Read ‘ Acute kidney injury; Challenges and opportunities’ article then the two additional articles ‘Acute kidney injury after cardiac surgery’ and ‘Optimal blood pressure

Journal of Critical Care (2013) 28, 389–396

Acute kidney injury after cardiac surgery according to Risk/Injury/Failure/Loss/End-stage, Acute Kidney Injury Network, and Kidney Disease: Improving Global Outcomes classifications☆,☆☆

Anthony J. Bastin MRCP, PhDa, Marlies Ostermann MD, PhDb,⁎, Andrew J. Slack MRCPa, Gerhard-Paul Diller MD, PhD c, Simon J. Finney MRCP, FRCA, PhDa, Timothy W. Evans MD, PhDa

aUnit of Critical Care, Imperial College, Royal Brompton Hospital, London, SW3 6NP, UK bDepartment of Nephrology and Critical Care, King's College London, Guys and St Thomas' Hospital, SE1 7EH, London, UK cAdult Congenital Heart Centre and Centre for Pulmonary Hypertension, Imperial College, Royal Brompton Hospital, SW3 6NP, London, UK

R b

0 h

Keywords: Acute kidney injury; Cardiac surgery; Renal replacement therapy; Intensive care unit; Cardiopulmonary bypass

Abstract Purpose: The epidemiology of acute kidney injury (AKI) after cardiac surgery depends on the definition used. Our aims were to evaluate the Risk/Injury/Failure/Loss/End-stage (RIFLE) criteria, the AKI Network (AKIN) classification, and the Kidney Disease: Improving Global Outcomes (KDIGO) classification for AKI post–cardiac surgery and to compare the outcome of patients on renal replacement therapy (RRT) with historical data. Methods: Retrospective analysis of 1881 adults who had cardiac surgery betweenMay 2006 and April 2008 and determination of the maximum AKI stage according to the AKIN, RIFLE, and KDIGO classifications. Results: The incidence of AKI using the AKIN and RIFLE criteria was 25.9% and 24.9%, respectively, but individual patients were classified differently. The area under the receiver operating characteristic curve for hospital mortality was significantly higher using the AKIN compared with the RIFLE criteria (0.86 vs 0.78, P = .0009). Incidence and outcome of AKI according to the AKIN and KDIGO classification were identical. The percentage of patients who received RRT was 6.2% compared with 2.7% in 1989 to 1990. The associated hospital mortality fell from 82.9% in 1989 to 1990 to 15.6% in 2006 to 2008.

☆ Sources of funding: A.J.B. was funded by the Dunhill Medical Trust, and this project was supported by the National Institute for Health Research (NIHR) espiratory Disease Biomedical Research Unit at Royal Brompton and Harefield NHS Foundation Trust and Imperial College London. The funding/supporting odies had no role in the design, collection, analysis, or interpretation of data or decision to submit for publication.

☆☆ Conflict of interest statement: None to declare. ⁎ Corresponding author. E-mail addresses: [email protected], [email protected] (M. Ostermann).

883-9441/$ – see front matter © 2013 Elsevier Inc. All rights reserved. ttp://dx.doi.org/10.1016/j.jcrc.2012.12.008

390 A.J. Bastin et al.

Conclusions: The AKIN classification correlated better with mortality than did the RIFLE criteria. Mortality of patients needing RRT after cardiac surgery has improved significantly during the last 20 years. © 2013 Elsevier Inc. All rights reserved.

1. Introduction

Acute kidney injury (AKI) after cardiac surgery is associated with increased mortality [1–4], a higher incidence of complications, a longer stay in the intensive care unit (ICU) and hospital, and increased health care costs [2,5–7]. Moreover, the highest mortality and complications are seen in patients who require renal replacement therapy (RRT). The reported incidence of AKI after cardiac surgery varies widely depending on the definition used.

