Share this article on:

Acute Kidney Injury After Abdominal Surgery: Incidence, Risk Factors, and Outcome

Long, Thorir E. MD; Helgason, Dadi MD; Helgadottir, Solveig MD; Palsson, Runolfur MD; Gudbjartsson, Tomas MD, PhD; Sigurdsson, Gisli H. MD, PhD; Indridason, Olafur S. MD, MHS; Sigurdsson, Martin I. MD, PhD

doi: 10.1213/ANE.0000000000001323
Patient Safety: Research Report

BACKGROUND: Acute kidney injury (AKI) is a serious complication after major surgical procedures. We examined the incidence, risk factors, and mortality of patients who sustained AKI after abdominal surgery in a large population-based cohort.

METHODS: All patients who underwent open and laparoscopic abdominal surgery (excluding genitourinary and abdominal vascular procedures), between 2007 and 2014 at the University Hospital in Reykjavik were identified and their perioperative serum creatinine (SCr) measurements used to identify AKI after surgery employing the Kidney Disease: Improving Global Outcome (KDIGO) criteria. Risk factors were evaluated using multivariate logistic regression analysis and 30-day mortality compared with a propensity score–matched control group.

RESULTS: During the 8-year period, a total of 11,552 abdominal surgeries were performed on 10,022 patients. Both pre- and postoperative SCr measurements were available for 3902 (33.8%) of the surgical cases. Of these, 264 (6.8%) were complicated by AKI; 172 (4.4%), 49 (1.3%), and 43 (1.1%) were classified as KDIGO stages 1, 2 and 3, respectively. The overall incidence of AKI for patients with available SCr values was 67.7 (99% confidence interval [CI], 57.7–78.6) per 1000 surgeries. In logistic regression analysis, independent risk factors for AKI were female sex (odds ratio [OR] = 0.68; 99% CI, 0.47–0.98), hypertension (OR = 1.75; 99% CI, 1.10–2.74), preoperative chronic kidney disease (OR= 1.68; 99% CI, 1.12–2.50), ASA physical status classification of IV (OR = 9.48; 99% CI, 3.66–29.2) or V (OR = 21.4; 99% CI, 5.28–93.6), and reoperation (OR = 4.30; 99% CI, 2.36–7.70). Patients with AKI had greater 30-day mortality (18.2% vs 5.3%; P < 0.001) compared with propensity score–matched controls.

CONCLUSIONS: AKI is an important complication of abdominal surgery. In addition to sex, hypertension, and chronic kidney disease, ASA physical status classification is an independent predictor of AKI. Individuals who develop AKI have substantially worse short-term outcomes, including higher 30-day mortality, even after correcting for multiple patient- and procedure-related risk factors.

From the *Internal Medicine, Surgical, and Perioperative Services, Landspitali–The National University Hospital of Iceland, Reykjavik, Iceland; §Faculty of Medicine, University of Iceland, Reykjavik, Iceland; and Department of Anesthesiology, Perioperative and Pain Medicine, Brigham and Women’s Hospital, Boston, Massachusetts.

Accepted for publication March 3, 2016.

Funding: Landspitali University Hospital Research Fund no. A-2014-030 internal grant.

The authors declare no conflicts of interest.

Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s website.

This study was previously presented, in abstract form at the American Society of Nephrology Kidney Week in November of 2015.

Reprints will not be available from the authors.

Address correspondence to Martin I. Sigurdsson, MD, PhD, Department of Anesthesiology, Perioperative and Pain medicine, Brigham and Women’s Hospital, 75 Francis St., Boston, MA 02115. Address e-mail to martiningi@gmail.com.

The estimated global volume of surgery is approximately 234 million cases per year.1 Acute kidney injury (AKI) is a serious perioperative complication and is associated with increased costs, prolonged hospital stay, and both short- and long-term mortality.2–4 Surgery is considered the leading cause of approximately one-third of in-hospital AKI.5 However, studies on postoperative AKI have been hampered by the lack of uniform definitions and different diagnostic criteria. Significant progress in this field has occurred in recent years, beginning in 2004 with the development and subsequent validation of the RIFLE definition and classification criteria for AKI,6 followed by the Acute Kidney Injury Network (AKIN) criteria in 2007.7 Ultimately, the Kidney Disease: Improving Global Outcomes (KDIGO) classification system, published in 2012,8 combined the 2 and defined AKI as an increase in serum creatinine (SCr) of 26.5 μmol/L over 48 hours or a 50% increase from baseline SCr, known or presumed to have occurred over 7 days, and/or urine output <0.5 mL/kg/h for 6 hours. Furthermore, severity staging of AKI was based on the magnitude of changes in SCr and urine output over time. A uniformed and standardized definition of AKI is essential for accurate diagnosis and prognostication of AKI, as well as the comparison of studies and, therefore, is important both in the clinical and in research settings. Moreover, well-defined criteria allow better analysis of risk factors, which will help prepare high-risk patients for surgery and hopefully prevent postoperative AKI.

