Secondary Logo

Acute Kidney Injury and Risk of Death After Elective Surgery

Prospective Analysis of Data From an International Cohort Study

Chaudery, Hannan, BSc*; MacDonald, Neil, MBBS; Ahmad, Tahania, MPH*; Chandra, Susilo, MD; Tantri, Aida, MD; Sivasakthi, Velayuthapillai, MD§; Mansor, Marzida, MD; Matos, Ricardo, MD; Pearse, Rupert M., MD*; Prowle, John R., MD* on behalf of the International Surgical Outcomes Study (ISOS) Group

doi: 10.1213/ANE.0000000000003923
Basic Science: Original Clinical Research Report
Free
SDC

BACKGROUND: Postoperative acute kidney injury (AKI) is associated with a high mortality rate. However, the relationship among AKI, its associations, and mortality is not well understood.

METHODS: Planned analysis of data was collected during an international 7-day cohort study of adults undergoing elective in-patient surgery. AKI was defined using Kidney Disease Improving Global Outcomes criteria. Patients missing preoperative creatinine data were excluded. We used multivariable logistic regression to examine the relationships among preoperative creatinine-based estimated glomerular filtration rate (eGFR), postoperative AKI, and hospital mortality, accounting for the effects of age, major comorbid diseases, and nature and severity of surgical intervention on outcomes. We similarly modeled preoperative associations of AKI. Data are presented as n (%) or odds ratios (ORs) with 95% confidence intervals.

RESULTS: A total of 36,357 patients were included, 743 (2.0%) of whom developed AKI with 73 (9.8%) deaths in hospital. AKI affected 73 of 196 (37.2%) of all patients who died. Mortality was strongly associated with the severity of AKI (stage 1: OR, 2.57 [1.3–5.0]; stage 2: OR, 8.6 [5.0–15.1]; stage 3: OR, 30.1 [18.5–49.0]). Low preoperative eGFR was strongly associated with AKI. However, in our model, lower eGFR was not associated with increasing mortality in patients who did not develop AKI. Conversely, in older patients, high preoperative eGFR (>90 mL·minute−1·1.73 m−2) was associated with an increasing risk of death, potentially reflecting poor muscle mass.

CONCLUSIONS: The occurrence and severity of AKI are strongly associated with risk of death after surgery. However, the relationship between preoperative renal function as assessed by serum creatinine-based eGFR and risk of death dependent on patient age and whether AKI develops postoperatively.

From the *William Harvey Research Institute, Barts and The London School of Medicine & Dentistry, Queen Mary University of London, London, United Kingdom

Department of Anaesthesia, Royal London Hospital, Barts Health NHS Trust, London, United Kingdom

Universitas Indonesia, Ciptomangunkusumo Hospital, Jakarta, Indonesia

§Ministry of Health of Malaysia, Putrajaya, Malaysia

Department of Anaesthesiology, Faculty of Medicine, University of Malaya, Kuala Lumpur, Malaysia

Unidade de Cuidados Intensivos Polivalente Neurocríticos, Hospital de S. José, Centro Hospitalar de Lisboa Central, E.P.E, Lisboa, Portugal.

Published ahead of print 3 October 2018.

Accepted for publication October 3, 2018.

Funding: This was an investigator-initiated study funded by Nestle Health Sciences through an unrestricted research grant, and by a National Institute for Health Research (United Kingdom) Professorship held by R.M.P. H.C. was part-supported for this research project by the John Snow Award jointly administered by the Royal College of Anaesthetics, the British Journal of Anaesthesia, and the National Institute of Academic Anaesthesia. This study was sponsored by Queen Mary University of London.

Conflicts of Interest: See Disclosures at the end of the article.

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.

Members of International Surgical Outcomes Study (ISOS) group are listed in Supplemental Digital Content, http://links.lww.com/AA/C649.

International Surgical Outcomes Study investigators were entirely responsible for study design, conduct, and data analysis. The authors had full data access and were solely responsible for data interpretation, drafting, and critical revision of the manuscript, and the decision to submit for publication.

