Abstract
Background/Aims
Acute kidney injury (AKI) following cardiac surgery is a complication associated with high rates of morbidity and mortality. We compared staging systems for the diagnosis of AKI after cardiac surgery, and assessed preoperative factors predictive of post-operative AKI.
Methods
Clinical data, surgical risk scores, procedure and clinical outcome were obtained on all 4,651 patients undergoing cardiac surgery to the Royal Infirmary of Edinburgh between April 2006 and March 2011, of whom 4,572 had sufficient measurements of creatinine before and after surgery to permit inclusion and analysis. The presence of AKI was assessed using the AKIN and RIFLE criteria.
Results
By AKIN criteria, 12.4% of the studied population developed AKI versus 6.5% by RIFLE criteria. Any post-operation AKI was associated with increased mortality from 2.2 to 13.5% (relative risk 7.0, p < 0.001), and increased inpatient stay from a median of 7 (IQR 4) to 9 (IQR 11) days (p < 0.05). Patients identified by AKIN, but not RIFLE, had a mean peak creatinine rise of 34% from baseline and had a significantly lower mortality com pared to RIFLE-’Risk’ AKI (mortality 6.1 vs. 9.7%; p < 0.05). Pre-operative creatinine, diabetes, NYHA Class IV dyspnoea and EuroSCORE-1 (a surgical risk score) all predicted subsequent AKI on multivariate analysis. EuroSCORE-1 outperformed any single demographic factor in predicting postoperative AKI risk, equating to an 8% increase in relative risk for each additional point.
Conclusion
AKI after cardiac surgery is associated with delayed discharge and high mortality rates. The AKIN and RIFLE criteria identify patients at a range of AKI severity levels suitable for trial recruitment. The utility of EuroSCORE as a risk stratification tool to identify high AKI-risk subjects for prospective intervention merits further study.
Keywords: Acute kidney injury, AKIN score, Cardiac surgery, EuroSCORE, RIFLE criteria
Introduction
Despite advances in percutaneous techniques, cardiac surgery remains the standard of care for the revascularisation of the most severe patterns of coronary artery disease [1,2]. Acute kidney injury (AKI) is a common complication of such surgery, and is associated with markedly increased rates of morbidity and mortality even after adjustment for pre-operative co-morbidities [3]. In addition to its impact on patient outcomes, AKI has marked health-economic implications, with extended inpatient stays and increases in hospital costs [4].
Despite the publication of two consensus statements from the Acute Dialysis Quality Initiative (RIFLE - Risk/ Injury/Failure/Loss/End-Stage) and Acute Kidney Injury Network (AKIN) consortia [5–7] offering standardised criteria for the diagnosis of AKI, there continues to be wide heterogeneity in the reported incidence of AKI after cardiac surgery ranging from as little as 1.2 to 39% [8–11] depending on the diagnostic criteria used. An understanding of incidence rates and outcomes is of key importance to gauge the success of current clinical practices, but also to facilitate appropriate powering of studies designed to investigate novel therapies and interventions targeting AKI prevention and treatment.
This work evaluated the incidence of AKI in a contemporary tertiary cardiac surgical unit using the AKIN and RIFLE staging criteria, assessing the impact of a diagnosis of AKI by either criteria on patient mortality and duration of inpatient stay. Patient factors present pre-operatively that predicted AKI on univariate and multivariate analysis were identified to facilitate the prospective identification of ‘high AKI-risk’ patients for recruitment into future clinical studies of AKI prophylaxis.
Subjects and Methods
Study Population
We identified 4,651 consecutive patients undergoing cardiac surgery in a 5-year period at the Royal Infirmary, Edinburgh, United Kingdom, between April 1, 2006 and March 31, 2011. A range of pre-operative, peri-operative and outcome details for each case had been prospectively recorded as part of unit contributions to an ongoing national audit of cardiac surgical outcomes, with appropriate hospital research ethics committee permissions in place. Follow-up was complete for all participants to time of death or hospital discharge.
Clinical Characteristics and Outcome
Clinical characteristics, cardiovascular risk factors and the urgency and nature of the surgery undertaken were documented. A pre-operative additive and logistic EuroSCORE-1 [12], a validated tool for the prediction of operative mortality, was available for each patient (see table 1 for breakdown of EuroSCORE-1). Preoperative and post-operative creatinine values were obtained through the TrakCare software application (InterSystems Corporation, Cambridge, Mass., USA), an electronic patient record system used by the Acute Hospitals Division of Lothian National Health Service Health Board, Scotland. The absolute creatinine rise in pmol/l and percentage change from baseline in the 48 h after surgery were calculated and used to establish the presence and severity of AKI, using the criteria proposed by AKIN [6] as follows:
Stage 1: increase in serum creatinine ≥26.5 μmol/l (≥0.3 mg/ dl) or increase to ≥150-199% (1.5- to 1.9-fold) from baseline
Stage 2: increase in serum creatinine to 200-299% (>2- to 2.9- fold) from baseline
Stage 3: increase in serum creatinine to ≥300% (>3-fold) from baseline or serum creatinine 354 µmol/l (≥4.0 mg/dl) with an acute rise of at least 44 µmol/l (0.5 mg/dl) or initiation of renal replacement therapy (RRT).