In AKI in general, the most commonly used definitions are the Risk/Injury/Failure/Loss/End-stage (RIFLE) criteria, which differentiate between 3 stages (Risk, Injury, Failure) and 2 outcome categories (Loss and End-stage renal disease), and the AKI Network (AKIN) classification, which essen- tially is a modified version of the RIFLE criteria [8,9] (Table 1). The AKIN classification differs from the RIFLE

Table 1 RIFLE, AKIN, and KDIGO classifications for AKI

Serum creatinine criteria a

RIFLE [10] RIFLE-Risk Increase serum creatinine to ≥1.5- to 2-fold from ba RIFLE-Injury Increase serum creatinine to N2-fold to 3-fold from RIFLE-Failure Increase serum creatinine to N3-fold from baseline,

≥354 μmol/L with an acute rise of at least 44 μmol RIFLE-Loss Complete loss of kidney function for N4 wk End-stage kidney disease

End-stage kidney disease N3 mo

AKIN [11] Definition: an abrupt (within 48 h) reduction in kidn absolute increase in serum creatinine of either ≥0.3 a percentage increase of ≥50% (1.5-fold from basel output (after exclusion of hypovolemia and obstructi

Stage 1 Increase serum creatinine ≥26 μmol/L (N0.3 mg/dL equal to 1.5- to 2-fold from baseline

Stage 2 Increase serum creatinine to more than 2- to 3-fold f Stage 3 Increase serum creatinine to more than 3-fold from b

to ≥354 μmol/L with an acute rise of at least 44 μm Individuals who receive RRT are considered to have irrespective of the stage they are in at the time of RR

KDIGO [13] Definition: AKI is diagnosed if serum creatinine ≥2 to ≥1.5-fold from baseline, which is known or presu preceding 7 d.

Stage 1 Rise in serum creatinine ≥26.5 μmol/L in 48 h, or r Stage 2 Rise in serum creatinine 2.0-2.9 times from baseline Stage 3 Rise in serum creatinine 3 times from baseline, or in

≥353.6 μmol/L, or initiation of RRT irrespective of a Acute kidney injury diagnosis based on change between 2 creatinine valu

window for RIFLE criteria.

criteria in several aspects: (a) a lower serum creatinine threshold for the diagnosis of AKI, (b) the classification of patients requiring RRT as AKIN stage 3 independent of serum creatinine, (c) the removal of estimated glomerular filtration rate (GFR) criteria, (d) a shorter time window for diagnosing AKI (48 hours instead of 7 days), and (e) the elimination of an assumption that patients with missing baseline creatinine values had normal preexisting renal function. Epidemiologic studies collectively enrolling more than 500 000 patients confirmed that the RIFLE and/or AKIN criteria were valid tools to diagnose and stage AKI. Joannidis et al [10] directly compared the RIFLE and AKIN criteria in 14 356 critically ill patients using changes of serum creatinine and urinary output during the first 48 hours of ICU admission without including a requirement of RRT in the analysis. Although the mortality of patients with AKI classified by either RIFLE or AKIN criteria was similar, the

Urine output criteria

seline, or GFR decrease N25% b0.5 mL kg−1 h−1 for N6 h baseline, or GFR decrease N50% b0.5 mL kg−1 h−1 for N12 h or serum creatinine to /L, or GFR decrease N75%

b0.3 mL kg−1 h−1 for 24 h or anuria for 12 h

ey function defined as an mg/dL (≥26.4 μmol/L) or ine) or a reduction in urine on) ) or increase to more than or b0.5 mL kg−1 h−1 for N6 h

rom baseline b0.5 mL kg−1 h−1 for N12 h aseline, or serum creatinine ol/L

b0.3 mL kg−1 h−1 for 24 h or anuria for 12 h

met the criteria for stage 3, T.

6.5 μmol/L for ≤48 h, or rises med to have occurred in the

ise 1.5-1.9 times from baseline b0.5 mL kg−1 h−1 for 6-12 h b0.5 mL kg−1 h−1 for ≥12h

crease in serum creatinine to serum creatinine

b0.3 mL kg−1 h−1 for ≥24 h or anuria for ≥12 h

es within a 48-hour period for AKIN classification and within a 1-week

391Acute kidney injury after cardiac surgery according to RIFLE, AKIN and KDIGO

2 classifications classified individual patients differently. The percentage of patients who were identified as non-AKI by the AKIN classification but fulfilled the RIFLE criteria for AKI was 10.5%. By contrast, 3.5% of patients were classified as non-AKI according to the RIFLE criteria but fulfilled the AKIN criteria for AKI. Mortality of this group of patients was nearly twice that of patients who did not have AKI by both criteria (25.2% vs 12.9%). These results suggest that both RIFLE and AKIN criteria are useful tools to identify patients with AKI despite their differences. In an attempt to standardize the definition of AKI, the Kidney Disease: Improving Global Outcomes (KDIGO) initiative recently produced the KDIGO classification, which essen- tially combines the RIFLE and AKIN criteria [11] (Table 1). To date, this classification has not been validated in critically ill patients, including patients post–cardiac surgery.

