Study design and participants
This retrospective, observational, single-center cohort study enrolled patients who underwent open surgical AAA repair from January 2007 to April 2019 at our institution. The exclusion criteria were preoperative estimated glomerular filtration rate (eGFR) < 15 ml/min/1.73 m2. For each patient, only the first surgical AAA repair was included in the analysis.
Surgical indications and procedure
Surgical indications were based on the 2011 Japanese Circulation Society Guidelines for Diagnosis and Treatment of Aortic Aneurysm and Aortic Dissection [14], as follows: a maximal transverse diameter ≥ 55 mm in males and ≥ 50 mm in females, an increase in the transverse diameter ≥ 5 mm/6 months, or an infected aneurysm.
For AAA, open repair was generally performed at our institution; however, the vascular team considered endovascular aneurysm repair (EVAR) if the AAA satisfied any of the following conditions: anatomical characteristics that made EVAR favorable [14], age > 75 years, history of open abdominal surgery, and history of smoking. In Japan, EVAR has been covered by health insurance since 2007.
Anesthetic management
In all patients, general anesthesia was performed with concomitant electrocardiography, pulse oximetry, non-invasive blood pressure measurement, and capnometry. An arterial cannula was inserted in the radial artery and a central venous catheter was inserted into the jugular vein. Arterial pressure and central venous pressure were continuously monitored. Anesthesia was managed by the attending anesthesiologist. Epidural anesthesia was also performed unless patients had a low platelet count or coagulopathy, or were undergoing antiplatelet or anticoagulation therapy. Patients were transferred to the surgical intensive care unit (ICU) after open AAA repair.
The attending anesthesiologist decided whether to administer an amino acid preparation, and if so, the dose. The amino acid preparation consisted of a standard mixture of amino acids (AMIPAREN Injection; Otsuka Pharmaceutical). Its composition is listed in Table 1.
Diagnostic criteria for AKI
AKI was defined by the Kidney Disease: Improving Global Outcomes criteria, as follows: an increase in serum creatinine of ≥ 0.3 mg/dl within 48 hours; an increase in serum creatinine to ≥ 1.5 times baseline within the previous 7 days; or urine volume < 0.5 ml/kg/h for 6 hours [15]. Baseline creatinine was defined as the most recent creatinine value before surgery.
Since urine output was no longer measured hourly after discharge from the ICU, we defined creatinine-AKI as an alternative to AKI diagnosed based only on creatinine. The criteria for creatinine were the same as above.
Data acquisition
We retrospectively collected the following data from medical records: age, sex, height, weight, American Society of Anesthesiologist physical status, location of the cross-clamp, anesthesia time, operative time, date of surgery, intraoperative blood loss, intraoperative urine output, intraoperative fluid volume, intraoperative transfusion volume, and intraoperative volume of amino acids administered, and serum creatinine (before surgery and on postoperative days (PODs) 0–7).
Statistical analysis
Continuous variables were summarized as medians and interquartile ranges. Categorical variables and ordinal variables were summarized as numbers and percentages (%).
We estimated the cumulative probability of incident postoperative AKI, which was the primary outcome of this study, using the product-limit estimator. Furthermore, we assessed the effect of amino acid infusion on postoperative AKI using Fine and Gray proportional hazards regression. In the regression model, we treated discharge from ICU as the competing risk event for AKI [16]. To avoid overfitting, we selected the following potential confounders a priori based on clinical priority for adjustment in the regression model: age, sex, body mass index, preoperative serum creatinine, location of the cross-clamp, and year of surgery. Non-linear associations between continuous variables and the outcome were considered using restricted cubic spline functions. We also assessed the impact of the treatment on creatinine-AKI using multivariable logistic regression. In addition to the covariates described above, American Society of Anesthesiologists physical status, anesthesia time, intraoperative fluid volume, and intraoperative transfusion volume were included because the number of subjects with the event was sufficient to avoid overfitting. In the regression models described above, all missing values were imputed using the multiple imputation method.
In addition, we used multivariable linear regression to examine the effect of amino acid infusion on the postoperative trajectory of serum creatinine. Since the outcome variable was measured at multiple time points repeatedly, we used a robust estimation method with a Huber-White variance–covariance matrix to account for the correlation between repeated measurements within each patient. Adjustments for covariates were conducted similar to adjustments for creatinine-AKI with multivariable logistic regression. For missing value imputation with the repeatedly measured outcome, serum creatinine values were imputed with the last observation carried forward method. Covariates were imputed via the single imputation method. All statistical analyses were performed using two-sided tests at the 5% significance level using R software version 4.1.0 (www.r-project.org).