Table 1. The additive EuroSCORE-1 scoring system.
| Criteria | Score | |
|---|---|---|
| Patient-related factors | ||
| Age | (per 5 years or part thereof over 60 years) | 1 |
| Sex | female | 1 |
| Chronic pulmonary disease | long-term use of bronchodilators or steroids for lung disease | 1 |
| Extracardiac arteriopathy | any one or more of the following: claudication, carotid occlusion or >50% stenosis, previous or planned intervention on the abdominal aorta, limb arteries or carotids | 2 |
| Neurological dysfunction | severely affecting ambulation or day-to-day functioning | 2 |
| Previous cardiac surgery | requiring opening of the pericardium | 3 |
| Serum creatinine | >2.3 mg/dl pre-operatively | 2 |
| Active endocarditis | patient still under antibiotic treatment for endocarditis at the time of surgery | 3 |
| Critical pre-operative state | any one or more of the following: ventricular tachycardia or fibrillation or aborted sudden death, pre-operative cardiac massage, pre-operative ventilation before arrival in the anaesthetic room, pre-operative inotropic support, intra-aortic balloon counter-pulsation or pre-operative acute renal failure (anuria or oliguria <10 ml/h) | 3 |
| Cardiac-related factors | ||
| Unstable angina | rest angina requiring i.v. nitrates until arrival in the anaesthetic room | 2 |
| LV dysfunction | moderate or LV ejection fraction 30 - 50% | 1 |
| poor or LV ejection fraction <30 | 3 | |
| Recent myocardial infarction | (<90 days) | 2 |
| Pulmonary hypertension | systolic PA pressure >60 mm Hg | 2 |
| Operation-related factors | ||
| Emergency | carried out on referral before the beginning of the next working day | 2 |
| Other than isolated CABG | major cardiac procedure other than or in addition to CABG | 2 |
| Surgery on thoracic aorta | for disorder of ascending, arch or descending aorta | 3 |
| Post-infarction septal rupture | 4 |
Patients are assessed and scored on pre-operative factors to generate a total additive EuroSCORE-1 [12]. Online calculators can be accessed at http://www.euroscore.org/calcold.html. Reproduced with permission of Sam Nashef. CABG = Coronary artery bypass grafting; PA = pulmonary artery.
These were compared to the RIFLE criteria for the diagnosis of early AKI [5], where the ‘Risk’ definition is based on a creatinine increase to ≥150-199% (1.5- to 1.9-fold) from baseline. Key outcome measures were diagnosis of new AKI, need for RRT, length of inpatient stay and inpatient death.
Statistical Analysis
Data were analysed using SPSS Statistics version 19.0 (IBM, Armonk, N.Y., USA). Univariate analysis to identify predictors of adverse clinical outcome was performed using Pearson’s χ2 tests for categorical variables and a 2-sample t test for continuous variables (e.g. age, weight and creatinine). Variables found to be related to outcomes and those that indicated significance (at the 10% level) were included in a multivariate binary logistic regression model. Statistical significance was set at 2-sided p < 0.05, with Bonferroni correction for multiple comparisons.
Results
Baseline Characteristics of the Studied Population
Data were collected from 4,651 consecutive patients undergoing cardiac surgery at the Royal Infirmary of Edinburgh between April 1, 2006 and March 31, 2011, of whom 4,572 had the necessary data relating to baseline renal function and post-operative renal function in the 48 h after surgery. Data were available for all other variables studied in a minimum of 95% of the patients. Baseline pre-operative characteristics stratified by the development of post-operative AKI are summarised in table 2. This illustrates an overall incidence of AKI (AKIN stages 1-3) of 12.8% over the 5-year period studied, with 1.4% of all patients in the study requiring RRT.
Table 2. Characteristics of patient cohorts with and without post-operative AKI.
| No AKI (n = 3,985) | Any AKI (n = 587) | P | |||
|---|---|---|---|---|---|
| mean ± SD/count | column, % | mean ± SD/count | column, % | ||
| Age, years | 66.6 ± 11.0 | 68.7 ± 10.9 | <0.001 | ||
| Sex | 0.846 | ||||
| Male | 2,819 | 70.7 | 413 | 70.4 | |
| Pre-operative creatinine, pmol/l | 94.5 ± 31.5 | 117.7 ± 68.7 | <0.001 | ||
| eGFR group (MDRD) | <0.001 | ||||
| GFR <60 | 1,206 | 31 | 316 | 55.1 | |
| GFR 60+ | 2,687 | 69 | 257 | 44.9 | |
| NYHA | <0.001 | ||||
| I | 998 | 25 | 105 | 17.9 | |
| II | 1,005 | 25.2 | 128 | 21.8 | |
| III | 1,639 | 41.1 | 240 | 40.9 | |