The aims of this study were, first, to assess and compare the use of the AKIN and RIFLE criteria in patients post– cardiac surgery; second, to compare the AKIN and RIFLE criteria with the KDIGO classification; and third, to compare the epidemiology of patients with severe AKI requiring RRT with that from 10 and 20 years earlier.

2. Methods

2.1. Study design, setting, and population

We retrospectively analyzed data from patients older than 16 years who underwent cardiac surgery necessitating cardiopulmonary bypass (CPB) in a tertiary referral center in London, UK, over a 2-year period fromMay 2006 to April 2008 inclusive. Patients were excluded if there was a need for a ventricular assist device or extracorporeal membrane oxygenation, cardiac transplantation, need for more than 1 episode of CPB during the same admission, a requirement for RRT before surgery, or death within 24 hours of surgery. These exclusions were selected to permit direct comparison with previously published data from our institution [15].

Cardiopulmonary bypass was carried out using a calibrated roller pump (Stöckert, Munich, Germany) at a flow rate of 2.4 L min−1 m−2 at 28°C to 32°C [12]. Mean arterial pressure during CPB was maintained between 60 and 65 mm Hg using vasoactive agents, if necessary. Venous oxygen saturation was maintained in excess of 65%. Shed mediastinal blood was washed (Cell Saver, Haemonetics, Mass) and returned. Patient exposure to free hemoglobin was minimized by checking occlusion pressures on roller pumps before each case and by weekly quality control of cell salvage equipment. Patients undergoing coronary artery bypass grafting (CABG) received antibiotic prophylaxis with cefuroxime for a total of 4 doses. For surgery involving valve replacement or repair, teicoplanin was added and the duration of therapy extended until removal of intravascular and urinary catheters. Gentamicin prophylaxis was reserved for penicillin-allergic patients.

The need for RRT was established by the consultant intensivist in charge of the patient's care. Modality of choice was continuous venovenous hemodiafiltration (Prisma CFM; Hospal, Lyon, France) using AN69 membranes (surface area 1 m2) via a 12Fr double-lumen catheter (Dualyse; Vygon, Ecouen, France) inserted into the internal jugular or femoral vein. The blood pump speed was set at 150 mL/min, aiming for ultrafiltration and dialysate rates totalling approximately 2 L/h. Unfractionated heparin was used routinely for anticoagulation. In patients with contraindications to hepa- rin, epoprostenol and/or nonpharmacologic measures were used to keep the circuit patent.

Details of patient demographics, type of surgery, laboratory data, and preoperative, perioperative, and postoperative management were retrieved from an automated, prospectively collected database (CareVue; Phillips, Groeningen, the Netherlands). The highest AKIN, RIFLE, and KDIGO stages in the first 7 days after surgery were calculated and recorded (Table 1). Glomerular filtration rate was estimated using the Modification of Diet in Renal Disease formula [13]. Baseline renal function was determined by using the most recent serum creatinine, which was either the creatinine value taken in preadmission clinic or on admission to hospital. Urine output criteria were not used because our database did not contain 6- or 12-hourly urine output data for all patients. Postoperative day 1 was defined as the period up to 8 AM on the day after surgery, day 2 as the period until 8 AM on the subsequent day, and so on. Length of stay in ICU and hospital were expressed in days, rounded up to a whole integer.

2.2. Ethics

The need for individual informed consent was waived because this was a retrospective analysis of data collected prospectively for routine care, and there was no breach of privacy or anonymity (UK National Research Ethics Service).

