Association between transient acute kidney injury and morbidity and mortality after lung transplantation: A retrospective cohort study

Journal of Critical Care 29 (2014) 1028–1034

Contents lists available at ScienceDirect

Journal of Critical Care journal homepage: www.jccjournal.org

Association between transient acute kidney injury and morbidity and mortality after lung transplantation: A retrospective cohort study☆ Pedro Fidalgo, MD a, b, Mohammed Ahmed, MSc a, Steven R. Meyer, MD, PhD c, Dale Lien, MD d, Justin Weinkauf, MD d, Ali Kapasi, MD d, Filipe S. Cardoso, MD a, e, Kathy Jackson, RN, BNSc f, Sean M. Bagshaw, MD, MSc a,⁎ a

Division of Critical Care Medicine, Faculty of Medicine and Dentistry, University of Alberta, 2-124E Clinical Sciences Building, Edmonton, Alberta, Canada Nephrology Department, Hospital Prof Dr Fernando Fonseca, Amadora, Portugal c Division of Cardiac Surgery, Department of Surgery, University of Alberta, Edmonton, Alberta, Canada d Division of Pulmonary Medicine, Department of Medicine, Faculty of Medicine and Dentistry, University of Alberta, Edmonton, Alberta, Canada e Gastroenterology Department, Hospital Prof Dr Fernando Fonseca, Amadora, Portugal f Lung Transplant Program, Alberta Health Services, Edmonton, Alberta, Canada b

a r t i c l e

i n f o

Keywords: Acute kidney injury Lung transplantation Recovery Transient AKI Mortality Chronic kidney disease

a b s t r a c t Purpose: Acute kidney injury (AKI) is a common occurrence after lung transplantation (LTx). Whether transient AKI or early recovery is associated with improved outcome is uncertain. Our aim was to describe the incidence, factors, and outcomes associated with transient AKI after LTx. Materials and Methods: We performed a retrospective cohort study of all adult recipients of LTx at the University of Alberta between 1990 and 2011. Our primary outcome transient AKI was defined as return of serum creatinine below Kidney Disease–Improving Global Outcome AKI stage I within 7 days after LTx. Secondary outcomes included occurrence of postoperative complications, mortality, and long-term kidney function. Results: Of 445 LTx patients enrolled, AKI occurred in 306 (68.8%) within the first week after LTx. Of these, transient AKI (or early recovery) occurred in 157 (51.3%). Transient AKI was associated with fewer complications including tracheostomy (17.2% vs 38.3%; P b .001), reintubation (16.4% vs 41.9%; P b .001), decreased duration of mechanical ventilation (median [interquartile range], 69 [41-142] vs 189 [63-403] hours; P b .001), and lower rates of chronic kidney disease at 3 months (28.5% vs 51.1%, P b .001) and 1 year (49.6% vs 66.7%, P = .01) compared with persistent AKI. Factors independently associated with persistent AKI were higher body mass index (per unit; odds ratio [OR], 0.91; 95% confidence interval, 0.85-0.98; P = .01), cyclosporine use (OR, 0.29; 0.12-0.67; P = .01), longer duration of mechanical ventilation (per hour [log transformed]; OR, 0.42; 0.21-0.81; P = .01), and AKI stages II to III (OR, 0.16; 0.08-0.29; P b .001). Persistent AKI was associated with higher adjusted hazard of death (hazard ratio, 1.77 [1.08-2.93]; P = .02) when compared with transient AKI (1.44 [0.93-2.19], P = .09) and no AKI (reference category), respectively. Conclusions: Transient AKI after LTx is associated with fewer complications and improved survival. Among survivors, persistent AKI portends an increased risk for long-term chronic kidney disease. © 2014 Elsevier Inc. All rights reserved.