Postoperative AKI has mostly been studied after cardiovascular surgery; the rate usually ranging between 11% and 31%, with the type of surgery (coronary bypass grafting versus valve repair or placement or combined procedure) explaining a substantial part of the difference.9–13 Other risk factors in these patients include advanced age, congestive heart failure, chronic obstructive pulmonary disease (COPD), longer cardiopulmonary bypass time, and preexisting chronic kidney disease (CKD).14 The few studies that exist on AKI after abdominal surgery have shown a difference in frequency ranging from 0.8% to 22.4%.13,15,16 This variability is likely related to differences in case mix in between studies but also discrepancies between diagnostic criteria for AKI. In addition, risk factors for AKI in this cohort of patients are poorly defined. It is, however, unlikely that the risk factors for AKI associated with cardiovascular surgery can be directly extrapolated to abdominal surgical procedures, given the absence of cardiopulmonary bypass and different pathophysiology caused by changes in intra-abdominal pressure and renal perfusion pressure during and after abdominal surgery.

We therefore examined the incidence of AKI after abdominal surgery, using the KDIGO criteria, in a cohort representing the great majority of an entire nation. In addition, we identified patient- and procedure-related risk factors for postoperative AKI and assessed the morbidity and mortality in these patients.

Back to Top | Article Outline

METHODS

Approval for the study was obtained from the IRB of Landspitali–The National University Hospital of Iceland, the National Bioethics Committee, and Data Protection Authority of Iceland. To analyze risk factors for AKI and patients outcomes, we estimated that approximately 200 patients with AKI would be adequate. Expecting an AKI incidence of at least 2% after all types of abdominal surgeries, the approximately 1500 abdominal surgical procedures performed each year in Iceland would yield more than adequate number of AKI patients over an 8-year study period. All patients who underwent abdominal surgery between January 1, 2007, and December 31, 2014, at Landspitali–The National University Hospital of Iceland were included. The University Hospital serves about 75% of the Icelandic population (329,100 inhabitants on January 1, 2015)17 and harbors the only center for nephrology and dialysis services in the country. Excluded from the study were individuals younger than 18 years of age and those requiring renal replacement therapy for end-stage renal disease before surgery.

Back to Top | Article Outline

Patient Demographics and Clinical Data

Data on all patients were collected from the electronic medical record system of the University Hospital. Surgical codes, based on the Nordic Medico-Statistical Committee (NOMESCO) Classification of Surgical Procedures (NCSP-IS, version 1.14, www.nowbase.org), were used to identify and group together abdominal operations. This classification system is based on 3 characters to define anatomical location and a 2-digit number that further defines the surgical procedure. Genitourinary and abdominal vascular procedures were excluded. All surgical procedures were divided into major surgery/minor surgery, laparoscopic/nonlaparoscopic, and emergency/nonemergency surgery for further analysis (Supplemental Digital Content, Item 1, http://links.lww.com/AA/B419). Information on patient age and sex, operative time (skin-to-skin), and the ASA physical status classification18 from the hospital surgical registry was collected. All International Classification of Diseases, Tenth Revision, diagnostic codes assigned to the patients were obtained to determine comorbidity (preoperative diagnoses) and short-term complications (discharge diagnoses) (Supplemental Digital Content, Item 2, http://links.lww.com/AA/B419). Only major complications were analyzed, as we expected minor complications to be underreported by diagnostic coding. Finally, the date of death was obtained.

All SCr measurements of the patients performed at the hospital’s clinical laboratory were obtained for the estimation of kidney function, both pre- and postoperatively.

Back to Top | Article Outline

Definitions

The incidence and severity of AKI were estimated using the SCr criteria of the KDIGO definition and classification system.8 For patients with available preoperative SCr within 30 days before surgery and postoperative SCr measurements in the first 7 days after surgery, we identified the highest SCr within 48 hours and within 7 days from surgery. Based on changes in SCr from baseline, AKI was classified into stage 1 (SCr ≥26.5 µmol/L above baseline or 1.5–1.9 times baseline), stage 2 (SCr 2.0–2.9 times baseline), or stage 3 (SCr ≥3 times baseline or postoperative SCr of ≥354 µmol/L with at least 26.5 µmol/L elevation from baseline, or initiation of renal replacement therapy).8

The preoperative baseline SCr, sex, and age were used to calculate preoperative estimated glomerular filtration rate (eGFR), using the Modification of Diet in Renal Disease study equation for standardized SCr.19 For further analysis of the cohort according to preexisting kidney function, patients were classified into 3 groups: eGFR <30, 30–60, and >60 mL/min/1.73 m2. CKD was defined as eGFR <60 mL/min/1.73 m2. The Revised Cardiac Risk Index preoperative risk factor score was generated using information on prior diagnoses of cerebrovascular insult, congestive heart failure, ischemic heart disease and type 1 diabetes, the presence of high-risk surgery (defined as open surgery entering at least the peritoneum), and a preoperative SCr level >176.8 μmol/L.20 Patients were considered to have a prior history of AKI if they had experienced an episode of AKI after previous surgery or diagnostic procedures, determined by the same KDIGO criteria.

