Acute Kidney Injury After Coronary Artery Bypass Grafting and Long-Term Risk of End-Stage Renal DiseaseCLINICAL PERSPECTIVE
Background—Acute kidney injury (AKI) is a common complication after coronary artery bypass grafting (CABG) and is associated with adverse outcomes. However, the relationship between AKI after CABG and the long-term risk of end-stage renal disease (ESRD) is unknown.
Methods and Results—This study included 29 330 patients who underwent primary isolated CABG in Sweden between 2000 and 2008. AKI was classified according to the Acute Kidney Injury Network (AKIN) classification: stage 1, >0.3 mg/dL (>26 μmol/L) or 50% to 100% increase; stage 2, 100% to 200% increase; and stage 3, >200% increase from the preoperative to postoperative serum creatinine level. Cox proportional hazards regression analysis was used to calculate hazard ratios with 95% confidence intervals for ESRD in AKIN stage 1 and stage 2 to 3. Postoperative AKI occurred in 13% of patients. During a mean follow-up of 4.3±2.4 years, 123 patients (0.4%) developed ESRD, including 50 (1.6%) in AKIN stage 1, 29 (5.2%) in AKIN stage 2 to 3, and 44 (0.2%) without AKI after CABG. After multivariable adjustment, the hazard ratio for ESRD was 2.92 (95% confidence interval, 1.87–4.55) for AKIN stage 1 and 3.81 (95% confidence interval, 2.14–6.79) for AKIN stage 2 to 3.
Conclusions—This nationwide study of patients who underwent CABG found that a small increase in the postoperative serum creatinine level was associated with an almost 3-fold increase in the long-term risk of ESRD after adjustment for a number of confounders, including preoperative renal function.
Acute kidney injury (AKI) is a common complication after coronary artery bypass grafting (CABG) and is associated with adverse outcomes.1,2 Several studies have investigated the impact of AKI on short- and long-term mortality.2–4 Approximately 0.6% to 5% of patients undergoing cardiac surgery will require dialysis in the immediate postoperative period,5–8 and these patients have a very high early mortality rate of ≈25% compared with 1% to 2% in patients who do not require immediate postoperative dialysis.6,7 In a cohort of patients who developed dialysis-requiring AKI during their hospital stay, the long-term risk of developing stage 4 or 5 chronic kidney disease was increased 28-fold.9
Clinical Perspective on p 2011
No previously reported studies have investigated whether AKI after cardiac surgery increased the long-term risk of end-stage renal disease (ESRD). One study of patients who underwent cardiac surgery found that more severe AKI was associated with an increased risk of progression of chronic kidney disease stage.10 Studies in various clinical settings found that patients with relatively small short-term changes in serum creatinine (s-Cr) levels had up to a 4-fold increase in the long-term risk of ESRD.10–13
In a recently published randomized, controlled trial, the risk of AKI and loss of renal function was investigated after off-pump versus on-pump CABG.14 The authors found that off-pump surgery reduced the risk of postoperative AKI but not loss of renal function at 1 year after surgery. In summary, studies of various cohorts found an increased risk of ESRD in patients with AKI. To the best of our knowledge, no previously reported studies investigated the relationship between AKI after CABG and the risk of ESRD. This study used a large nationwide cohort of patients who underwent primary isolated CABG to investigate the long-term risk of ESRD in patients who developed postoperative AKI. The study also investigated whether the strength of the association between AKI and ESRD was related to the preoperative renal function or the sex of the patient.
Patients were recruited from the Swedish Web-System for Enhancement and Development of Evidence-Based Care in Heart Disease Evaluated According to Recommended Therapies (SWEDEHEART) registry.15 This registry includes all patients who have undergone coronary angiography, percutaneous coronary intervention, or cardiac surgery in Sweden since 1992, with complete coverage of the whole country. Agreement between information in the registry and the medical records was reported to be between 93% and 97%.15 The present study included all patients who underwent CABG in Sweden between 2000 and 2008. The exclusion criteria are shown in Figure 1. Patients were excluded if they had undergone prior cardiac surgery, had undergone vascular or valvular surgery concurrently with CABG, had missing preoperative or postoperative s-Cr levels, had undergone surgery within 24 hours of the decision to operate, had a preoperative estimated glomerular filtration rate (eGFR) of <15 mL·min−1·1.73 m−2, or were dialysis dependent before surgery. Diabetes mellitus, hypertension, chronic obstructive pulmonary disease, and hyperlipidemia were recorded if patients were receiving ongoing pharmacological treatment for these conditions.
Patients were usually reviewed by a cardiologist a few weeks after surgery and then followed up by their family doctor.