2.3. Statistical analysis

Data were analyzed using GraphPad Prism version 4.02 (GraphPad Software, San Diego, CA, USA). Data were tested for normality using the Kolmogorov-Smirnov test. Normally distributed data were expressed as mean and SD, and nonnormally distributed data were expressed as median and interquartile range. Logistic regression analysis includ- ing calculation of odds ratio (OR) and 95% confidence intervals (CIs) and receiver operating characteristics curves were used to assess the association between maximum stage of AKI and hospital mortality. Areas under the receiver operating characteristics curve (AUC) were calculated, and differences between AUCs were compared using a nonpara- metric algorithm [14]. The relationship between maximum stage of AKI and length of stay in ICU and hospital was assessed by a negative binomial regression model. P values were calculated 2 sided, and analyses were performed using

392 A.J. Bastin et al.

R version 2.12.2 (R Foundation for Statistical Computing, Vienna, Austria) and Medcalc version 12.0.4 (MedCalc Software, Ostend, Belgium). Current data on the incidence and outcome of patients treated with RRT were compared with data from 1989 to 1990 [14] and 1997 to 1998 data [15] using Fisher exact test or a 1-sample t test, respectively.

3. Results

3.1. Baseline characteristics

During the 24-month study period, 1922 adult patients underwent cardiac surgery with CPB. We excluded patients who received a ventricular assist device (n = 8), had treatment with extracorporeal membrane oxygenation (n = 2), required RRT preoperatively (n = 7), had more than 1 episode of

Table 2 Baseline characteristics of 1881 patients undergoing cardiac surgery

Parameter Median (IQR) or n (%)

Age (y) 66 (56-74) Male sex 1340 (71.2) EuroSCORE a 5 (2-7) Logistic EuroSCORE a 3.49 (1.76-6.88) Redo surgery 222 (11.8) Nonelective surgery 341 (18.1) Preoperative IABP 14 (0.7) Preoperative creatinine (μmol/L) 92 (80-107) Preoperative eGFR (mL min−1 1.73 m−2)

68 (56-80)

Preoperative eGFR (mL min−1 1.73 m−2) ≥90 210 (11.2) 60-89 1045 (55.6) 45-59 424 (22.6) 30-44 166 (8.8) 15-29 33 (1.8) b15 2 (0.1) Procedure CABG only 895 (47.6) AVR only 288 (15.3) MVR/Repair only 165 (8.8) AVR + CABG 161 (8.6) Surgery involving aorta ± AVR 65 (3.5) Other 307 (16.3) Bypass time (min) b 98 (79-128) Cross-clamp time (min) b 59 (44-88) Circulatory arrest, n (%) b 24 (1.3)

IQR indicates interquartile range; AVR, aortic valve replacement; MVR, mitral valve replacement; eGFR, estimated GFR; IABP, intra-aortic balloon pump.

a EuroSCORE (European System for Cardiac Operative Risk Evaluation) not applicable to 194 patients with adult congenital heart disease. Data missing for 6 patients.

b Data missing for 65 (3.5%) patients.

CPB during the same admission (n = 5), and who died within 24 hours of surgery (n = 19). The remaining 1881 patients were included in the analysis. Baseline characteris- tics and operative details are shown in Table 2. One third of patients had a preoperative estimated GFR of less than 60 mL min−1 1.73 m−2 consistent with the diagnosis of chronic kidney disease stage 3 or worse.

3.2. Incidence of AKI

The incidence of AKI after cardiac surgery according to the AKIN and RIFLE criteria was 25.9% and 24.9%, respectively (Table 3). Most patients with AKI had maximum AKI stage on the second day postsurgery, but more than 40% of episodes of AKI occurred later (Fig. 1). The proportion of patients with AKIN stage 1 and RIFLE- Risk was also similar at 16.9% and 17.9%, respectively. However, there was a greater proportion of patients with AKIN stage 3 compared with RIFLE-Failure, mainly because of all patients on RRT were classified as having AKIN stage 3 independent of serum creatinine results. When applying the RIFLE criteria to this cohort, the proportion of patients in each RIFLE stage was 42/336 (12.5%) for RIFLE- Risk, 41/98 (41.8%) for RIFLE-Injury, and 35/35 (100%) for RIFLE-Failure, respectively. The number of patients who received RRT in each of the RIFLE categories was 42/336 (12.5%) for RIFLE-Risk, 40/98 (40.8%) for RIFLE-Injury, and 17/35 (48.6%) for RIFLE-Failure. Renal replacement therapy was started in 18 patients before the serum creatinine had risen to meet the RIFLE criteria for AKI, mainly for severe acidosis, hyperkalemia, and/or fluid overload.