Back to Top | Article Outline

Statistical Analysis and Data Configuration

All data were processed using custom scripts written in the JAVA programming language by one of the study authors. Scripts were tested by comparing the program output with manual examination of data from a random subset of patients. All statistics and data visualization were performed in R, version 3.0.3 (R Foundation for Statistical Computing, Vienna, Austria). For each part of the study, we considered P < 0.01 as statistically significant and report 99% confidence intervals (CIs) where appropriate.

The 99% CIs of AKI incidence rates were estimated using the exact method with a binominal model, using the Hmisc package. The differences in baseline characteristics, comorbid conditions, preoperative assessment factors, and surgery-related factors were compared between the AKI and the non-AKI groups. Continuous variables were compared using either the 2-sample t test/Wilcoxon rank sum test or analysis of variance/Kruskal-Wallis test based on normality of residuals for the data. Categorical variables were compared using either the χ2 test or Fisher exact test. The change in incidence over time was assessed with a Poisson model. Univariate and multivariate logistic regression analysis was used to evaluate patient- and procedure-related risk factors for postoperative AKI. Linearity of age and surgery duration as a continuous variable included in the AKI model were assessed by plotting the logit of AKI (log(p/(1 − p), where p is the proportion of patients with AKI) in groups of individuals with increasing age or surgery duration (Supplemental Digital Content, Items 3 and 4, http://links.lww.com/AA/B419). The duration of the surgical procedure was dichotomized at 100 minutes after observing a sharp increase in the logit of AKI around that duration (Supplemental Digital Content, Item 4, http://links.lww.com/AA/B419).

An initial multivariate logistic model was created using all preoperatively known variables demonstrating a univariate association with AKI (at P < 0.01). The model was further refined by dropping variables one by one in a stepwise manner, optimizing the Akaike information criterion, using the step function in R. Several other models were also assessed. The use of actual, instead of estimated, duration of surgery has been shown to bias models designed to support preoperative decision making because the actual duration of surgery is not known preoperatively.21 Similarly, the estimated but not the actual performed procedure is not known beforehand. These variables were therefore not included in the model, which included only preoperative variables. Nevertheless, additional models including these variables were constructed and are presented in the Supplemental Digital Content (Item 5, http://links.lww.com/AA/B419). The area under the curve of the receiving operating characteristic was compared between models that included preoperatively known variables only and models created using postoperatively known variables with a univariate association with AKI (at P < 0.01), using the DeLong test from the pROC package in R. The goodness of fit of all models was assessed using the Hosmer and Lemeshow test from the Resource Selection package in R. We also tested for interaction between age and ASA physical status classification, age and preoperative CKD, and between the ASA physical status classification and preoperative CKD (Supplemental Digital Content, Item 5, http://links.lww.com/AA/B419). Hospital length of stay (LOS) as a function of AKI was examined using a generalized linear model, adjusting for preoperative LOS and time until peak SCr level. Residuals in the models were visualized by a Q-Q plot.

To compare outcomes of AKI, a control group was created employing the propensity score–matching method. Each patient with AKI was assigned a control patient from the group without AKI, using the nearest neighbor method with the MatchIt package in R.22 This method matches a control patient from the non-AKI group to each patient with AKI based on the closest match to a distance measurement for the pair. The distance measurement is an estimate of the probability of having AKI based on a logistic regression model of AKI, using preselected matching parameters. The following parameters were matched: age, sex, baseline eGFR, ASA physical status classification, year of surgery, type of surgery, emergency surgery, operative time, diagnosis of myocardial infarction or sepsis during hospitalization, and previous history of ischemic heart disease, congestive heart failure, hypertension, diabetes mellitus, COPD, liver disease, and both malignant and benign neoplasms. The AKI and propensity score–matched groups were compared using standardized differences of the matched variables (Supplemental Digital Content, Items 6 and 7, http://links.lww.com/AA/B419).23 Contributions to the propensity score were analyzed and visualized by multiple factor analysis using the FactoMineR package on an input matrix of all variables included in the propensity matching (Supplemental Digital Content, Item 8, http://links.lww.com/AA/B419). Survival was assessed by Kaplan-Meier survival curves and 30-day survival compared between AKI cases and the propensity score–matched control group and between the different time periods using logistic regression.

