 2020, approved this single centre retrospective cohort analysis (Reference: P2020/031). Data collection was performed by Z.M in our institution between February 11th and April 1st 2020.
We included all consecutive adult patients (≥ 18 years old) who: 1) had undergone elective high-risk abdominal surgery (including hepatobiliary surgery, pancreatectomy, gastrectomy, oesophagectomy, cancer debulking, and cystectomy) under general anaesthesia between January 1st, 2014, and April 30th, 2019. Patients who had had major vascular surgery were also included if the surgery involved an abdominal incision (e.g., aorto bifemoral bypass and abdominal aortic aneurysm surgery); 2) had a plasma creatinine value measured before surgery, within 7 days after surgery, and at a later follow up visit (6 months to 2 years after surgery).
Patients who received dialysis in the preoperative period, those with chronic kidney disease (prede ned as a baseline creatinine level > 1.5 mg/dl), those who had emergency surgery and patients who had another surgical procedure in the two years following their rst surgery (unless it was a redo surgery in the same admission) were excluded. Patients who had suprarenal clamping during their vascular surgery were also excluded as this clamping phase can seriously impact renal function.
For each eligible patient, we recorded, from our hospital health records, the plasma creatinine concentration prior to surgery (the most recent result available in the three months before surgery), the highest creatinine concentration during the seven postoperative days, and the creatinine concentration at long-term follow up (between 6 months and 2 years; if multiple creatinine values were available, the measurement closest to one year following surgery was always selected). If no long-term creatinine measurement was available in the hospital database, attempts were made to contact the patients and/or their general practitioners to obtain any values that had been measured elsewhere.
The change in creatinine concentration between the preoperative and the postoperative period was used to classify patients according to a "modi ed" KDIGO classi cation in which the urine output criteria were not considered [17]. Mild AKI (KDIGO stage 1) was characterised by an increase in creatinine of ≥ 0.3 mg/dl within 48-hours or 1.5-1.9 times baseline; moderate AKI (KDIGO stage 2) by an increase in creatinine of 2-2.9 times baseline; and severe (KDIGO stage 3) was characterised by an increase in creatinine 3 times baseline or to ≥ 4.0 mg/dl). To simplify our statistical analysis because of the low occurrence rate, AKI stages 2 and 3 were combined into a single category (2-3), leaving us with three nal groups (no AKI vs AKI stage 1 vs AKI stage 2-3).
Long-term renal injury was de ned using the difference between the preoperative creatinine concentration and the long-term follow-up measurement. We used the same KDIGO classi cation system to stage longterm renal injury as we used for the immediate postoperative period.

Statistical Analysis
Distribution of continuous data was analysed using a Kolmogorov-Smirnov test. Normally distributed data are presented as means ± standard deviation and were compared between groups using a one-way analysis of variance. Non-normally distributed data are presented as medians (interquartiles ranges) and were compared using a Kruskall-Wallis test. Dichotomous variables are presented as crude numbers and percentages and were compared between groups using a Chi-square test. Modelling of the risk of longterm renal injury was performed using the same approach as Turan et al [20] including early AKI and all covariates listed in Tables 1 and 2 in a logistic (binomial) model. The risk of developing long-term renal injury is presented as an odds ratio with 95% con dence intervals. Statistical analyses were done using Minitab 16 (Paris, France and Medcalc Software LTD, Ostend, Belgium) and R (www.r-project.org)