Reprints will not be available from the authors.

Address correspondence to John R. Prowle MD, Adult Critical Care Unit, Royal London Hospital, London, E1 1BB, United Kingdom. Address e-mail to j.prowle@qmul.ac.uk.

See Editorial, p 841

KEY POINTS

  • Question: How are preoperative estimated glomerular filtration rate and postoperative acute kidney injury (AKI) associated with hospital mortality in the 30 days after surgery?
  • Findings: After accounting for comorbid disease and severity of surgery, AKI remained strongly associated with risk of death while the association between estimated glomerular filtration rate and mortality was U-shaped.
  • Meaning: AKI is strongly associated with risk of death after major surgery; patients with chronic kidney disease are only at very high risk of death if they develop AKI.

More than 300 million surgical procedures are performed worldwide every year.1 Estimates of morbidity and mortality vary.2–4 However, recent data suggest that as many as 75 million patients will experience a postoperative complication, leading to 2 million deaths each year.5,6 Acute kidney injury (AKI) is now widely considered to be a key postoperative complication, linked to increased risks of infection, prolonged hospital stay, and mortality.7–12

The findings of a large registry study of 3.6 million surgical patients demonstrated an AKI incidence of 12%,9 while a recent meta-analysis of studies of AKI after major abdominal surgery described a pooled incidence of 13%.7 The severity and nature of the operative procedure, age, comorbid disease, and intercurrent conditions such as sepsis have all been associated with postoperative AKI; however, across all settings, chronic kidney disease (CKD) is consistently identified as a leading association with the development of AKI.9,10,13 Given this strong association with AKI, it is not surprising that preexisting CKD is also strongly associated with postoperative mortality.14–17 However, it remains uncertain whether this relationship depends on the development of AKI or represents an independent risk of mortality associated with the CKD phenotype.

We performed a prospectively planned analysis of the incidence and association of AKI with postoperative outcomes using data collected during the International Surgical Outcomes Study (ISOS), which described patient outcomes after elective surgery in 27 low-, middle-, and high-income countries.6,18,19 Based on previous single-center findings,12 we hypothesized that the relationship between CKD diagnosis and death would be associated with the development of AKI. Our primary objective was to determine the association between the occurrence and severity of AKI and hospital mortality, using a multivariable model to account for potentially confounding patient and surgical variables, including preoperative creatinine-based estimated glomerular filtration rate (eGFR).

Back to Top | Article Outline

METHODS

Data Collection

This was a prospectively defined analysis of data collected during an international 7-day cohort study of adult patients undergoing elective in-patient surgery. Patients undergoing day-case, radiological, and emergency procedures were excluded. The methods of this study have been described previously.6 The study was approved by the Yorkshire & Humber Research Ethics Committee (Reference: 13/YH/0371) with waiver requirement for written informed consent as an anonymized analysis of routinely collected patient data. Data were collected until hospital discharge and censored at 30 days. Only hospitals returning valid data describing ≥20 patients, countries with ≥10 participating hospitals, and patients with complete mortality data were eligible for inclusion in the core ISOS dataset. Postoperative AKI was graded by local investigators according to the Kidney Disease Improving Global Outcomes 2012 AKI guidelines.20 eGFR was calculated from age, sex, preoperative serum creatinine, and ethnicity using the CKD-Epidemiology Research Group creatinine formula.21

Back to Top | Article Outline

Objectives

Our primary objective was to assess the independent association between hospital mortality in the 30 days after surgery (outcome) and exposure to postoperative AKI, accounting for prior renal function and other available variables potentially associated with postoperative death or AKI in multivariable logistic regression. Our secondary objective was to investigate postoperative AKI as an outcome variable, examining its independent associations with preoperative clinical and demographic data and nature of surgery as exposure variables, which enabled us to better understand patient characteristics associated with AKI that could potentially confound assessment of the relationship between AKI and mortality.