| IV | 343 | 8.6 | 114 | 19.4 | |
| Previous MI | <0.001 | ||||
| None | 2,689 | 67.5 | 364 | 62 | |
| 1 | 1,031 | 25.9 | 157 | 26.7 | |
| 2 or more | 257 | 6.4 | 66 | 11.2 | |
| Unknown | 8 | 0.2 | 0 | 0 | |
| Last MI | 0.031 | ||||
| MI <6 h | 0 | 0.0 | 1 | 0.2 | |
| MI >90 days | 644 | 16.2 | 105 | 17.9 | |
| MI 1 - 30 days | 441 | 11.1 | 81 | 13.8 | |
| MI 31 - 90 days | 186 | 4.7 | 31 | 5.3 | |
| MI 6 - 24 h | 17 | 0.4 | 3 | 0.5 | |
| No previous MI | 2,696 | 67.7 | 366 | 62.4 | |
| LV ejection fraction | <0.001 | ||||
| Good (>50%) | 1,772 | 44.5 | 228 | 38.8 | |
| Fair (30 - 49%) | 646 | 16.2 | 107 | 18.2 | |
| Poor (<30%) | 179 | 4.5 | 50 | 8.5 | |
| Not measured | 1,378 | 34.6 | 202 | 34.4 | |
| Hypertension | <0.001 | ||||
| No hypertension | 1,023 | 25.7 | 90 | 15.3 | |
| Treated or BP >140/90 on 1 occasion | |||||
| before admission | 2,767 | 69.4 | 456 | 77.7 | |
| Claudication | 0.128 | ||||
| No | 3,472 | 87.4 | 498 | 85.1 | |
| Yes | 501 | 12.6 | 87 | 14.9 | |
| Carotid occlusion | 0.629 | ||||
| No | 3,873 | 98.2 | 575 | 98.5 | |
| Yes | 72 | 1.8 | 9 | 1.5 | |
| Diabetes mellitus | <0.001 | ||||
| Not diabetic | 3,251 | 81.6 | 422 | 71.9 | |
| Diabetic | 734 | 18.4 | 165 | 28.1 | |
| Weight, kg | 81.2 ± 21.5 | 83.5 ± 17.2 | 0.013 | ||
| Smoking | 0.984 | ||||
| Current smoker | 567 | 14.2 | 84 | 14.3 | |
| Ex-smoker | 2,025 | 50.8 | 300 | 51.1 | |
| Never smoked | 1,393 | 35.0 | 203 | 34.6 | |
| Cardiogenic shock | 0.02 | ||||
| No | 3,944 | 99.0 | 572 | 97.4 | |
| Yes | 41 | 1 | 15 | 2.6 | |
| Pre-operative IABP | <0.001 | ||||
| No | 3,813 | 95.7 | 533 | 90.8 | |
| Yes | 172 | 4.3 | 54 | 9.2 | |
| Cardiac procedure | <0.001 | ||||
| CABG | 2,300 | 57.7 | 276 | 47.0 | |
| CABG + other | 25 | 0.6 | 6 | 1.0 | |
| CABG + valve | 482 | 12.1 | 91 | 15.5 | |
| CABG + valve + other | 26 | 0.7 | 9 | 1.5 | |
| Other | 89 | 2.2 | 34 | 5.8 | |
| Valve | 933 | 23.4 | 157 | 26.7 | |
| Valve + other | 130 | 3.3 | 14 | 2.4 | |
| Additive EuroSCORE | 4.8 ± 3.1 | 6.5 ± 3.6 | <0.001 | ||
Summary of key pre-operative characteristics of subjects stratified by the presence of post-operative AKI. p values quoted were calculated using Pearson χ 2 tests and unpaired t tests with Bonferroni correction. CABG = Coronary artery bypass grafting; IABP = intraaortic balloon pump; MI = myocardial infarction.
Impact of AKI on Length of Inpatient Stay and Mortality following Elective and Emergency Cardiac Surgery
The impact of degree of AKI on length of inpatient stay and risk of mortality is summarised in table 3. This demonstrates that with as little as a 26.5-µmol/l absolute rise (and <100% relative rise) in peak post-operative serum creatinine (AKIN stage 1), there is an increase in length of inpatient stay (7 vs. 9 days; p < 0.05; no AKI vs. AKIN 1). Of note, mortality rose by >3-fold in AKIN 1 compared to patients without AKI (2.2 vs. 7.5%, No AKI vs. AKIN 1; p < 0.05), with a further rise to 12.5% with AKIN stage 2 (p < 0.05 vs. all groups). AKIN stage 3 (including patients requiring RRT) was associated with a >20-fold increased risk of death compared to those with no postoperative AKI (inpatient mortality 2.2 vs. 50%, no AKI vs. AKIN 3; p < 0.05).
Table 3. Patient outcomes stratified by AKIN classification of AKI.
| No AKI (n = 3,985) | AKIN stage 1 (n = 455) | AKIN stage 2 (n = 56) | AKIN stage 3 (n = 76) | |||||
|---|---|---|---|---|---|---|---|---|
| mean ± SD/count | column | mean ± SD/count | column | mean ± SD/count | column | mean ± SD/count | column | |
| Age, years | 66.6 ± 11.0 | 68.6 ± 11.0a | 67.7 ± 11.5 | 70.2 ± 9.6a | ||||
| Additive EuroSCORE | 4.9 ± 3.1 | 6.2 ± 3.6a | 7.3 ± 3.9a | 7.5 ± 3.7a, b | ||||
| Median length of stay, days | 7 (IQR 4) | 9 (IQR 8)c | 11 (IQR 13)c | 17 (IQR 34)c | ||||
| Died as inpatient | 87 | 2.2% | 34c | 7.5% | 7c | 12.5% | 38c | 50% |
AKIN stage 1: increase in serum creatinine ≥26.5 μmol/l (0.3 mg/dl) or increase to ≥150–199% from baseline. AKIN stage 2: increase in serum creatinine to 200–299% (>2- to 2.9-fold) from baseline. AKIN stage 3: increase in serum creatinine to ≥300% (≥3-fold) from baseline or serum creatinine ≥354 μmol/l (4.0 mg/dl) with an acute rise of at least 44.25 μmol/l (0.5 mg/dl) or initiation of RRT. Stepwise increases in median length of stay and inpatient mortality were noted with increasing AKIN grade.
p < 0.05 when compared to no AKI
p < 0.05 when compared to AKIN 1
p < 0.001 all groups.