The study complied with the guidelines of the Declaration of Helsinki and was approved by the regional ethics review board in Stockholm.
Definitions of Renal Function and AKI
eGFR was calculated from the simplified Modification of Diet in Renal Disease study equation.16 AKI was classified according to the Acute Kidney Injury Network (AKIN) classification: stage 1, >0.3 mg/dL (>26 μmol/L) or 50% increase; stage 2, 100% to 200% increase; and stage 3, >200% increase from the preoperative to the postoperative s-Cr level recorded in the SWEDEHEART registry. Patients with an increase of <0.3 mg/dL (<26 μmol/L) or <50% or a decrease in the postoperative s-Cr level were used as the reference group.17
The postoperative s-Cr level recorded in the SWEDEHEART registry is the highest level measured during the postoperative hospital stay. In a subgroup of 483 patients who underwent CABG at 1 center (Karolinska University Hospital) and developed postoperative AKI, the highest postoperative s-Cr level was measured at 48 hours after surgery in 93% of patients.4
All Swedish citizens have a unique personal identification number.18 The Swedish National Board of Health and Welfare created a database for this study using these personal identification numbers and the Swedish National Inpatient Registry, which includes all patients who have been hospitalized in Sweden since 1987. Information on ESRD was collected from the Swedish Renal Registry, which is a nationwide database of patients who have received renal replacement therapy in Sweden since 1991 and includes the dates of the start of chronic dialysis and renal transplantation. The Swedish Renal Registry includes >95% of patients on dialysis and almost 100% of patients who underwent renal transplantation.19,20 This study used the date of the first renal replacement therapy to define the start of ESRD. Survival status was determined with the Swedish personal identification number and the continuously updated Total Population Register at Statistics Sweden.18
Information on hospitalization before surgery with a principal discharge diagnosis of myocardial infarction, stroke, heart failure, or chronic obstructive pulmonary disease was obtained from the National Inpatient Register.
Patient characteristics are described with frequencies and percentages for categorical variables and means and standard deviations for continuous variables. Patients contributed person-time in days from the time of surgery and were censored at the time of death resulting from any cause, the start of renal replacement therapy, or the end of the follow-up period (December 31, 2008), whichever occurred first. The Cox proportional hazards regression model was used to analyze unadjusted and adjusted associations between AKI and the primary outcome, ESRD, and the secondary outcomes, ESRD or death. The hazard ratios (HRs) with 95% confidence intervals (CIs) for ESRD were calculated for each group of AKI, using patients with an increase of <0.3 mg/dL (<26 μmol/L) or a decrease in the postoperative s-Cr level as the reference category.
We included all the variables listed in Table 1 as covariates in the final statistical model for the primary analyses.
We used the cumulative incidence function to calculate the proportion of patients who developed ESRD, thereby accounting for the competing risk of death, that is, the possibility that death may occur before ESRD. The Kaplan-Meier method was used to calculate the cumulative incidence of the combined outcome of ESRD or death. Some values were missing from the study data set, including left ventricular ejection fraction in 7.7% of patients, diabetes mellitus status in 26%, and peripheral vascular disease status in 7.4%. Multiple imputation by chained equations was used to handle missing data.21 Twenty-five data sets were imputed, and estimates from these data sets were combined by use of standard methods. In addition to the main analysis of the multiply imputed data set, complete-case analysis was performed using only patients without missing values for model covariates (n=20 652). The results of the complete-case analyses were consistent with those of the main analyses and therefore are not reported. Stata version 13.1 (StataCorp LP, College Station, TX) was used for data management and all statistical analyses.
The selection of the study population is shown in Figure 1. Patient characteristics are presented in Table 1. A total of 29 330 patients with a mean age of 67 years were included, of whom 21% were women. Postoperative AKI occurred in 13% of patients, including 11% in patients in AKIN stage 1 and 1.9% in patients in AKIN stage 2 to 3. A total of 101 patients underwent postoperative temporary venovenous hemodialysis in the cardiac surgery intensive care unit. Preoperatively, 20% of the patients had an eGFR of <60 mL·min−1·1.73 m−2, but only 1% had an eGFR of 15 to 30 mL·min−1·1.73 m−2. Larger changes in postoperative s-Cr levels were associated with older age, lower preoperative eGFR, lower left ventricular ejection fraction, and prior myocardial infarction, stroke, heart failure, and diabetes mellitus (Table 1).