Back to Top | Article Outline

Data Analysis

We used the Strengthening the Reporting of Observational Studies in Epidemiology statement guidelines for reporting observational studies.22 All analyses were performed using R version 3.3.1 (R Foundation for Statistical Computing, Vienna, Austria; http://R-project.org). Additional data cleaning included removal of patients with biologically implausible creatinine values (<10 µmol·L−1) and patients with missing data for age, sex, or creatinine. For univariable comparison between AKI categories and between in-hospital deaths and survivors, continuous variables were compared using the Wilcoxon rank-sum test, or for multiple categories with the Kruskal–Wallis H test. Categorical variables were compared with the χ2 test. Data are presented as median and interquartile range, number (%), or odds ratios (ORs) with 95% confidence intervals (CIs). Finally, hemoglobin values <2 g·dL−1 were regarded as data entry errors and not used in analysis, and to avoid overinfluential outliers, eGFR values were capped at maximum of 180 mL·minute−1·1.73 m−2.

For our primary analysis, we examined the independent association between mortality within 30 days of surgery and postoperative AKI in a multivariable logistic regression model. For mortality modeling, all available biologically plausible potential associations were included in an initial maximal model before variable selection; these were age, sex, current smoker, American Society of Anesthesiologists (ASA) physical status score, preoperative comorbidities (ischemic heart disease, heart failure, diabetes mellitus, metastatic cancer, cirrhosis, stroke, and chronic obstructive pulmonary disease or asthma), surgical procedure category, grade of surgery (minor/intermediate/major), laparoscopic surgery, preoperative hemoglobin, preoperative eGFR, and postoperative AKI. Detailed data on intraoperative events and physiologically monitoring were not available in the ISOS dataset. We did not consider other postoperative complications in the modeling process because of the complex interrelationship with development of AKI. Missing predictor variable values were handled by case-wise deletion. To minimize the inflation of degrees of freedom in our models, surgical procedure categories were merged as follows: upper gastrointestinal, lower gastrointestinal, and hepatobiliary into “abdominal surgery”; cardiac and thoracic surgery into “cardiothoracic”; and breast, neurosurgery, head and neck, gynecology, plastics, and cutaneous surgery were merged into “other.” Orthopedic, urology and kidney, and vascular surgery were retained as independent categories. The largest category (orthopedics) was designated the reference surgical category. Similarly, ASA grades I and II were merged and used as the reference category.

Because we expected a nonlinear relationship between preoperative renal function and postoperative outcomes, we prespecified inclusion of eGFR as a continuous predictor fitted to a restricted cubic spline with 5 knots. In addition, to account for changes in renal function with age, we prespecified examination of an interaction term between age and eGFR in mortality-model development. Bootstrap stability was used to select risk factors into the model. This approach estimates the proportion of times a risk factor would be selected into the model if a different random sample of patients were taken each time. Backward variable selection, based on minimization of the Akaike Information Criterion as the threshold for exclusion (equivalent to a P value for exclusion of .15), was performed within each bootstrap sample. Risk factors selected using this backward selection approach in least two-thirds of bootstrap samples were included in a minimum model. The interaction between glomerular filtration rate and age (selected on the basis of biological plausibility) was then added to the model, with the criterion for selection of the interaction was a P < .05 in at least two-thirds of bootstrap samples.23 To facilitate model selection and presentation of CIs for model predictions, the primary mortality analysis was performed as a single-level model; however, in a sensitivity analysis also examined the association between AKI and in-hospital mortality using a mixed-effects logistic regression model with a random intercept for country. Finally, we also assessed whether the association between AKI and mortality differed according to a national income status (high- versus middle-income countries) using an interaction test. Modeling of preoperative predictors of the development of AKI was performed in a similar fashion using a single-level logistic regression including national income status.