Back to Top | Article Outline

RESULTS

Figure 1

Figure 1

A total of 11,552 abdominal surgeries were performed on 10,022 patients during the 8-year study period. Of those, preoperative SCr was available for 9220 procedures (79.8%, Fig. 1; Supplemental Digital Content, http://links.lww.com/AA/B419). In 3902 (33.8%) of the surgical procedures, both pre- and postoperative SCr measurements were performed, and these patients could therefore be assessed for AKI (Fig. 1). When patients with available pre- and postoperative SCr were compared with those without one or both these values, the former group had a significantly higher ASA physical status classification, shorter duration of surgery, longer hospital stay, and higher 30-day mortality (Supplemental Digital Content, Items 9 and 10, http://links.lww.com/AA/B419). In the remaining analysis, we report the incidence of AKI per surgical procedure for the subset of patient with both pre- and postoperative SCr measurements available.

Back to Top | Article Outline

Incidence

We found 264 (6.8%) cases of AKI among 236 patients, of whom 172 (4.4%), 49 (1.3%), and 43 (1.1%) were classified as AKI stages 1, 2, and 3, respectively (Table 1). The incidence of AKI was 12.8% after explorative laparotomy, 12.7% after esophageal surgery, and 9.6% and 9.4% after splenic and gastric surgical procedures, respectively (Table 1). The majority of the AKI patients met the diagnostic criteria for AKI in the first 3 postoperative days, with 151 (57.2%), 62 (23.5%), and 19 (7.2%) of the cases detected on postoperative day 1, 2, or 3, respectively.

Table 1

Table 1

When only patients with available pre- and postoperative SCr were examined, the incidence of AKI was 67.7 (99% CI, 57.7–78.6) per 1000 surgeries. Assuming that none of the individuals missing either pre- or postoperative SCr experienced AKI, the incidence of AKI would be 22.9 (99% CI, 19.4–26.7) per 1000 surgeries. The incidence of postoperative AKI was not significantly different between the first and the second half of the study period; it was 75.1 (99% CI, 60.9–93.6) per 1000 surgeries in 2007–2010 and 60.3 (99% CI, 60.9–93.6) per 1000 surgeries in 2011 to 2014 (P = 0.08).

Back to Top | Article Outline

Risk Factors

In Table 2, clinical characteristics, preoperative diagnoses, and perioperative factors between patients who did or did not develop AKI are compared for the subset of patients that had both pre- and postoperative SCr measurements available. Patients who underwent surgical procedures associated with postoperative AKI were older (69 vs 62 years, P < 0.001), had greater number of comorbid conditions, and their surgical procedures took longer time and were more complex (Table 2). Furthermore, individuals with AKI were more likely to have a prior history of postoperative AKI (17.4% vs 5.1%, P < 0.001). There was an increased frequency of AKI with more advanced CKD at baseline or 13%, 18%, and 22% for stages 3, 4, and 5, respectively (P = 0.001).

Table 2

Table 2

Table 3

Table 3

The univariate analysis identified multiple potential patient- and procedure-related risk factors associated with AKI (Table 2). Several multivariate models of AKI after abdominal surgery were created from a basic model using only preoperatively known variables exhibiting a univariate association with AKI (Supplemental Digital Content, Item 5, http://links.lww.com/AA/B419). In our final multivariate model including only variables available to the clinician preoperatively, female sex (odds ratio [OR] = 0.68; 99% CI, 0.47–0.98), hypertension (OR = 1.75; 99% CI, 1.10–2.74), preoperative CKD (OR = 1.68; 99% CI, 1.12–2.50), ASA physical status classification of IV (OR = 9.48; 99% CI, 3.66–29.2) or V (OR = 21.4; 99% CI, 5.28–93.6), and reoperation (OR = 4.30; 99% CI, 2.36–7.70) were independently associated with postoperative AKI (Table 3).

Back to Top | Article Outline

Outcomes of Patients with Postoperative AKI

Analysis of the subset of patients with the available pre- and postoperative SCr values revealed that a total of 111 (42%) AKI patients were admitted to an intensive care unit and 31 (11.7%) patients required transient renal replacement therapy during their hospitalization. Patients who developed postoperative AKI had a longer median hospital stay (16 days [interquartile range {IQR}, 7–34] vs 6 days [IQR, 4–11]; P < 0.001) compared with those without AKI. After adjusting for the duration of preoperative hospital stay (0.5 days longer LOS per day of preoperative hospital stay; P < 0.001) and time until peak SCr was reached (2.9 days longer LOS for each postoperative day until peak SCr level; P < 0.001), AKI was still associated with a longer LOS (13.6 days longer LOS for patients with AKI compared with those without AKI; P < 0.001).

The rate of in-hospital complications was higher in patients with postoperative AKI compared with patients without AKI, including the rate of perioperative acute myocardial infarction, pneumonia, and sepsis (P < 0.001; Table 4). Furthermore, the rate of these complications was higher with increasing severity of AKI (data not shown).