During the mean (4.3±2.4 years) and median (4.3 years; quartiles 1 and 3, 2.2 and 6.3 years), follow-up time, 123 patients (0.4%) developed ESRD, of whom 82 (67%) started hemodialysis, 40 (33%) started peritoneal dialysis, and 1 (0.8%) underwent kidney transplantation. The mean age at the start of dialysis or transplantation was 68±10 years. Patients who developed ESRD had a lower eGFR, higher rate of diabetes mellitus, and lower left ventricular ejection fraction at the time of surgery than those who did not develop ESRD. The incidence of ESRD increased with advancing AKIN stage. ESRD occurred in 44 patients (0.2%) without AKI, 50 (1.6%) in AKIN stage 1, and 29 (5.2%) in AKIN stage 2 to 3 (Table 2). The incidence rate of ESRD was 0.4 per 1000 person-years (95% CI, 0.30–0.53) in patients without AKI, 3.68 (95% CI, 2.79–4.86) in patients in AKIN stage 1, and 13.2 (95% CI, 9.14–18.9) in patients in AKIN stage 2 to 3. After multivariable adjustment, the HR for ESRD was 2.92 (95% CI, 1.87–4.55) in patients in AKIN stage 1 and 3.81 (95% CI, 2.14–6.79) in patients in AKIN stage 2 to 3.
During a mean follow-up period of 4.3±2.4 years, 4487 patients (15%) died. Death or ESRD occurred in 3554 patients (14%) without AKI, in 840 patients (27%) in AKIN stage 1, and in 216 patients (38%) in AKIN stage 2 to 3 (Table 2). The incidence rate of death or ESRD was almost 4-fold higher in patients in AKIN stage 2 to 3 than in patients in AKIN stage 1. The multivariable-adjusted HR for death or ESRD during the follow-up period was 1.34 (95% CI, 1.23–1.45) in patients in AKIN stage 1 and 2.32 (95% CI, 2.01–2.68) in patients in AKIN stage 2 to 3 (Table 2).
Figure 2 shows the cumulative incidence of ESRD over 8 years, taking death into account as a competing event. The cumulative incidence of ESRD increased with advancing AKIN stage. This difference was apparent throughout the follow-up period. Figure 3 shows the cumulative incidence of the combined outcome of ESRD or death during the 8-year follow-up period.
Relationship Between AKI and ESRD According to Preoperative Renal Function
In further subgroup analyses, patients were stratified according to preoperative renal function, and HRs for ESRD were calculated according to the presence or absence of AKI after adjustment for potential confounders (Table 3). There were few cases of ESRD among patients with normal preoperative renal function and postoperative AKI. In patients with normal preoperative renal function, the HR for ESRD in patients with postoperative AKI was 6.24 (95% CI, 1.94–20.1) compared with patients without AKI. In patients with a reduced preoperative eGFR, the relative risks of ESRD were lower, whereas the absolute risk of ESRD increased markedly. The relative risk of death or ESRD combined was similar for each level of preoperative renal function (Table 3).
Relationship Between AKI and ESRD According to Sex
The relationship between AKI and ESRD according to sex is shown in Table 4. There was a stronger association between AKI and ESRD in women (HR, 3.51; 95% CI, 1.46–8.40) than in men (HR, 2.97; 95% CI, 1.85–4.78), but these CIs overlap.
In a nationwide cohort of patients who underwent CABG during a 9-year period, we found that even a small postoperative increase of >0.3 mg/dL or 50% to 100% in the postoperative s-Cr level was associated with almost a 3-fold increase in the long-term risk of ESRD. We also found that a small postoperative increase in the s-Cr level was associated with an increased risk of long-term mortality.