Back to Top | Article Outline

RESULTS

Description of Dataset

The ISOS study predefined core dataset described 44,814 patients from 474 hospitals in 27 countries.6 From this, we excluded a further 16 patients with incomplete demographic data, 8291 patients with no recorded preoperative serum creatinine, and 150 patients with implausible baseline creatinine data, leaving 36,357 patients from 19 high-income countries (22,186 patients) and 7 middle-income countries (14,171 patients) for this analysis (Supplemental Digital Content, Figure 1, http://links.lww.com/AA/C649).

Back to Top | Article Outline

Univariable Analyses

Table 1

Table 1

Table 2

Table 2

Preoperative clinical and demographic characteristics are presented by AKI category in Table 1 and operative and anesthetic details in Table 2. Overall, 196 of 36,357 patients (0.5%) died within 30 days, 743 patients (2.0%) developed AKI, and 73 of 743 (9.8%) patients with AKI died in hospital with 37% of deaths after surgery being preceded by an AKI diagnosis. In the 13,918 patients classified as having major surgery, the incidence of AKI was 4%. The incidence of AKI reported in middle-income countries was 0.8% compared to 2.9% in high-income countries (P < .001); however, when AKI occurred, patients in middle-income countries were more likely to die: 24 of 107 (27%) vs 49 of 636 (7%) (P < .001). Mortality increased with AKI severity, and of the 143 patients who experienced AKI stage III, 41 (29%) died in hospital. Univariable analysis, based in hospital mortality outcome, revealed the clinical significance of AKI as a postoperative complication, with unadjusted ORs for hospital mortality of 8.88 (95% CI, 4.87–16.2), 33.3 (20.3–54.6), and 116.9 (78.1–175.1) for AKI stages I, II, or III versus no AKI (Supplemental Digital Content, Table 1, http://links.lww.com/AA/C649). The categorical variable next most strongly associated with death was ASA grade.

Back to Top | Article Outline

Multivariable Mortality Model

Figure 1

Figure 1

We examined the relationship between postoperative AKI and hospital mortality, adjusting for the confounding effect of other variables independently associated with risk of death after surgery. After variable selection, age, eGFR, ASA grade, preoperative hemoglobin, smoking status, metastatic cancer, stroke, surgical specialty, laparoscopic surgery, and severity of surgery were retained in the final logistic regression model (Figure 1; Supplemental Digital Content, Box 1, http://links.lww.com/AA/C649). In our final model, we found a significant nonlinear relationship between eGFR and mortality. In addition, a significant interaction between age and eGFR (linear component) was retained in our final model. Mortality remained strongly associated with the severity of AKI (stage 1: OR, 2.57 [1.3–5.0]; stage 2: OR, 8.6 [5.0–15.1]; stage 3: OR, 30.1 [18.5–49.0]) (Figure 1). Finally, as a sensitivity analysis, we examined a model with forced inclusion of any preoperative variables that were significantly associated with development of AKI that were not included in our mortality model after selection (previous diagnoses of heart failure or cirrhosis); in this analysis, the association between mortality and AKI was essentially unchanged (stage 1: OR, 2.50 [1.3–4.8]; stage 2: OR, 8.52 [4.9–14.9]; stage 3: OR, 29.6 [18.1–48.4]).

Back to Top | Article Outline

Relationship Among eGFR, Age, AKI, and Mortality

Figure 2

Figure 2

Figure 3

Figure 3

The relationship among preoperative eGFR, age, and AKI in our final model is illustrated in Figure 2. All AKI stages remained associated with hospital mortality. Both extremes of preoperative eGFR were associated with risk of death. However, in the absence of AKI, patients with low baseline eGFR did not confer a high mortality risk (Figure 2A). Higher eGFR (>90 mL·minute−1·1.73 m−2) was associated with increasing mortality risk among patients at or above the median age. Younger patients with low eGFR at baseline were at an increased risk of death approaching that of older patients without baseline renal dysfunction (Figure 2). This complex relationship between eGFR and risk of death only becomes fully apparent after inclusion of an interaction term between age and eGFR and of AKI diagnosis in the modeling process, as illustrated in Figure 3.