Table 4

Table 4

The 30-day mortality after abdominal surgery was significantly greater in patients with AKI (17.8% vs 2.1%; P < 0.001). The patients who did not survive 30 days were older (73 vs 63 years; P < 0.001) and were more likely to undergo emergency surgery (78.2% vs 40.8%; P < 0.001) compared with those who survived. Patients with AKI who reached their peak SCr value >2 days after the surgery had a comparable 30-day mortality with those patients who had their peak SCr within the first 2 postoperative days (15% vs 21%; P = 0.17). Patients who required renal replacement therapy (default to stage 3 AKI) had both a higher 30-day mortality (32% vs 12%; P = 0.01) and a longer hospital stay (34 vs 14 days; P < 0.001) compared with patients with stage 1 AKI.

Back to Top | Article Outline

Survival Comparison with a Propensity Score–Matched Control Group

Figure 2

Figure 2

A propensity score–matched control group was successfully created from the group of patients without AKI for 247 episodes of postoperative AKI in the subset of patients with available pre- and postoperative SCr measurements (Supplemental Digital Content, Items 6–8, http://links.lww.com/AA/B419). All nonmatched patients (n = 17) had missing components of the propensity score (ASA physical status classification). The AKI patients had a prolonged hospital stay compared with the propensity score–matched control group (16 vs 10 days; P < 0.001). After adjusting for the duration of preoperative hospital stay (0.3 days longer LOS per day of preoperative hospital stay; P = 0.01) and time until the peak SCr was reached (4.3 days longer LOS for each postoperative day until peak SCr; P < 0.001), AKI was no longer associated with a longer LOS (5.7 days longer for patients with AKI versus those without AKI; P = 0.02). The median (IQR) hospital stay was 14 (6–28), 26 (8–43), and 14 (7–45) days for stages 1, 2, and 3 AKI, respectively, compared with 10 (6–22) days for the propensity score–matched control group. Excluding patients who died during the hospitalization, the median (IQR) hospital stay was 14 (7–28), 32 (16–55), and 40 (18–71) days for stages 1, 2, and 3 AKI, respectively. There was a dose-response relationship between mortality and severity of AKI, with a 30-day mortality of 12.3%, 19.1%, and 43.2% for the AKI stages 1, 2, and 3 groups, respectively, compared with 5.3% of the propensity score–matched control group (P < 0.001; Fig. 2).

Back to Top | Article Outline

DISCUSSION

In this large population-based cohort study of patients undergoing abdominal surgery, AKI was independently associated with 30-day mortality after correction for comorbidities and procedure-related factors using propensity score matching. Furthermore, the ASA physical status classification independently predicted risk for the development of AKI.

The literature on the incidence of AKI after abdominal surgery is scarce. In the present study, we found that an incidence of AKI was 67.7 per 1000 surgeries. This rate likely overestimates the true incidence of AKI, as it includes only patients who had both pre- and postoperative SCr measurements available. The lack of pre- and postoperative SCr in more than half of the patients is an important limitation of our study. All patients had their surgery performed at the same hospital with a limited number of physicians providing their care. Because Iceland has a public healthcare system, there are very few direct referrals to surgeons; rather patients get referred to the University Hospital’s surgical service. Although this creates some randomness in the pairing of patient, procedure, and treating physician, information on the treating physician could not be included in our modeling, which could potentially generate bias. Patients without either pre- or postoperative SCr measurements had a lower ASA physical status classification, shorter operative time, shorter hospital stay, and lower 30-day mortality. This indicates that these patients were healthier and more likely to have undergone less extensive operations, thereby reducing their risk of developing AKI. Assuming that none of the patients who were without either pre- or postoperative SCr had experienced AKI, the lowest possible incidence of AKI would be 22.9 per 1000 surgeries. Similarly, this likely underestimates the true incidence of AKI, as some of these patients may have developed mild AKI.

In comparison, Kim et al.13 examined AKI after intra-abdominal surgery in a large multicenter study, reporting a rate of 1.1%. Patients who underwent explorative laparotomy had the highest AKI incidence, which is in line with our findings. Similarly, a study by Kheterpal et al.24 revealed an AKI rate of 1.0%. However, direct comparison of these studies and the present one is difficult, as AKI was defined as an elevation of SCr >176.8 μmol/L above baseline or requirement for renal replacement therapy within 30 days after surgery in the previous studies. Interestingly, Teixeira et al.16 examined postoperative AKI after abdominal surgery using the KDIGO criteria and found an incidence of 22.4%, which is approximately 10-fold higher than the overall rate in our study. The difference between studies is most likely explained by limiting the inclusion to patients undergoing major surgery in a tertiary care setting and who were predicted to stay in hospital for at least 48 hours. In a study from South Korea on esophageal surgery,3 the rate of AKI was 35.3% compared with our rate of 12.7%, the difference most likely explained by the large proportion of high-risk cases in the South Korean investigation. The incidence of AKI after gastric surgery in our cohort was 9.4%, which is slightly higher than the rate of 6.9% observed in a single-center study on AKI after radical gastrectomy that defined AKI as an increase in SCr by 50% over 3 days.25 Another study examining AKI after gastric surgery among cancer patients, using the KDIGO criteria, observed a rate of 14.4%.2