Comparison of RIFLE and AKIN Criteria for Diagnosis of Early-Stage AKI
A key difference between the RIFLE and AKIN systems for the diagnosis of AKI relates to the additional criteria of a rise of ≥26.5 µmol/l in serum creatinine leading to the diagnosis of AKIN stage 1. We re-examined our dataset of early AKI (peak creatinine rise <100% over baseline) and classified patients as ‘no AKI’, AKIN stage 1a (creatinine rise of ≥26.5 µmol/l, but <50% baseline) and AKIN stage 1b/RIFLE-’Risk’ (peak creatinine rise ≥50%, <100% over baseline). Use of the AKIN criteria resulted in an additional 279 patients being diagnosed with AKIN stage 1 (total n = 455) compared to RIFLE-’Risk’ (n = 176), increasing ‘early AKI’ diagnoses from 3.8 to 10.0% of the study population. The characteristics of these ‘early’ AKI groups are summarised in table 4. Individuals who were RIFLE negative but AKIN 1a positive had a significantly lower risk of death than AKIN 1b/RIFLE-’Risk’ (mortality 9.7 vs. 6.1 vs. 2.2%, AKIN 1b vs. AKIN 1a vs. No AKI; p < 0.05 AKIN 1b vs. both groups).
Table 4. Comparison of early AKI diagnosis and patient outcome using AKIN and RIFLE criteria.
| No AKI (n = 3,985) | AKIN 1a/RIFLE negative (n = 279) | AKIN 1b/RIFLE-’Risk’ (n = 176) | ||||
|---|---|---|---|---|---|---|
| mean ± SD/count | column, % | mean ± SD/count | column, % | mean ± SD/count | column, % | |
| Age, years | 66.6 ± 11.0 | 68.6 ± 10.6a | 68.2 ± 11.7 | |||
| Pre-operative creatinine, pmol/l | 94.5 ± 31.5 | 127.1 ± 79.8b | 105.9 ± 44.4b | |||
| Peak percentage creatinine rise | −4.2 ± 18.7% | 34.1 ± 8.9%b | 65.7 ± 13.3%b | |||
| Median length of stay, days | 7 (IQR 4) | 9 (IQR 8)a | 9 (IQR 8)a | |||
| Died as inpatient | 87 | 2.2 | 17a | 6.1 | 17a | 9.7 |
Patients with peak creatinine rises <100% above pre-operative baseline were classified into groups stratified by positivity by both RIFLE-’Risk’ and AKIN stage 1 criteria (peak creatinine ≥50% above baseline), or those positive only by the extended AKIN criteria (increase in serum creatinine of 26.5 μmol/l, referred to as AKIN ‘1a’). Significant differences between groups on χ2 testing or comparison of column means are highlighted by superscript.
p < 0.05 vs. no AKI
p < 0.05 vs. all groups.
Factors Associated with the Development of AKI following Cardiac Surgery
Univariate analysis of pre-operative characteristics was undertaken to assess those associated with subsequent development of AKI. The EuroSCORE composite score of predicted surgical mortality risk was also analysed to assess its potential utility as a predictor of AKI risk (for components of the additive EuroSCORE see table 1). This demonstrated that patient factors such as increasing age, pre-existing CKD (identified by a baseline eGFR <60 ml/min), NYHA IV dyspnoea, LV ejection fraction <30%, diabetes mellitus, a history of hypertension and admission weight, and pre-operative factors such as the presence of cardiogenic shock, pre-operative intra-aortic balloon pump use and a procedure other than coronary artery bypass grafting alone were all associated with subsequent AKI (table 2).
Patients with AKI also had higher pre-operative EuroSCORES (additive EuroSCORE 4.8 ± 3.1 vs. 6.5 ± 3.6, no AKI vs. AKI; p < 0.05). Suffering any AKI was associated with a >5-fold increase in risk of death (inpatient mortality 2.2 vs. 13.5%, no AKI vs. AKI; p < 0.001).
Multivariate Analysis of Pre-Operative Factors Predicting Subsequent AKI
In order to assess the key factors in predicting AKI outcome after surgery, all pre-operative factors significant on univariate analysis were entered into a binary logistic regression model.
After multiple regression, pre-operative creatinine [relative risk (RR) of AKI 1.009 per µmol/l (95% CI: 1.0061-1.011), p < 0.001], diabetes [RR 1.533 (95% CI: 1.239-1.897), p < 0.001], NYHA IV dyspnoea [RR 1.873 (95% CI: 1.354-2.591), p < 0.001], weight [RR 1.005 per kg (95% CI: 1.001-1.010), p = 0.01], cardiac procedure other than valve or cardiac bypass surgery [RR 3.807 (95% CI: 1.7978.067), p < 0.001] and additive EuroSCORE-1 [RR 1.083 (95% CI: 1.036-1.132), p < 0.001] were significant predictors of the development of post-operative AKI. It is of note that EuroSCORE-1 achieved the most significant correlate on multivariate analysis, with each additional risk point equating to an 8% increase in RR of AKI. Furthermore, on receiver operating characteristic analysis of EuroSCORE-1 as a predictor for post-operative AKI, the area under the curve was 0.634 (95% CI: 0.609-0.658, p < 0.001) whilst the Hosmer-Lemeshow test value was 0.08.
Impact of AKI and RRT on Outcome
Increasing severity grade of AKI was associated with significant stepwise increases in mortality and in length of hospital stay (table 3; p < 0.05 between all groups for mortality, p < 0.05 AKIN 3 vs. AKIN 1 and p < 0.05 AKIN 1/2/3 vs. no AKI). The 62 patients with AKIN stage 3 who required RRT exhibited the highest mortality (54.8 vs. 28.6%, AKIN-3 + RRT vs. AKIN-3 no RRT; p < 0.001). Pre-operative factors and peri-operative factors associated with mortality were assessed using univariate analysis (table 5). Notably, many of the factors significantly associated with AKI risk also predicted risk of death and contributed to the score of the additive EuroSCORE-1.