Back to Top | Article Outline

Country and National Income Variation

When modeling was repeated using a mixed-effects logistic regression model with a random intercept for country, ORs for fixed-effect covariates and model predictions remained near identical to the single-level model (Supplemental Digital Content, Figure 2, Table 2, http://links.lww.com/AA/C649). In interaction testing of our primary model, AKI diagnoses were associated with higher risk of death in middle-income countries (significant interaction between AKI and national income status [P < .001]; Supplemental Digital Content, Figure 3, Table 2, http://links.lww.com/AA/C649). No low-income countries in ISOS provided preoperative creatinine data, and hence only middle- and high-income countries were included in this substudy.

Back to Top | Article Outline

Multivariable Modeling for AKI

We examined the preoperative variables associated with development of AKI. After variable selection, the final model included the following variables: gender, ASA grade, heart failure, cirrhosis, severity of surgery, laparoscopic surgery, preoperative hemoglobin, surgical procedure category, country income category, and preoperative eGFR (Supplemental Digital Content, Figure 4 and full specification of model in Box 2, http://links.lww.com/AA/C649). In this model, preoperative eGFR was strongly associated with AKI with an inflection point followed by progressively increasing risk of AKI with eGFR below 90 (Supplemental Digital Content, Figure 5, http://links.lww.com/AA/C649). There were lower odds of AKI diagnosis in middle- versus high-income countries.

Back to Top | Article Outline

DISCUSSION

Our findings support AKI as a sentinel postoperative complication strongly associated with death after surgery. We found that in patients with previous CKD, low eGFR was associated with only a small increase in predicted mortality in absence of AKI, highlighting AKI as a potentially pivotal complication and bridging preoperative characteristics and outcomes. As an additional important finding, we demonstrated that creatinine-base eGFR above 90 mL·minute−1·1.73 m−2 at older ages was associated with increased predicted risk of death.

Back to Top | Article Outline

Comparison With Previous Studies

Rates of AKI in the ISOS cohort were low, at 2% in all patients and 4% in more severe surgery compared to reported of 6%–30% in previous studies in major surgery.7,9–12 The lower incidence of AKI may reflect a younger age group, lower incidence of comorbid disease, and different surgical case-mix in the global ISOS population. In addition, routine postoperative creatinine measurement could be less frequent in middle-income countries, reducing potential to diagnose incidental, mild AKI. Overall, AKI stage 1 comprised only half of all AKI events in ISOS, compared to three-quarters of cases in other large reports.9,10 Despite these differences in AKI incidence, the association between increasing AKI severity and mortality mirrors that seen in existing retrospective studies in high-income countries. In a study of 161,185 US veterans undergoing major surgery, there was an adjusted rate ratio of 6.7 for hospital mortality with any stage of postoperative AKI, with an 8-fold greater risk in patients experiencing stage 3 AKI compared with stage 1 AKI.9 Similarly, in a study of 39,000 patients undergoing major surgery in Germany, adjusted ORs for hospital mortality of 2.2, 3.8, and 15.6 were reported for AKI stages 1, 2, and 3, respectively.11

While the associations between CKD and postoperative AKI, and AKI and death are not novel findings,15–17,24,25 most previous investigators have not considered the impact of AKI when exploring associations between CKD and mortality, nor have they explored eGFR as a continuous risk but rather dichotomized into arbitrary CKD categories, reducing statistical power. Furthermore, in most studies examining preoperative renal function, eGFR >90 was considered as a reference category, precluding exploration of the association between higher eGFR and outcome.15,16,24 We also accounted for 2 important confounding factors in our modeling. First, serum creatinine concentration is directly related to the mass of muscle from which it is derived, so that apparently high eGFR could reflect cachexia and poor physical condition. Second, because both renal function and muscle mass tend to decline with age, the relationship between creatinine-based eGFR and risk of death is likely to be age dependent. Our finding of a U-shaped relationship between eGFR and risk of death is in keeping with a large retrospective study in abdominal surgery where the highest category preoperative eGFR was associated with increased mortality.17 Because a truly high glomerular filtration rate is likely to be biologically implausible in older patients, this grouping is likely to identify patients with low muscle mass, who are independently at increased risk of postoperative death. Finally, our analysis also demonstrates that younger patients with CKD have predicted risks of death approaching those of much older patients with the same level of renal function. This suggests that younger patients with advanced CKD represent a very high–risk group compared to their age-matched peers with normal renal function, in keeping with the association between CKD and a prematurely aged phenotype,26 and/or the presence of other serious comorbid conditions in this group.