In the present study, the rate of AKI after hepatic and other biliary surgery (excluding cholecystectomy) was 8.3%, a rate similar to the 7.6% observed by Cho et al.26 after hepatobiliary surgery, using the AKIN criteria. However, direct comparison is difficult as the study by Cho et al. included both cholecystectomy, which is classified as high-volume and low-risk procedure and liver transplantation that is associated with a very high rate of postoperative AKI, or in the 63% to 95% range.27–29

We did not observe changes in the incidence of AKI between the first and the second half of the study period, but our study might be underpowered to detect subtle changes in incidence over a relatively short time interval. The proportion of laparoscopic procedures was similar in both periods, but a change in surgical practice toward less invasive procedures during the later years of the study cannot be ruled out. Improved perioperative care, including the initiation of lung-protective ventilation in the operating room, more aggressive treatment of patients with signs of infection instigated by the Surviving Sepsis Campaign Guidelines, and a more judicious intraoperative fluid management could also have affected the incidence of AKI during the study period.

As might be expected for the subset of patients with both pre- and postoperative SCr available, those with postoperative AKI had a higher rate of major in-hospital complications and a longer hospital stay. These findings are consistent with previous reports in both general AKI and postoperative AKI cohorts.2–4 In our study, AKI after abdominal surgery had a negative impact on 30-day mortality, which was 18.2% in the AKI patients compared with 5.3% in the propensity score–matched control group. This is consistent with studies by both Teixeira et al.16 and Kheterpal et al.,15 who observed a rate of 20.8% and 15%, respectively. Furthermore, Kim et al.13 described a rather higher 30-day mortality of 31% of patients with AKI after nonvascular intra-abdominal surgery. However, their AKI definition includes only more severe forms of AKI, and in our analysis, we used propensity score matching to account for a complex combination of patient comorbidities and surgery-related factors when comparing mortality in the AKI and non-AKI cohorts. Although the present study demonstrates a strong association between postoperative AKI and adverse outcomes, a causal relationship cannot be determined based on our findings. In fact, it is possible that the development of AKI is a marker of low physiologic reserve or part of a systemic response to tissue injury and inflammation, which, in turn, could contribute to worse prognosis.

In the current study, important independent preoperative variables associated with AKI included CKD, higher preoperative ASA physical status classification, and reoperation. Several pre- and perioperative risk factors for developing AKI after abdominal surgery have been reported, including age, obesity, hypertension, COPD, elevated preoperative SCr level, and more Revised Cardiac Risk Index risk factors, although studies have yielded somewhat discordant results.13,15,16,30 After multivariate adjustment, female sex, hypertension, preoperative CKD, ASA physical status classification of IV or V, and reoperation were identified as independent risk factors for AKI. The strong association of the commonly used perioperative mortality and cardiac morbidity risk indices with AKI is of special interest and further demonstrates their usefulness in the overall perioperative risk assessment. This is particularly important in light of the association of AKI with less favorable overall outcome and higher mortality. Identifying patients at high risk of AKI based on readily available information on sex, preoperative SCr, ASA physical status classification, and if the surgical procedure is a reoperation can provide valuable guidance for the perioperative physician when communicating to surgical colleagues and, in particular, to patients and their families on the risk of AKI after surgery.

A major strength of this study is the inclusion of all SCr measurements performed in the great majority of patients undergoing abdominal surgery in a whole nation. Furthermore, mortality data were complete for all patients. Importantly, we defined AKI using the widely accepted creatinine-based KDIGO criteria, which facilitates comparison of studies. Unfortunately, we did not have reliable information on urine output for the patient cohort. Using urine output data in addition to SCr to define AKI detects a greater number of AKI cases. However, most of the additional AKI cases are mild, as is indicated by a comparable 30-day survival between the group who had AKI diagnosed only based on urine output and the group that did not sustain AKI.31 This suggests that adding information on urine output will likely identify more cases of mild AKI with more favorable outcome.

A limitation of our study is its retrospective design. However, this was partially mitigated by propensity score matching for comorbidities, surgery type, and major outcomes, which included postoperative sepsis and acute myocardial infarction.

Back to Top | Article Outline

CONCLUSIONS

In this large and well-characterized cohort, AKI was noted to complicate 6.8% of all abdominal surgery. Furthermore, we found that AKI after abdominal surgery is associated with a higher 30-day mortality, even after correcting for multiple patient-related comorbidities by propensity score matching. Interestingly, the ASA physical status classification seems to predict AKI, even after adjusting for eGFR and other previously known risk factors. Our results should encourage vigilance in attempts to prevent the occurrence of this important postoperative complication.

Back to Top | Article Outline

DISCLOSURES

Name: Thorir E. Long, MD.

Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.

Attestation: Thorir E. Long has seen the original study data, reviewed the analysis of the data, approved the final manuscript, and is the author responsible for archiving the study files.