Table 5. Characteristics of patient cohorts stratified by inpatient mortality.
| Alive (n = 4,406) | Died as inpatient (n = 166) | p | |||
|---|---|---|---|---|---|
| mean ± SD/count | column, % | mean ± SD/count | column, % | ||
| Pre-operative creatinine | 96.6 ± 37.5 | 120.5 ± 66.2 | <0.001 | ||
| Age, years | 66.8 ± 11 | 70.4 ± 11 | <0.001 | ||
| Sex | <0.05 | ||||
| Female | 1,273 | 28.9 | 67 | 40.4 | |
| Male | 3,133 | 71.1 | 99 | 59.6 | |
| Dyspnoea status | <0.001 | ||||
| NYHA I | 1,073 | 24.4 | 30 | 18.1 | |
| NYHA II | 1,110 | 25.2 | 23 | 13.9 | |
| NYHA III | 1,807 | 41 | 72 | 43.4 | |
| NYHA IV | 416 | 9.4 | 41 | 24.7 | |
| Previous MI | <0.05 | ||||
| None | 2,954 | 67 | 99 | 56.9 | |
| One | 1,141 | 25.9 | 47 | 28.3 | |
| Two or more | 304 | 6.9 | 19 | 11.4 | |
| Unknown | 7 | 0.2 | 1 | 0.6 | |
| Active endocarditis | <0.001 | ||||
| yes | 87 | 2 | 13 | 7.8 | |
| Last MI | <0.001 | ||||
| MI <6 h | 0 | 0 | 1 | 0.6 | |
| MI >90 days | 724 | 16.4 | 25 | 15.1 | |
| MI 1 – 30 days | 492 | 11.2 | 30 | 18.1 | |
| MI 31 – 90 days | 209 | 4.7 | 8 | 4.8 | |
| MI 6 – 24 h | 18 | 0.4 | 2 | 1.2 | |
| No previous MI | 2,962 | 67.2 | 100 | 60.2 | |
| Hypertension | <0.01 | ||||
| No hypertension | 1,084 | 24.6 | 29 | 17.5 | |
| Treated or BP >140/90 on >1 occasion | |||||
| prior to admission | 3,102 | 70.4 | 121 | 72.9 | |
| Unknown | 220 | 5 | 16 | 9.6 | |
| Smoking | 0.93 | ||||
| Current smoker | 626 | 14.2 | 25 | 15.1 | |
| Ex-smoker | 2,240 | 50.8 | 85 | 51.2 | |
| Never smoked | 1,540 | 35 | 56 | 33.7 | |
| Diabetic? | 0.143 | ||||
| Not diabetic | 3,547 | 80.5 | 126 | 75.9 | |
| Diabetic | 859 | 19.5 | 40 | 24.1 | |
| Carotid occlusion | 0.985 | ||||
| No | 4,285 | 98.2 | 163 | 98.2 | |
| Yes | 78 | 1.8 | 3 | 1.8 | |
| Claudication | <0.05 | ||||
| No | 3,836 | 87.3 | 134 | 80.7 | |
| Yes | 556 | 12.7 | 32 | 19.3 | |
| Additive EuroSCORE | 4.9 ± 3.1 | 9.0 ± 3.8 | <0.001 | ||
| Logistic EuroSCORE (% predicted mortality) | 5.9 ± 8.3 | 18.9 ± 18.2 | <0.001 | ||
| Cardiogenic shock | <0.001 | ||||
| No | 4,363 | 99 | 153 | 92.2 | |
| Yes | 43 | 1 | 13 | 7.8 | |
| Pre-operative IABP | <0.001 | ||||
| No | 4,206 | 95.5 | 200 | 84.5 | |
| Yes | 200 | 4.5 | 26 | 15.7 | |
| Cardiac procedure | <0.001 | ||||
| CABG | 2,525 | 57.3 | 51 | 30.7 | |
| CABG + other | 26 | 0.6 | 5 | 3 | |
| CABG + valve | 540 | 12.3 | 33 | 19.9 | |
| CABG + valve + other | 28 | 0.6 | 7 | 4.2 | |
| Other | 100 | 2.3 | 23 | 13.9 | |
| Valve | 1,055 | 23.9 | 35 | 21.1 | |
| Valve + other | 132 | 3.0 | 12 | 7.2 | |
| Operative priority | <0.001 | ||||
| Elective | 3,288 | 74.6 | 89 | 53.6 | |
| Emergency | 132 | 3 | 24 | 14.5 | |
| Salvage | 2 | 0 | 3 | 1.8 | |
| Urgent | 984 | 22.3 | 50 | 30.1 | |
| Bypass time, min | 109 ± 48 | 160.7 ± 92.7 | <0.001 | ||
| Cross-clamp time, min | 79.3 ± 89.5 | 97.4 ± 54.8 | <0.05 | ||
| Bloods transfused total (red cell concentrate units) | 2.1 ± 38.7 | 4.4 ± 5.8 | 0.45 | ||
| Post-operative AKI | <0.001 | ||||
| No | 3,906 | 88.7 | 97 | 58.4 | |
| Yes | 500 | 11.3 | 69 | 41.6 | |
Summary of key pre-operative and peri-operative characteristics of subjects stratified by post-operative survival vs. mortality. p values quoted were calculated using Pearson χ2 tests and unpaired t tests with Bonferroni correction. CABG = Coronary artery bypass grafting; IABP = intra-aortic balloon pump; MI = myocardial infarction.