Back to Top | Article Outline

Strengths and Weaknesses

As a secondary analysis of ISOS, the strengths of this study include the large number of consecutive patients prospectively enrolled worldwide. A simple data set of objective, routinely collected, and easily transcribed variables minimized missing data and maximized data fidelity. Due to the short period of data collection, the study was unaffected by changes in patterns of care and differences in definition of AKI. There are, however, also limitations to our analysis. While ISOS dataset is international, there were varying degrees of participation in middle- and high-income countries. The low incidence of AKI suggests inclusion of only more clinically overt AKI, so that risk associated with AKI may be smaller if AKI surveillance is better. Thus, due to differences in AKI diagnosis rate and smaller number of patients, we cannot attribute the larger association between AKI and risk of death in middle-income countries to differences in care. Furthermore, definitions of surgical specialties and severity of surgery fail to capture detailed information on the procedures, and these may vary greatly around the world. Overall, further studies in diverse surgical populations are required to confirm and generalize subgroup analyses.

Because AKI diagnoses were determined locally, we have no information on breakdown of AKI definitions by urine output or creatinine criteria; however, because few patients are likely to have had hourly urine output recording, our results are likely to predominantly reflect creatinine-based diagnosis. Nor do we have details on timing of AKI diagnosis after surgery. In addition, many potential AKI risk factors are unavailable in the ISOS dataset, most importantly intraoperative variables including duration of surgery and physiological variables during and after anesthesia. Furthermore, while we have suggested that the relationship between higher eGFR and mortality at older age may relate to low muscle mass, we have no height or weight data to corroborate this hypothesis. Finally, despite the strength of the association seen between AKI and risk of death in this and other studies, retrospective associations cannot fully answer questions of causality. However, the strength of the observed effect, the biological gradient of the association between AKI severity and risk of death, the consistency of these results with a range of other studies,8–11,27 and plausible mechanisms involving “organ cross-talk”28 satisfy many of the epidemiological criteria for causation.29

Back to Top | Article Outline

CONCLUSIONS

These findings confirm AKI as a clinically meaningful and easily measurable surgical outcome strongly associated with risk of death. In an international surgical cohort, preoperative CKD was the strongly associated with AKI; however, the majority risk of death in CKD appears specifically associated with occurrence of AKI. In addition, abnormally high eGFR (a low serum creatinine) among older patients was also associated with increasing risk of death.

Back to Top | Article Outline

DISCLOSURES

Name: Hannan Chaudery, BSc.

Contribution: This author helped conceive and plan the International Surgical Outcomes Study–Acute Kidney Injury (ISOS-AKI) substudy; undertake detailed study design and statistical analysis planning; perform data cleaning and preparation, and primary statistical analysis; write the manuscript; prepare figures; and review and approve the final manuscript.

Conflicts of Interest: None.

Name: Neil MacDonald, MBBS.

Contribution: This author helped conceive and plan the ISOS-AKI substudy, review and approve the study design and statistical analysis plan, write the manuscript, and review and approve the final manuscript.

Conflicts of Interest: None.

Name: Tahania Ahmad, MPH.

Contribution: This author helped conceive and plan the ISOS-AKI substudy, perform data cleaning and preparation, review and approve the study design and statistical analysis plan, review the analysis and results, and review and approve the final manuscript.

Conflicts of Interest: None.

Name: Susilo Chandra, MD.

Contribution: This author helped coordinate the ISOS study and collect ISOS data, conceive and plan the ISOS-AKI substudy, and review and approve the final manuscript.