Name: Dadi Helgason, MD.

Contribution: This author helped design the study, conduct the study, and write the manuscript.

Attestation: Dadi Helgason has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.

Name: Solveig Helgadottir, MD.

Contribution: This author helped design the study, conduct the study, and write the manuscript.

Attestation: Solveig Helgadottir has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.

Name: Runolfur Palsson, MD.

Contribution: This author helped design the study and write the manuscript.

Attestation: Runolfur Palsson reviewed the analysis of the data and approved the final manuscript.

Name: Tomas Gudbjartsson, MD, PhD.

Contribution: This author helped design the study, conduct the study, and write the manuscript.

Attestation: Tomas Gudbjartsson has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.

Name: Gisli H. Sigurdsson, MD, PhD.

Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.

Attestation: Gisli H. Sigurdsson has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.

Name: Olafur S. Indridason, MD, MHS.

Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.

Attestation: Olafur S. Indridason has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.

Name: Martin I. Sigurdsson, MD, PhD.

Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.

Attestation: Martin I. Sigurdsson has seen the original study data, reviewed the analysis of the data, approved the final manuscript, and is the author responsible for archiving the study files.

This manuscript was handled by: Sorin J. Brull, MD, FCARCSI (Hon).

Back to Top | Article Outline

ACKNOWLEDGMENT

This research was supported by the Landspitali University Hospital Research Fund A-2014-030.