Multivariate analysis of all factors associated with inpatient mortality on univariate analysis was undertaken. Only EuroSCORE-1 [adjusted RR 1.39 per point (95% CI: 1.14-1.72), p = 0.001], baseline creatinine [RR 1.004 per µmol/l (95% CI: 1.000-1.008), p < 0.05], any post-operative AKI [RR 2.11 (95% CI: 1.27-3.53), p < 0.005] bypass time [RR 1.019 per min (95% CI: 1.013-1.025), p < 0.001] and requirement for RRT [RR 14.98 (95% CI: 6.54-34.28), p < 0.001] remained significant independent predictors of inpatient mortality.
Discussion
These data demonstrate that AKI remains a common sequela of modern cardiac surgical practice and is associated with significant morbidity and mortality. Whilst the majority of AKI resolves without the requirement for RRT, even in its mildest form (AKIN stage 1) it is associated with a >3-fold increase in mortality rates, rising to greater than 50% mortality when renal replacement was required. The pre-operative serum creatinine level, a diagnosis of diabetes, NHYA stage IV dyspnoea, weight and cardiac procedure other than coronary bypass or valve surgery all predicted risk of subsequent AKI on multivariate analysis, with the composite pre-operative risk score additive EuroSCORE-1 offering the strongest correlation with risk of subsequent AKI.
Our data highlights the impact that diagnostic criteria for AKI have on the reported incidence and outcome of AKI. Until the last decade there was a lack of any true consensus on the definition of AKI, short of the absolute requirement for RRT. This has contributed to widespread variation in the reported incidence of AKI after cardiac surgery, with recent reports ranging from 4.3% (using a definition of >100% creatinine increase) [13] to as high as 39% [8] (using a definition of a rise of ≥0.3 mg/dl in creatinine). Despite the wide variation in reported total AKI incidence between our data (12.8%) and the study of Englberger et al. [13] (4.3%), it is noteworthy that rates of AKI requiring RRT were similar at 1.4 and 1.7%, respectively.
The advent of the RIFLE and AKIN criteria [5,6] have led to greater uniformity in the diagnosis of early AKI; however, as we demonstrate, at their lowest diagnostic threshold, AKIN scoring more than doubles the incidence of ‘early’ AKI compared to RIFLE criteria. Of particular interest are those patients fulfilling only the ‘mildest’ of AKIN stage 1 entry criteria and who are negative by RIFLE assessment (here referred to as AKIN stage 1a). These patients have a mortality rate that was not statistically different from those without AKI in the group sizes studied, and less than half that of those patients positive by both the RIFLE and AKIN definitions, calling into question the clinical significance of the finding in these patients.
Whilst it has been suggested that this subgroup of patients represents a misclassification of AKI [13], we would propose it is of diagnostic importance for two reasons. Firstly, the comparatively low mortality rates in ‘no AKI’ and ‘AKIN stage 1a’ mean that our study population size did not deliver power to detect a significant difference in mortality at the 80% confidence level. If confirmed in further studies, the percentage difference between the two (204% higher relative mortality in AKIN stage 1a) would be of clinical importance, and similar work in larger cohorts has demonstrated significant mortality differences at these or similar levels in both cardiac surgery [13] and in general hospitalized patients [14]. Secondly, this phase of early AKI, with low associated mortality, represents a more desirable point for therapeutic intervention with the aim of preventing progression and the attendant mortality burden of higher AKI stages.
Cardiac surgery, and its unavoidable haemodynamic stresses on the perfusion of the kidneys, represents a prototype of a predicable acute renal injury. Attention has been directed to the utility of novel urinary biomarkers such as neutrophil gelatinase-associated lipocalin and kidney injury molecule-1 in this setting as tools to permit the early identification of significant AKI [15–17]. Whilst these assays allow earlier identification of post-operative AKI [17], it is unproven whether this is of solely prognostic interest or can translate into improved outcomes. It may be that identifying patients at elevated AKI risk before surgery represents the most realistic opportunity to improve patient outcome by protecting the kidney via prophylactic interventions [18]. It is worthwhile acknowledging that whilst many pre-emptive interventions have shown efficacy in animal models of AKI [19–21], attempts to improve outcome by translating interventions such as diuretic or dopamine infusion to clinical practice have failed to replicate their success in rodent models [22]. A recent retrospective registry study has suggested that use of pre-operative statins was associated with a reduced risk of post-surgical AKI after multivariate analysis of risk factors [23]. Whilst this finding echoes reports of statin-induced protection in rodent models of AKI [24], whether this effect will persist in prospective studies remains to be established, as does the underlying mechanism in man.
A number of AKI risk estimation systems have been proposed based on pre-operative and peri-operative patient factors. The Cleveland Score [10,25] and Society of Thoracic Surgeons (STS) Bedside Risk Tool [26] have been developed specifically for the prediction of AKI in cardiac surgery patients, and subsequently externally validated. However, their primary outcome is the need for RRT rather than AKI by the AKIN or RIFLE criteria. As the clinical importance of even mild post-operative AKI in this patient cohort becomes apparent, so does the need for identifying those patients at risk. Models do exist to predict non-RRT requiring AKI [8,27,28], but similarly these aim to identify patients at risk of ‘severe’ AKI - the definition of which varies from a serum creatinine >2 mg/ dl (with an increase of at least 0.7 mg/dl from baseline) or RRT, to a post-operative eGFR <30 ml/min/1.73 m2. None use the AKIN or RIFLE definitions for AKI, thereby again potentially underestimating those who could be identified as being at risk of AKI.