Several previous epidemiological studies in various clinical settings reported an association between AKI and ESRD.10,11,22–24 AKI requiring dialysis in the immediate postoperative period is an independent risk factor for mortality.5 Even though only ≈0.6% of patients need dialysis during the postoperative period, this is an important prognostic factor for both all-cause mortality and long-term dialysis.6 A previous study reported that 8% of patients who survived >3 months after postoperative dialysis developed ESRD during 5.5 years of follow-up.6 In another cohort of patients who underwent coronary angiography and developed AKI, the risk of ESRD was >4-fold higher in patients in AKIN stage 1 and almost 12-fold higher in patients in AKIN stage 2 to 3 compared with patients without AKI.12 That cohort was slightly younger than ours but had approximately the same preoperative eGFR and a similar proportion of patients with diabetes mellitus. That study reported an overall progression to ESRD of 0.6% compared with 0.4% in our cohort and a death rate of 7.5% during 20 months of follow-up compared with 15% during 4.3 years of follow-up in our cohort.12 Another study found that AKI after coronary angiography was associated with a long-term decline in renal function and that the risk of progression to ESRD was 12-fold higher in patients with mild AKI than in patients without AKI.11
Among 108 000 general intensive care unit patients in Denmark, 2.8% developed dialysis requiring AKI during their hospital stay. The 5-year risk of ESRD in these patients was 3.8% compared with 0.3% in patients who did not require dialysis.22 Even though the setting was different in the Danish study, the risk of ESRD was similar to that in our cohort. In the Danish study, the association between AKI and ESRD was stronger in patients with normal kidney function than in those with chronic kidney disease. However, the definition of chronic kidney disease was not based on eGFR. In this study, patients were stratified according to preoperative renal function, and the relative risk of ESRD was higher in patients with a higher preoperative eGFR. The reason may be that healthy kidneys are able to tolerate greater trauma in terms of factors that may result in AKI (such as hypotension, hypoxia, and prolonged cardiopulmonary bypass time) before the s-Cr level increases compared with more vulnerable kidneys in patients with a lower preoperative eGFR. However, the absolute risk of ESRD was several times higher in patients with a lower preoperative eGFR than in patients with normal or moderately reduced renal function.
Our study is an observational study, and our findings cannot establish a potential causal relationship between AKI and loss of renal function. AKI could be a marker of risk rather than a risk factor in itself. Recently, the risk of AKI was investigated in patients randomized to either on-pump or off-pump primary isolated CABG surgery.14 In patients who underwent off-pump CABG, the relative risk of AKI was reduced by 17%, but there was no difference between the 2 groups in the reduction of eGFR 1 year after surgery. Interestingly, the need for postoperative dialysis was higher in patients who underwent off-pump CABG. However, the follow-up ended at 1 year, and the authors did not report the long-term need for renal replacement therapy.
Our study population and the baseline characteristics of this population were obtained from the SWEDEHEART registry, which has been found to have high validity.15 This registry includes a well-defined nationwide cohort with extensive baseline information. Outcome data were obtained from the Swedish Renal Registry and the national Total Population Registry, which include almost complete coverage of the whole country, thereby ensuring that no patients were lost to follow-up.19 In addition, the study cohort was relatively large, and the analyses included several potential confounding factors. The preoperative eGFR was available for all subjects. This is one of the most important confounding factors because it is an important predictor of adverse outcomes such as worsening of renal function and all-cause mortality.25,26 Another strength of this study is that it included only patients who underwent primary isolated nonemergent CABG. Patients who undergo emergency surgery often have altered hemodynamics, are more likely to receive potent drugs and fluids that may affect the s-Cr level, and are more likely to have a complicated postoperative course, which may affect long- and short-term outcomes.
The risk of death competes with the risk of ESRD, and the vast majority of patients with chronic kidney disease will die before they develop ESRD.27 This is reflected by the number of deaths in this study being several times higher than the number of cases of ESRD.
This study has some limitations that should be considered in the interpretation of the results. Although this was a nationwide study that included all patients who underwent CABG during a 9-year period in Sweden, relatively few patients developed ESRD during the follow-up period, resulting in wide CIs. Considering the small number of events, we believe that our results should be interpreted cautiously. The relative risk of ESRD increased almost 3-fold in patients in AKIN stage 1, but the absolute risk of ESRD was relatively small. In addition, patients who underwent valvular heart surgery were not included in the study, and our results may not be generalizable to cardiac surgery patients not undergoing primary isolated CABG. However, our study population was similar in terms of age and comorbidities to other cohorts of patients who underwent CABG in the United States or United Kingdom.8,28 We therefore believe that our data have high external validity and may be applicable to other countries with similar standards of health care. We were unable to determine whether the s-Cr level returned to baseline before patients were discharged from hospital. However, previous studies reported that postoperative AKI is associated with adverse long-term outcomes even if the s-Cr returns to the preoperative level.28,29 Information on preoperative proteinuria, which could affect outcome, was also not available. Another limitation was that data on diabetes mellitus were missing in 26% of the study population. We investigated potential differences in patient characteristics and prognoses between patients with and without information on diabetes mellitus (online-only Data Supplement) but found no evidence supporting a substantial difference. We used multiple imputation to handle missing data to reduce bias and to retain statistical power. We did not have information on diabetes mellitus managed with diet alone, which may have led to misclassification in terms of the presence or absence of diabetes mellitus in some patients. However, we believe that such misclassification would not have had a substantial effect on the study results because this potential bias would lead to dilution of the association between AKI and ESRD. Furthermore, no information was available on concurrent medications.