Back to Top | Article Outline

REFERENCES

1. Weiser TG, Regenbogen SE, Thompson KD, Haynes AB, Lipsitz SR, Berry WR, Gawande AA. An estimation of the global volume of surgery: a modelling strategy based on available data. Lancet 2008;372:139–44.
2. Kim CS, Oak CY, Kim HY, Kang YU, Choi JS, Bae EH, Ma SK, Kweon SS, Kim SW. Incidence, predictive factors, and clinical outcomes of acute kidney injury after gastric surgery for gastric cancer. PLoS One 2013;8:e82289.
3. Lee EH, Kim HR, Baek SH, Kim KM, Chin JH, Choi DK, Kim WJ, Choi IC. Risk factors of postoperative acute kidney injury in patients undergoing esophageal cancer surgery. J Cardiothorac Vasc Anesth 2014;28:936–42.
4. Chertow GM, Burdick E, Honour M, Bonventre JV, Bates DW. Acute kidney injury, mortality, length of stay, and costs in hospitalized patients. J Am Soc Nephrol 2005;16:3365–70.
5. Uchino S, Kellum JA, Bellomo R, Doig GS, Morimatsu H, Morgera S, Schetz M, Tan I, Bouman C, Macedo E, Gibney N, Tolwani A, Ronco CBeginning and Ending Supportive Therapy for the Kidney (BEST Kidney) Investigators. Beginning and Ending Supportive Therapy for the Kidney (BEST Kidney) InvestigatorsAcute renal failure in critically ill patients: a multinational, multicenter study. JAMA 2005;294:813–8.
6. Bellomo R, Ronco C, Kellum JA, Mehta RL, Palevsky PAcute Dialysis Quality Initiative workgroup. Acute Dialysis Quality Initiative workgroupAcute renal failure—definition, outcome measures, animal models, fluid therapy and information technology needs: the Second International Consensus Conference of the Acute Dialysis Quality Initiative (ADQI) Group. Crit Care 2004;8:R204–12.
7. Mehta RL, Kellum JA, Shah SV, Molitoris BA, Ronco C, Warnock DG, Levin AAcute Kidney Injury Network. Acute Kidney Injury NetworkAcute Kidney Injury Network: report of an initiative to improve outcomes in acute kidney injury. Crit Care 2007;11:R31.
8. Disease K. Kidney Disease: Improving Global Outcomes (KDIGO) Acute Kidney Injury Work Group. KDIGO Clinical Practice Guideline for Acute Kidney Injury. Kidney Int 2012;Supplement:1–138.
9. Karkouti K, Wijeysundera DN, Yau TM, Callum JL, Cheng DC, Crowther M, Dupuis JY, Fremes SE, Kent B, Laflamme C, Lamy A, Legare JF, Mazer CD, McCluskey SA, Rubens FD, Sawchuk C, Beattie WS. Acute kidney injury after cardiac surgery: focus on modifiable risk factors. Circulation 2009;119:495–502.
10. Grayson AD, Khater M, Jackson M, Fox MA. Valvular heart operation is an independent risk factor for acute renal failure. Ann Thorac Surg 2003;75:1829–35.
11. Swaminathan M, Hudson CC, Phillips-Bute BG, Patel UD, Mathew JP, Newman MF, Milano CA, Shaw AD, Stafford-Smith M. Impact of early renal recovery on survival after cardiac surgery-associated acute kidney injury. Ann Thorac Surg 2010;89:1098–104.
12. Robert AM, Kramer RS, Dacey LJ, Charlesworth DC, Leavitt BJ, Helm RE, Hernandez F, Sardella GL, Frumiento C, Likosky DS, Brown JRNorthern New England Cardiovascular Disease Study Group. Northern New England Cardiovascular Disease Study GroupCardiac surgery-associated acute kidney injury: a comparison of two consensus criteria. Ann Thorac Surg 2010;90:1939–43.
13. Kim M, Brady JE, Li G. Variations in the risk of acute kidney injury across intraabdominal surgery procedures. Anesth Analg 2014;119:1121–32.
14. Chertow GM, Lazarus JM, Christiansen CL, Cook EF, Hammermeister KE, Grover F, Daley J. Preoperative renal risk stratification. Circulation 1997;95:878–84.
15. Kheterpal S, Tremper KK, Englesbe MJ, O’Reilly M, Shanks AM, Fetterman DM, Rosenberg AL, Swartz RD. Predictors of postoperative acute renal failure after noncardiac surgery in patients with previously normal renal function. Anesthesiology 2007;107:892–902.
16. Teixeira C, Rosa R, Rodrigues N, Mendes I, Peixoto L, Dias S, Melo MJ, Pereira M, Bicha Castelo H, Lopes JA. Acute kidney injury after major abdominal surgery: a retrospective cohort analysis. Crit Care Res Pract 2014;2014:132175.
17. Population by sex and age 1841–2015. Statistics Iceland, 2015Available at: http://www.statice.is.
18. Saklad M. Grading of patients for surgical procedures. Anesthesiology 1941;2:281–4.
19. Levey AS, Coresh J, Greene T, Stevens LA, Zhang YL, Hendriksen S, Kusek JW, Van Lente FChronic Kidney Disease Epidemiology Collaboration. Chronic Kidney Disease Epidemiology CollaborationUsing standardized serum creatinine values in the modification of diet in renal disease study equation for estimating glomerular filtration rate. Ann Intern Med 2006;145:247–54.
20. Lee TH, Marcantonio ER, Mangione CM, Thomas EJ, Polanczyk CA, Cook EF, Sugarbaker DJ, Donaldson MC, Poss R, Ho KK, Ludwig LE, Pedan A, Goldman L. Derivation and prospective validation of a simple index for prediction of cardiac risk of major noncardiac surgery. Circulation 1999;100:1043–9.
21. Dexter F, Dexter EU, Ledolter J. Statistical grand rounds: Importance of appropriately modeling procedure and duration in logistic regression studies of perioperative morbidity and mortality. Anesth Analg 2011;113:1197–201.
22. Ho D, Imai K, King G, Stuart E. Matching as nonparametric preprocessing for reducing model dependence in parametric causal inference. Polit Anal 2007;15:199–236.
23. Austin PC. An introduction to propensity score methods for reducing the effects of confounding in observational studies. Multivariate Behav Res 2011;46:399–424.
24. Kheterpal S, Tremper KK, Heung M, Rosenberg AL, Englesbe M, Shanks AM, Campbell DA Jr. Development and validation of an acute kidney injury risk index for patients undergoing general surgery: results from a national data set. Anesthesiology 2009;110:505–15.
25. Zhang J, Feng G, Yang Y, Zhang P, Pu C, Zhao G. Acute kidney injury after radical gastrectomy: a single center study. Int Urol Nephrol 2014;46:973–7.
26. Cho E, Kim SC, Kim MG, Jo SK, Cho WY, Kim HK. The incidence and risk factors of acute kidney injury after hepatobiliary surgery: a prospective observational study. BMC Nephrol 2014;15:169.
27. Klaus F, Keitel da Silva C, Meinerz G, Carvalho LM, Goldani JC, Cantisani G, Zanotelli ML, Duro Garcia V, Keitel E. Acute kidney injury after liver transplantation: incidence and mortality. Transplant Proc 2014;46:1819–21.
28. Fraley DS, Burr R, Bernardini J, Angus D, Kramer DJ, Johnson JP. Impact of acute renal failure on mortality in end-stage liver disease with or without transplantation. Kidney Int 1998;54:518–24.
29. McCauley J, Van Thiel DH, Starzl TE, Puschett JB. Acute and chronic renal failure in liver transplantation. Nephron 1990;55:121–8.
30. Kelz RR, Reinke CE, Zubizarreta JR, Wang M, Saynisch P, Even-Shoshan O, Reese PP, Fleisher LA, Silber JH. Acute kidney injury, renal function, and the elderly obese surgical patient: a matched case-control study. Ann Surg 2013;258:359–63.
31. Kellum JA, Sileanu FE, Murugan R, Lucko N, Shaw AD, Clermont G. Classifying AKI by urine output versus serum creatinine level. J Am Soc Nephrol 2015;26:2231–8.

Supplemental Digital Content

Back to Top | Article Outline
© 2016 International Anesthesia Research Society