We have demonstrated that, in addition to its primary application in mortality risk prediction, increasing points on the additive EuroSCORE are predictive of subsequent AKI risk. The ease of calculation (and availability of online calculators) coupled with the large cohorts of retrospective EuroSCORE data in registries adds to its potential utility for comparing AKI incidence across centres and the focused recruitment of ‘high AKI-risk’ patients into prospective clinical studies, given the score can be calculated in the outpatient setting prior to elective surgery.
Whilst scoring systems such as the additive EuroSCORE-1 can allow some estimation of future AKI risk, regrettably there is currently no clinically validated intervention that can be applied to those patients to prevent progression to higher stage AKI, requirement for RRT or mortality. Accordingly, a major driver for better predictors of AKI risk should be to identify suitable patients for further clinical trials aimed at improving outcome for this devastating clinical problem.
Acknowledgements
The authors would like to express thanks to Mr. Sam Nashef of Papworth Hospital, Cambridge, for his permission to reproduce the scoring criteria for the additive EuroSCORE-1.
Footnotes
Disclosure Statement
D.A.F. was funded by a Clinical Lecturer’s startup grant from the Academy of Medical Sciences and the Wellcome Trust.
References
- 1.Kushner FG, Hand M, Smith SC, Jr, King SB, 3rd, Anderson JL, Antman EM, Bailey SR, Bates ER, Blankenship JC, Casey DE, Jr, Green LA, et al. 2009 focused updates: ACC/AHA guidelines for the management of patients with ST-elevation myocardial infarction (updating the 2004 guideline and 2007 focused update) and ACC/AHA/SCAI guidelines on percutaneous coronary intervention (updating the 2005 guideline and 2007 focused update): a report of the American College of Cardiology Foun-dation/American Heart Association Task Force on Practice Guidelines. Circulation. 2009;120:2271–2306. doi: 10.1161/CIRCULATIONAHA.109.192663. [DOI] [PubMed] [Google Scholar]
- 2.Wijns W, Kolh P, Danchin N, Di Mario C, Falk V, Folliguet T, Garg S, Huber K, James S, Knuuti J, Lopez-Sendon J, et al. Guidelines on myocardial revascularization. Eur Heart J. 2010;31:2501–2555. doi: 10.1093/eurheartj/ehq277. [DOI] [PubMed] [Google Scholar]
- 3.Chertow GM, Levy EM, Hammermeister KE, Grover F, Daley J. Independent association between acute renal failure and mortality following cardiac surgery. Am J Med. 1998;104:343–348. doi: 10.1016/s0002-9343(98)00058-8. [DOI] [PubMed] [Google Scholar]
- 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;16:3365–3370. doi: 10.1681/ASN.2004090740. [DOI] [PubMed] [Google Scholar]
- 5.Bellomo R, Ronco C, Kellum JA, Mehta RL, Palevsky P. Acute 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–R212. doi: 10.1186/cc2872. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Mehta RL, Kellum JA, Shah SV, Molitoris BA, Ronco C, Warnock DG, Levin A. Acute kidney injury network: report of an initiative to improve outcomes in acute kidney injury. Crit Care. 2007;11:R31. doi: 10.1186/cc5713. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Bagshaw SM, George C, Bellomo R. A comparison of the RIFLE and AKIN criteria for acute kidney injury in critically ill patients. Nephrol Dial Transplant. 2008;23:1569–1574. doi: 10.1093/ndt/gfn009. [DOI] [PubMed] [Google Scholar]
- 8.Brown JR, Kramer RS, MacKenzie TA, Coca SG, Sint K, Parikh CR. Determinants of acute kidney injury duration after cardiac surgery: an externally validated tool. Ann Thorac Surg. 2012;93:570–576. doi: 10.1016/j.athoracsur.2011.11.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Hashemzadeh K, Hashemzadeh S, Dehdilani M. Risk factors and outcomes of acute renal failure after open cardiac surgery. Asian Car-diovasc Thorac Ann. 2012;20:275–280. doi: 10.1177/0218492312436402. [DOI] [PubMed] [Google Scholar]
- 10.Thakar CV, Arrigain S, Worley S, Yared JP, Paganini EP. A clinical score to predict acute renal failure after cardiac surgery. J Am Soc Nephrol. 2005;16:162–168. doi: 10.1681/ASN.2004040331. [DOI] [PubMed] [Google Scholar]
- 11.Karkouti K, Wijeysundera DN, Yau TM, Callum JL, Cheng DC, Crowther M, Dupuis JY, Fremes SE, Kent B, Laflamme C, Lamy A, et al. Acute kidney injury after cardiac surgery: focus on modifiable risk factors. Circulation. 2009;119:495–502. doi: 10.1161/CIRCULATIONAHA.108.786913. [DOI] [PubMed] [Google Scholar]
- 12.Nashef SA, Roques F, Michel P, Gauducheau E, Lemeshow S, Salamon R. European System for Cardiac Operative Risk Evaluation (EuroSCORE) Eur J Cardiothorac Surg. 1999;16:9–13. doi: 10.1016/s1010-7940(99)00134-7. [DOI] [PubMed] [Google Scholar]