Conflicts of Interest: None.

Name: Aida Tantri, MD.

Contribution: This author helped coordinate the ISOS study and collect ISOS data, conceive and plan the ISOS-AKI substudy, and review and approve the final manuscript.

Conflicts of Interest: None.

Name: Velayuthapillai Sivasakthi, MD.

Contribution: This author helped coordinate the ISOS study and collect ISOS data, conceive and plan the ISOS-AKI substudy, and review and approve the final manuscript.

Conflicts of Interest: None.

Name: Marzida Mansor, MD.

Contribution: This author helped coordinate the ISOS study and collect ISOS data, conceive and plan the ISOS-AKI substudy, and review and approve the final manuscript.

Conflicts of Interest: None.

Name: Ricardo Matos, MD.

Contribution: This author helped coordinate the ISOS study and collect ISOS data, conceive and plan the ISOS-AKI substudy, and review and approve the final manuscript.

Conflicts of Interest: None.

Name: Rupert M. Pearse, MD.

Contribution: This author helped coordinate the ISOS study and collect ISOS data, conceive and plan the ISOS-AKI substudy, review and approve the study design and statistical analysis plan, review the analysis and results, edit and revise the final draft, and review and approve the final manuscript.

Conflicts of Interest: R. M. Pearse holds research grants and has given lectures and/or performed consultancy work for Nestle Health Sciences, BBraun, Medtronic, Glaxo Smithkline, Intersurgical, and Edwards Lifesciences, and he is a member of the associate editorial board of the British Journal of Anaesthesia.

Name: John R. Prowle, MD.

Contribution: This authorhelped conceive and plan the ISOS-AKI substudy, undertake detailed study design and statistical analysis planning, perform primary statistical analysis, review the analysis and results, write the manuscript, prepare figures, edit and revise the final draft, and review and approve the final manuscript.

Conflicts of Interest: J. R. Prowle has given lectures and/or performed consultancy work for Nikkiso Europe GmbH and Baxter Inc.

This manuscript was handled by: Alexander Zarbock, MD.