Results
Among the 1482 patients who underwent high-risk abdominal surgery between January 1st 2014 and April 30th 2019, 815 patients met the inclusion criteria and were thus included in our study. The main reason for exclusion was lack of postoperative or long-term creatinine values (Fig. 1).
Among the 81 patients who developed mild postoperative AKI, 19 patients (23.5%) had persistent mild or moderate-to-severe renal injury one year after surgery, compared to 64 (9.1%) of those who had no postoperative AKI (P < 0.001) (Fig. 2). Among the 28 patients (3.4%) who developed moderate to severe AKI postoperatively, 10 (35.7%) had some degree of long-term renal dysfunction. Patients who developed mild AKI after surgery therefore had a threefold higher chance of developing long-term renal injury compared to patients without postoperative AKI (odds ratio [95% CI] of 3.1 [1.7-5.5]; P = 0.0001). In patients with postoperative AKI KDIGO stage 2-3, the odds ratio for development of long-term renal dysfunction was 5.6 (95% CI 2.5-12.6; P < 0.0001).
Occurrence of postoperative AKI was associated with older age, higher baseline creatinine level and presence of comorbid conditions, notably a history of chronic hypertension, myocardial infarction, atrial brillation, or chronic obstructive pulmonary disease (Table 1). Patients who developed postoperative AKI had longer surgery times, received more uid and had a higher estimated blood loss during surgery compared to patients without postoperative AKI (Table 2). Patients with postoperative AKI also had more postoperative complications and a longer hospital length of stay (Table 3). Importantly, long-term creatinine values were measured around one year in median among all groups (Table 3; P = 0.190)

Discussion
In a cohort of 815 patients who underwent high-risk abdominal surgery, more than one fth of patients (21%) who developed mild postoperative AKI had mild renal injury long term, and 2.5% developed moderate to severe long-term renal injury. This observation demonstrates that even a slight increase in postoperative creatinine can be important and should not be neglected. Stated a different way, development of mild postoperative AKI more than tripled the odds of having renal injury one year after surgery compared to patients without postoperative AKI.
Our results are in agreement with the only available study which examines the association of mild AKI with long-term renal injury [20]. This study, recently published by Turan et al, utilized a large database from the Cleveland Clinic that included more than 15,000 patients who underwent a variety of noncardiac surgical procedures ranging from low to high-risk. Interestingly, postoperative AKI was a complication in only 3% of their study population compared to 13.4% in our study. This is not surprising as we included only patients who had had high-risk abdominal surgery that carries a greater risk of postoperative renal dysfunction than low-risk abdominal surgery. Moreover, major surgery is a well-known contributing factor for postoperative AKI [22]. This increased risk is likely due to larger uid shifts, blood losses and a relatively high incidence of perioperative hypotension in these patients, all of which can compromise renal blood ow [23][24][25][26]. Moreover, these types of surgical procedure are more often performed in elderly patients who are at a greater risk of having comorbidities that predispose to development of AKI. This study has some limitations that should be taken into consideration. Firstly, our study was observational, retrospective, and single-centre and included a relatively small sample size, largely because of the high proportion of patients without long-term follow up creatinine concentrations.
Therefore, a causal relation cannot be proven. Secondly, we only included patients who underwent highrisk abdominal surgery [27] so that the data cannot be extrapolated to other types of surgery (neurosurgical, cardiac, etc). Thirdly, as urine output was not used for the AKI classi cation ("modi ed" KDIGO classi cation), this may have led to an "underestimation" of the incidence of postoperative AKI in our study cohort. Lastly, our logistic regression only took into account perioperative variables so that we did have information on speci c events during the long-term follow-up (oncological evolution or cardiovascular problems).

Conclusions
Although mild increases in postoperative plasma creatinine concentration are frequently considered to have little long-term clinical signi cance, we found that patients undergoing high-risk abdominal surgery who developed a mild increase in plasma creatinine concentration had a much higher incidence of longterm renal dysfunction. Clinicians should not neglect "minor" disturbances in renal function after surgery as they may persist or even worsen during long-term follow up. Jean-Louis Vincent is Editor-in-Chief of Critical Care. He has no other con icts related to this article.
The other authors have no con icts of interest related to this article. V: Data analysis and editing the nal manuscript.
R: Study design, statistical analysis and editing the nal manuscript.
VdL: Study design and conception, statistical analysis and editing the nal manuscript.