The incidence and staging of AKI according to the KDIGO classification were identical to the AKIN classifi- cation for all 1881 patients, with no patients reclassified to a different stage. Data for the AKIN and KDIGO classifica- tions are therefore presented in the same column in Table 3. The agreement between AKIN and RIFLE staging of AKI is summarized in Table 4.

3.3. Outcome

There were 24 in-hospital deaths (1.3%), of which 19 occurred in patients with AKIN stage 3. Mortality increased in a stepwise fashion across RIFLE stages Risk, Injury, and Failure (Table 3). In a univariate logistic regression analysis, there was also a significant association between AKIN and RIFLE criteria and hospital mortality (OR, 4.3 [95% CI, 2.9- 6.3; P b .0001] for AKIN criteria and 2.7 [95% CI, 1.8-3.9; P b .0001] for RIFLE criteria). The AUC for hospital mortality was significantly higher using the AKIN classifi- cation (0.86; 95% CI, 0.85-0.88) compared with the RIFLE criteria (OR, 0.78; 95% CI, 0.76-0.80; P = .0009).

Length of stay in ICU and hospital increased in a stepwise fashion across AKIN stages 1 to 3 and RIFLE-Risk to Failure (Table 3). Negative binomial regression analysis confirmed a

Table 3 AKI stage, hospital mortality, and hospital and ICU length of stay according to RIFLE, AKIN, and KDIGO classifications (n = 1881)

Parameter Classification system, n (%) or median (IQR)

KDIGO and AKIN classification RIFLE classification

Maximum AKI stage No AKI 1394 (74.1) No AKI 1412 (75.1) AKIN stage 1 317 (16.9) RIFLE-Risk 336 (17.9) AKIN stage 2 34 (1.8) RIFLE-Injury 98 (5.2) AKIN stage 3 136 (7.2) RIFLE-Failure 35 (1.9) Any AKI stage 487 (25.9) Any AKI 469 (24.9)

Hospital mortality No AKI 4 (0.3) No AKI 5 (0.4) AKIN stage 1 1 (0.3) RIFLE-Risk 13 (3.8) AKIN stage 2 0 (0) RIFLE-Injury 4 (4.1) AKIN stage 3 19 (14.0) RIFLE-Failure 2 (5.7) Any AKI 20/487 (4.1%) Any AKI 19/469 (4.1%)

LOS hospital (d) No AKI 7 (6-9) No AKI 7 (6-10) AKIN stage 1 9 (7-13) RIFLE-Risk 9 (7-14) AKIN stage 2 14 (9-22) RIFLE-Injury 18 (10.8-29) AKIN stage 3 24 (13-46) RIFLE-Failure 20 (11-44)

LOS ICU (d) No AKI 1 (1-2) No AKI 1 (1-2) AKIN stage 1 2 (1-3) RIFLE-Risk 2 (1-4) AKIN stage 2 4 (1-8) RIFLE-Injury 6 (3-14) AKIN stage 3 13 (6-27) RIFLE-Failure 7 (5-23)

IQR indicates interquartile range; LOS, length of stay.

393Acute kidney injury after cardiac surgery according to RIFLE, AKIN and KDIGO

significant association between AKI and length of stay in hospital (r = 0.40 [P b .0001] for AKI as per AKIN classification and r = 0.41 [P b .0001] for AKI according to RIFLE criteria). There was a similar association with even closer correlation between AKI and length of stay in ICU (r = 0.74 [P b .0001] for AKI according to AKIN classification and r = 0.74 [P b .0001] for AKI as per RIFLE criteria).

Patients who developed AKI were significantly older, had a lower preoperative estimated GFR, had been on CBP for a significantly longer duration, and had previous cardiac

Fig. 1 Post-operative day when maximum AKIN stage occurred. Abbreviations: AKIN = acute kidney injury network.

surgery more often compared with patients who did not develop AKI (Table 5).