Back to Top | Article Outline

REFERENCES

1. Weiser TG, Haynes AB, Molina GEstimate of the global volume of surgery in 2012: an assessment supporting improved health outcomes. Lancet. 2015;385suppl 2S11.
2. Kable AK, Gibberd RW, Spigelman ADAdverse events in surgical patients in Australia. Int J Qual Health Care. 2002;14:269–276.
3. Mullen R, Scollay JM, Hecht G, McPhillips G, Thompson AMDeath within 48 h–adverse events after general surgical procedures. Surgeon. 2012;10:1–5.
4. Pearse RM, Moreno RP, Bauer P, et alEuropean Surgical Outcomes Study (EuSOS) Group for the Trials Groups of the European Society of Intensive Care Medicine and the European Society of Anaesthesiology. Mortality after surgery in Europe: a 7 day cohort study. Lancet. 2012;380:1059–1065.
5. Weiser TG, Regenbogen SE, Thompson KDAn estimation of the global volume of surgery: a modelling strategy based on available data. Lancet. 2008;372:139–144.
6. International Surgical Outcomes Study Group. Global patient outcomes after elective surgery: prospective cohort study in 27 low-, middle- and high-income countries. Br J Anaesth. 2016;117:601–609.
7. O’Connor ME, Kirwan CJ, Pearse RM, Prowle JRIncidence and associations of acute kidney injury after major abdominal surgery. Intensive Care Med. 2016;42:521–530.
8. Bihorac A, Yavas S, Subbiah SLong-term risk of mortality and acute kidney injury during hospitalization after major surgery. Ann Surg. 2009;249:851–858.
9. Grams ME, Sang Y, Coresh JAcute kidney injury after major surgery: a retrospective analysis of Veterans Health Administration Data. Am J Kidney Dis. 2016;67:872–880.
10. Korenkevych D, Ozrazgat-Baslanti T, Thottakkara PThe pattern of longitudinal change in serum creatinine and 90-day mortality after major surgery. Ann Surg. 2016;263:1219–1227.
11. Kork F, Balzer F, Spies CDMinor postoperative increases of creatinine are associated with higher mortality and longer hospital length of stay in surgical patients. Anesthesiology. 2015;123:1301–1311.
12. O’Connor ME, Hewson RW, Kirwan CJ, Ackland GL, Pearse RM, Prowle JRAcute kidney injury and mortality 1 year after major non-cardiac surgery. Br J Surg. 2017;104:868–876.
13. Kheterpal S, Tremper KK, Heung MDevelopment and validation of an acute kidney injury risk index for patients undergoing general surgery: results from a national data set. Anesthesiology. 2009;110:505–515.
14. Prowle JR, Kam EP, Ahmad T, Smith NC, Protopapa K, Pearse RMPreoperative renal dysfunction and mortality after non-cardiac surgery. Br J Surg. 2016;103:1316–1325.
15. Gaber AO, Moore LW, Aloia TACross-sectional and case-control analyses of the association of kidney function staging with adverse postoperative outcomes in general and vascular surgery. Ann Surg. 2013;258:169–177.
16. Mases A, Sabaté S, Guilera N, et alANESCARDIOCAT Group. Preoperative estimated glomerular filtration rate and the risk of major adverse cardiovascular and cerebrovascular events in non-cardiac surgery. Br J Anaesth. 2014;113:644–651.
17. Cloyd JM, Ma Y, Morton JM, Kurella Tamura M, Poultsides GA, Visser BCDoes chronic kidney disease affect outcomes after major abdominal surgery? Results from the National Surgical Quality Improvement Program. J Gastrointest Surg. 2014;18:605–612.
18. Kahan BC, Koulenti D, Arvaniti K, et alInternational Surgical Outcomes Study (ISOS) Group. Critical care admission following elective surgery was not associated with survival benefit: prospective analysis of data from 27 countries. Intensive Care Med. 2017;43:971–979.
19. Ahmad T, Bouwman RA, Grigoras I, et alUse of failure-to-rescue to identify international variation in postoperative care in low-, middle- and high-income countries: a 7-day cohort study of elective surgery. Br J Anaesth. 2017;119:258–266.
20. Kidney Disease: Improving Global Outcomes (KDIGO). Section 2: AKI definition. Kidney Int Suppl (2011). 2012;2:19–36.
21. Levey AS, Stevens LA, Schmid CH, et alCKD-EPI (Chronic Kidney Disease Epidemiology Collaboration). A new equation to estimate glomerular filtration rate. Ann Intern Med. 2009;150:604–612.
22. von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JPSTROBE Initiative. Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. BMJ. 2007;335:806–808.
23. Sauerbrei WThe use of resampling methods to simplify regression models in medical statistics. J R Stat Soc Ser C Appl Stat. 1999;48:313–329.
24. Prowle JR, Kam EP, Ahmad T, Smith NC, Protopapa K, Pearse RMPreoperative renal dysfunction and mortality after non-cardiac surgery. Br J Surg. 2016;103:1316–1325.
25. Blitz JD, Shoham MH, Fang YPreoperative renal insufficiency: underreporting and association with readmission and major postoperative morbidity in an academic medical center. Anesth Analg. 2016;123:1500–1515.
26. Kooman JP, Kotanko P, Schols AM, Shiels PG, Stenvinkel PChronic kidney disease and premature ageing. Nat Rev Nephrol. 2014;10:732–742.
27. Grams ME, Sang Y, Coresh JCandidate surrogate end points for ESRD after AKI. J Am Soc Nephrol. 2016;27:2851–2859.
28. Doi K, Rabb HImpact of acute kidney injury on distant organ function: recent findings and potential therapeutic targets. Kidney Int. 2016;89:555–564.
29. Hill ABThe environment and disease: association or causation? Proc R Soc Med. 1965;58:295–300.

Supplemental Digital Content

Back to Top | Article Outline
© 2019 International Anesthesia Research Society