This study found a strong association between postoperative AKI and the development of ESRD for up to 9 years after surgery, even in patients with a small postoperative increase in s-Cr level. ESRD is a serious condition that affects both quality of life and mortality and leads to high healthcare costs. Our current knowledge of the importance of even a small increase in the postoperative s-Cr level suggests that patients with postoperative AKI should be closely monitored to optimize other known risk factors for ESRD such as blood pressure and proteinuria.
This large nationwide study found that AKI after CABG, even with only a small increase in the s-Cr level, was associated with an increased long-term risk of ESRD after adjustment for a number of confounders, including preoperative renal function.
We are grateful to the steering committee of SWEDEHEART for providing us with the data for this study.
Sources of Funding
This study was supported by grants from the Swedish Heart and Lung Foundation (Dr Rydén), the Swedish Society of Medicine (Drs Sartipy and Holzmann), the Karolinska Institutet Foundations and Funds (Dr Sartipy), and the Mats Kleberg Foundation (Dr Sartipy).
Continuing medical education (CME) credit is available for this article. Go to http://cme.ahajournals.org to take the quiz.
The online-only Data Supplement is available with this article at http://circ.ahajournals.org/lookup/suppl/doi:10.1161/CIRCULATIONAHA.114.010622/-/DC1.
- Received April 16, 2014.
- Accepted September 12, 2014.
- © 2014 American Heart Association, Inc.
- Lassnigg A,
- Schmidlin D,
- Mouhieddine M,
- Bachmann LM,
- Druml W,
- Bauer P,
- Hiesmayr M
- Hobson CE,
- Yavas S,
- Segal MS,
- Schold JD,
- Tribble CG,
- Layon AJ,
- Bihorac A
- Ivert T,
- Holzmann MJ,
- Sartipy U
- Conlon PJ,
- Stafford-Smith M,
- White WD,
- Newman MF,
- King S,
- Winn MP,
- Landolfo K
- James MT,
- Ghali WA,
- Knudtson ML,
- Ravani P,
- Tonelli M,
- Faris P,
- Pannu N,
- Manns BJ,
- Klarenbach SW,
- Hemmelgarn BR
- Garg AX,
- Devereaux PJ,
- Yusuf S,
- Cuerden MS,
- Parikh CR,
- Coca SG,
- Walsh M,
- Novick R,
- Cook RJ,
- Jain AR,
- Pan X,
- Noiseux N,
- Vik K,
- Stolf NA,
- Ritchie A,
- Favaloro RR,
- Parvathaneni S,
- Whitlock RP,
- Ou Y,
- Lawrence M,
- Lamy A
- Jernberg T,
- Attebring MF,
- Hambraeus K,
- Ivert T,
- James S,
- Jeppsson A,
- Lagerqvist B,
- Lindahl B,
- Stenestrand U,
- Wallentin L
- 20.↵SNR. Swedish Renal Registry 2012. http://www.medscinet.net/snr/. Accessed March 22, 2014.
- Kangasniemi OP,
- Mahar MA,
- Rasinaho E,
- Satomaa A,
- Tiozzo V,
- Lepojärvi M,
- Biancari F
- Loef BG,
- Epema AH,
- Smilde TD,
- Henning RH,
- Ebels T,
- Navis G,
- Stegeman CA
Acute kidney injury (AKI) is a common complication after coronary artery bypass grafting and increases the risk of short- and long-term morbidity and mortality. In several clinical settings, AKI has been related to an increased long-term risk of end-stage renal disease (ESRD). However, there is a lack of data on the association between AKI and ESRD after coronary artery bypass grafting. We included 29 330 patients who underwent a first isolated coronary artery bypass grafting between 2000 and 2008 in Sweden. During a mean follow-up of 4.3 years, 123 patients developed ESRD. We calculated the relative risk for ESRD in patients with AKI stages 1 and 2–3 compared with patients with no AKI and controlled for differences in baseline characteristics, including preoperative renal function. There was an almost 3-fold (hazard ratio, 2.96; 95% confidence interval, 1.92–4.58) increase in risk for ESRD for patients in AKI stage 1 and an almost 4-fold (hazard ratio, 3.77; 95% confidence interval, 2.14–6.64) increase in risk in patients with AKI stage 2–3. AKI after coronary artery bypass grafting is associated with a considerable increase in the long-term risk of ESRD. We believe that patients with AKI after coronary artery bypass grafting should be followed up more closely in terms of renal function and that renoprotective measures such as optimal blood pressure control and treatment with angiotensin-converting-enzyme inhibitors or angiotensin-receptor blockers should be instituted early to prevent or postpone the need for renal replacement therapy.