3.4. Comparison with historical data

Of the 1881 patients, 117 (6.2%) received RRT within 7 days of surgery. Their hospital mortality was 16.2%. In addition, 5 further patients required RRT after day 7. This prevalence rate of 6.5% (122/1881) was used as a comparator to previously published data from our institution from 1989 to 1990 and 1997 to 1998 [14,15]. Comparison confirmed a significant increase in the proportion of patients treated with RRT after cardiac surgery from 2.7% in 1989 to 1990 [16] to 6.5% (P b .0001). The associated hospital mortality in this cohort fell from 82.9% in 1989 to 1990 [14] to 53.8% in 1997 to 1998 [15] and 15.6% in 2006 to 2008 (P b .0001). More detailed analysis showed a trend toward earlier initiation of RRT, with a significantly higher proportion of patients who were started on RRT before serum urea was greater than 30 mmol/L and/or serum creatinine was greater than 300 μmol/L (Table 6).

4. Discussion

In this large, single-center cohort of patients post–cardiac surgery, the AKIN classification correlated better with hospital mortality than did the RIFLE criteria. The recently developed KDIGO classification did not alter the incidence or staging of AKI compared with using the AKIN classification. Three previous studies evaluated both the

Table 4 Agreement between RIFLE, AKIN, and KDIGO classifications of AKI in 1881 patients after cardiac surgery

AKI according to AKIN/KDIGO classification, n (%)

No AKI AKIN 1 AKIN 2 AKIN 3 Total

AKI according to RIFLE criteria, n (%) No AKI 1311 (69.7) 83 (4.4) 0 18 (1.0) 1412 (75.1) RIFLE-Risk 83 (4.4) 211 (11.2) 0 42 (2.2) 336 (17.9) RIFLE-Injury 0 23 (1.2) 34 (1.8) 41 (2.2) 98 (5.2) RIFLE-Failure 0 0 0 35 (1.9) 35 (1.9) Total 1394 (74.1) 317 (16.9) 34 (1.8) 136 (7.2) 1881 (100)

394 A.J. Bastin et al.

AKIN and the RIFLE classifications in patients after cardiac surgery [15–17]. Haase and colleagues [16] prospectively applied the AKIN and RIFLE criteria to 282 patients and showed a relatively high incidence of AKI at 44.7% and 45.8%, respectively. The AUC for in-hospital mortality was similar (0.94 and 0.91, respectively). The authors concluded that the AKIN classification offered no advantage over the RIFLE criteria for classifying AKI after cardiac surgery. Robert and colleagues [17] interrogated a large database of more than 25 000 patients undergoing cardiac surgery in 8 centers and also concluded that there was little difference in the incidence of AKI using the AKIN and RIFLE criteria (30% and 31%, respectively) and prediction of associated mortality (AUC, 0.79 and 0.78, respectively). Finally, Englberger and colleagues [18] compared the AKIN and RIFLE criteria in 4836 patients in a single center and found a significantly higher incidence of AKI according to the AKIN classification (26.3% vs 18.9%) but a similar AUC for in- hospital mortality (0.82 and 0.80, respectively). The authors attributed this difference to “overdiagnosis” of AKI in the early postoperative period as a result of hemodilution during CPB and lower serum creatinine on the first postoperative day serving as a reference point in the moving 48-hour AKIN diagnostic window. This overdiagnosis accounted for almost 10%. None of these patients had an increase in serum creatinine of 0.3 mg/dL or greater (26 μmol/L) from preoperative baseline values within the 7-day study period and would not have fulfilled the criteria for AKI if the early low postoperative value had been ignored [18].

Table 5 Preoperative and intraoperative characteristics of patients und and KDIGO)

Parameter, median (IQR) or n (%) AKIN/KDIGO

0 (n = 1340)

Age (y) 64 (54-73) Redo surgery 162 (11.6) Nonelective surgery 239 (17.1) Preoperative creatinine (μmol/L) 89 (78-101) Preoperative estimated GFR, (mL min−1 1.73 m−2) 71 (60-82) Duration of CPB (min) 95 (77-119) MAP during CPB (mm Hg) 65 (62-68)

IQR indicates interquartile range; MAP, mean arterial pressure. a Kruskal-Wallis test or χ2 test.