- 13.Englberger L, Suri RM, Li Z, Casey ET, Daly RC, Dearani JA, Schaff HV. Clinical accuracy of RIFLE and Acute Kidney Injury Network (AKIN) criteria for acute kidney injury in patients undergoing cardiac surgery. Crit Care. 2011;15:R16. doi: 10.1186/cc9960. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Coca SG, Peixoto AJ, Garg AX, Krumholz HM, Parikh CR. The prognostic importance of a small acute decrement in kidney function in hospitalized patients: a systematic review and meta-analysis. Am J Kidney Dis. 2007;50:712–720. doi: 10.1053/j.ajkd.2007.07.018. [DOI] [PubMed] [Google Scholar]
- 15.Bennett M, Dent CL, Ma Q, Dastrala S, Grenier F, Workman R, Syed H, Ali S, Barasch J, Devarajan P. Urine NGAL predicts severity of acute kidney injury after cardiac surgery: a prospective study. Clin J Am Soc Nephrol. 2008;3:665–673. doi: 10.2215/CJN.04010907. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Mishra J, Dent C, Tarabishi R, Mitsnefes MM, Ma Q, Kelly C, Ruff SM, Zahedi K, Shao M, Bean J, Mori K, et al. Neutrophil gelatinase-associated lipocalin (NGAL) as a biomarker for acute renal injury after cardiac surgery. Lancet. 2005;365:1231–1238. doi: 10.1016/S0140-6736(05)74811-X. [DOI] [PubMed] [Google Scholar]
- 17.Han WK, Wagener G, Zhu Y, Wang S, Lee HT. Urinary biomarkers in the early detection of acute kidney injury after cardiac surgery. Clin J Am Soc Nephrol. 2009;4:873–882. doi: 10.2215/CJN.04810908. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Okusa MD, Molitoris BA, Palevsky PM, Chinchilli VM, Liu KD, Cheung AK, Weisbord SD, Faubel S, Kellum JA, Wald R, Chertow GM, et al. Design of clinical trials in acute kidney injury: a report from an NIDDK workshop -prevention trials. Clin J Am Soc Nephrol. 2012;7:851–855. doi: 10.2215/CJN.12811211. [DOI] [PubMed] [Google Scholar]
- 19.Goldfarb M, Rosenberger C, Ahuva S, Rosen S, Heyman SN. A role for erythropoietin in the attenuation of radiocontrast-induced acute renal failure in rats. Ren Fail. 2006;28:345–350. doi: 10.1080/08860220600591420. [DOI] [PubMed] [Google Scholar]
- 20.Heyman SN, Brezis M, Greenfeld Z, Rosen S. Protective role of furosemide and saline in radiocontrast-induced acute renal failure in the rat. Am J Kidney Dis. 1989;14:377–385. doi: 10.1016/s0272-6386(89)80171-4. [DOI] [PubMed] [Google Scholar]
- 21.Denton MD, Chertow GM, Brady HR. ‘Renaldose’ dopamine for the treatment of acute renal failure: scientific rationale, experimental studies and clinical trials. Kidney Int. 1996;50:4–14. doi: 10.1038/ki.1996.280. [DOI] [PubMed] [Google Scholar]
- 22.Lassnigg A, Donner E, Grubhofer G, Presterl E, Druml W, Hiesmayr M. Lack of renopro-tective effects of dopamine and furosemide during cardiac surgery. J Am Soc Nephrol. 2000;11:97–104. doi: 10.1681/ASN.V11197. [DOI] [PubMed] [Google Scholar]
- 23.Molnar AO, Coca SG, Devereaux PJ, Jain AK, Kitchlu A, Luo J, Parikh CR, Paterson JM, Siddiqui N, Wald R, Walsh M, et al. Statin use associates with a lower incidence of acute kidney injury after major elective surgery. J Am Soc Nephrol. 2011;22:939–946. doi: 10.1681/ASN.2010050442. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Gueler F, Park JK, Rong S, Kirsch T, Lindschau C, Zheng W, Elger M, Fiebeler A, Fliser D, Luft FC, Haller H. Statins attenuate ischemia-reperfusion injury by inducing heme oxygenase-1 in infiltrating macrophages. Am J Pathol. 2007;170:1192–1199. doi: 10.2353/ajpath.2007.060782. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Demirjian S, Schold JD, Navia J, Mastracci TM, Paganini EP, Yared JP, Bashour CA. Predictive models for acute kidney injury following cardiac surgery. Am J Kidney Dis. 2012;59:382–389. doi: 10.1053/j.ajkd.2011.10.046. [DOI] [PubMed] [Google Scholar]
- 26.Mehta RH, Grab JD, O’Brien SM, Bridges CR, Gammie JD, Haan CK, Ferguson TB, Peterson FD. Bedside tool for predicting the risk of postoperative dialysis in patients undergoing cardiac surgery. Circulation. 2006;114:2208–2216. doi: 10.1161/CIRCULATIONAHA.106.635573. [DOI] [PubMed] [Google Scholar]
- 27.Aronson S, Fontes ML, Miao Y, Mangano DT. Risk Index for perioperative renal dysfunc-tion/failure: critical dependence on pulse pressure hypertension. Circulation. 2007;115:733–742. doi: 10.1161/CIRCULATIONAHA.106.623538. [DOI] [PubMed] [Google Scholar]
- 28.Palomba H, de Castro I, Neto ALC, Lage S, Yu L. Acute kidney injury prediction following elective cardiac surgery: AKICS Score. Kidney Int. 2007;72:624–631. doi: 10.1038/sj.ki.5002419. [DOI] [PubMed] [Google Scholar]