Three further studies compared the AKIN classification and RIFLE criteria in mixed populations of critically ill patients and showed either no difference [19,20] or a higher sensitivity and better prediction of mortality when using the RIFLE criteria [10]. There are several potential explanations for these discrepancies between different studies. First, in most studies, baseline creatinine values were not available, and it was assumed that preexisting renal function was normal. This approach ignores the fact that a proportion of AKI patients may have preexisting chronic kidney disease and may overdiagnose AKI. Second, in some studies, data on the use of RRT were either missing or not included [10,19], which means that patients who were started on RRT before the serum creatinine had risen enough to meet the criteria for AKI would be labeled as non-AKI. Third, even when only including patients post–cardiac surgery, the incidence and prognosis of AKI may be affected by nonpatient factors such as type of surgery and degree of urgency. As an example, data from the Society of Thoracic Surgeons National Database (2002-2004) showed that only 1.4% of patients needed acute RRT after cardiac surgery [21]. This lower rate compared with our data could be explained by the fact that in our cohort, twice as many patients had undergone complex cardiac surgery (aortic or mitral valve surgery or combined valve and CABG surgery), and only 47% had only CABG surgery. Fourth, the incidence and staging of AKI depend on the duration of observation and whether the diagnosis of AKI is based on serum creatinine alone or whether urine criteria are included.

ergoing cardiac surgery, classified according to AKI stage (AKIN

stage P a

1 (n = 310) 2 (n = 34) 3 (n = 132)

70.5 (61.5-77.5) 72 (62-76) 71.5 (63-77) b .0001 30 (9.4) 3 (8.6) 27 (20.1) .01 56 (17.6) 11 (31.4) 35 (26.1) .01 104 (89-125) 95 (79-112) 118 (89-158) b .0001 59 (47-71) 62 (51-81) 51 (34-64) b .0001 105 (83-137) 97 (82-154) 123 (97-163) b .0001 65 (62-69) 66 (62-71) 65 (61-69) .07

Table 6 Comparison of data from patients receiving RRT over a 20-year period

Parameter 1989-1990 a

(n = 1300) 1997-1998 b

(n = 2329) 2006-2008 (n = 1881)

P c

Incidence, n (%) 35 (2.7) 39 (1.7) 122 ⁎ (6.5) b .0001 In-hospital mortality (%) 82.9 53.8 15.6 b .0001 Age (y), mean (range) 56 (24-74) 65.3 (17-86) 68.2 (24-91) .02 Renal function at initiation of RRT, mean (SD) Serum urea (mmol/L) 30 (13) 25.3 (17.1) 13.0 (6.2) b .0001 Serum creatinine (μmol/L) 362 (141) 352 (147) 226 (101) b .0001 Parameters at initiation of RRT, n (%) Oliguria (urine output b20 mL/h) NA 23 (59) 49 (40) .04 Acidaemia (arterial pH b7.25) NA 6 (15) 38 (31) .06 Hyperkalemia (serum K+ N6 mmol/L) NA 8 (21) 20 (16) .63 Serum urea N30 mmol/L or serum creatinine N300 μmol/L

NA 22 (56) 32 (26) .0008

Duration of RRT in survivors (d), mean (SD) NA 11 (10.1) 6.7 (6.5) b .0001

NA indicates not available. a Baudouin et al [14]. b Ostermann et al [15]. c Fisher exact test or 1-sample t test (compared with 1997-1998 data). ⁎ Including 5 patients who started RR

Collepals.com Plagiarism Free Papers

Are you looking for custom essay writing service or even dissertation writing services? Just request for our write my paper service, and we'll match you with the best essay writer in your subject! With an exceptional team of professional academic experts in a wide range of subjects, we can guarantee you an unrivaled quality of custom-written papers.

Get ZERO PLAGIARISM, HUMAN WRITTEN ESSAYS

Why Hire Collepals.com writers to do your paper?

Quality- We are experienced and have access to ample research materials.

We write plagiarism Free Content

Confidential- We never share or sell your personal information to third parties.

Support-Chat with us today! We are always waiting to answer all your questions.