Abbreviations: AKI, acute kidney injury; BMI, body mass index; CI, confidence intervals; CKD, chronic kidney disease; CPB, cardiopulmonary bypass; CyA, cyclosporine; ECMO, extracorporeal membrane oxygenation; eGFR, estimated glomerular filtration rate; FEV1/ FVC, ratio forced expiratory volume 1 second and forced vital capacity; HR, hazard ratio; ICU, intensive care unit; IQR, interquartile range; KDIGO, Kidney Disease–Improving Global Outcome; LAS, lung allocation score; LTx, lung transplantation; OR, odds ratio; RRT, renal replacement therapy; sCr, serum creatinine criteria; Tac, tacrolimus. ☆ Disclosures/Conflicts: On behalf of all authors, the corresponding author states that there is no conflict of interest. ⁎ Corresponding author at: Division of Critical Care Medicine, Faculty of Medicine and Dentistry, University of Alberta, Clinical Sciences Building, 2-124E, 8440-112 Street NW, Edmonton, Alberta, Canada T6G2B7. Tel.: +1 780 492 9162; fax: +1 780 4921500. E-mail addresses: fi[email protected] (P. Fidalgo), [email protected] (M. Ahmed), [email protected] (S.R. Meyer), [email protected] (D. Lien), [email protected] (J. Weinkauf), [email protected] (A. Kapasi), [email protected] (F.S. Cardoso), [email protected] (K. Jackson), [email protected] (S.M. Bagshaw). http://dx.doi.org/10.1016/j.jcrc.2014.07.024 0883-9441/© 2014 Elsevier Inc. All rights reserved.

1. Introduction Acute kidney injury (AKI) occurs in approximately 50% to 70% of patients after lung transplantation (LTx) [1,2] and portends an increased risk for short- and long-term morbidity and mortality [3]. Although the long-term impact of AKI on survival has been well characterized [4], there is limited information on the modifying impact of transient AKI on renal outcomes. Most studies focused on renal recovery after AKI have used independence from renal replacement therapy (RRT) to define recovery [5]. However, this definition is limited by omitting those patients with milder forms of AKI not receiving RRT. Recovery from less severe AKI has been associated with improved long-term survival among patients undergoing cardiac surgery [6–8]. To date, few studies have

P. Fidalgo et al. / Journal of Critical Care 29 (2014) 1028–1034

described the impact of renal recovery after AKI following LTx [9,10], and no study has specifically evaluated the effect of transient AKI. We therefore hypothesized that transient AKI would be associated with improved short- and long-term outcomes in patients undergoing LTx compared with those not recovering function. Accordingly, our objectives were to (1) describe the incidence of transient AKI in a large cohort receiving LTx, (2) describe factors associated with transient AKI, and (3) describe the association of transient AKI on the occurrence of postoperative complications, mortality, kidney function, and resource use. 2. Methods The reporting of this study follows the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement for observational studies [11]. The local health research ethics board at the University of Alberta approved this study prior to commencement. The requirement for individual informed consent was waived. 2.1. Study design, setting, and population We performed a retrospective population-based cohort study of all adult patients receiving LTx at the University of Alberta between 1990 and 2011. The LTx program at the University of Alberta, the second largest in Canada [12] with an average of 30 to 35 lung transplants per year in the last decade, is the referral base for LTx in the provinces of Alberta and Saskatchewan (population ~ 5.2 million). Inclusion criteria were as follows: (1) age ≥18 years and (2) received any of single, double, or combined heart-lung transplantation. Exclusion criteria were as follows: (1) death within 24 hours of LTx; (2) preoperative advanced chronic kidney disease (CKD) stage 4 or 5, defined by the Kidney Disease–Improving Global Outcome (KDIGO) CKD classification [13]; (3) documented preoperative AKI or receipt of RRT within 1 month prior to LTx; and (4) data unavailable on vital status. 2.2. Surgical technique Cardiopulmonary bypass (CPB) was used in most cases (76%), with increasing frequency in the more recent LTx recipients. Standardized postoperative care was used in all patients, with hemodynamic monitoring performed in a specialized cardiovascular surgical intensive care unit. 2.3. Immunosuppression protocol All patients received induction therapy consisting of corticosteroids (intravenous methylprednisolone 0.5 g administered intraoperatively before reperfusion of the allograft followed by tapering postoperatively), antilymphocyte therapy (antithymocyte globulin [ATGAM, Pfizer, United States] 10 mg kg−1 d−1 for 5-10 days in 308 patients, rabbit antithymocyte globulin [Thymoglobulin, Sanofi, Canada] 1.5 mg kg−1 d−1 for 5-10 days in 11 patients, or dacluzimab 1 mg/kg on days 1 and 4 in 83 patients), and azathioprine 100 mg/d or mycophenalate mofetil 3 g/d. In 2000, mycophenalate mofetil replaced azathioprine in the baseline regimen for all patients. Calcineurin inhibitors were introduced on days 2 to 5 using either tacrolimus (Tac; target trough 10-15 ng/mL) or cyclosporine (CyA; target trough 250-350 ng/mL) prior to the introduction of Tac. Although CyA was the predominant calcineurin inhibitor used until 2006, Tac was the most commonly prescribed thereafter. The standard perioperative antimicrobial regimen consisted of prophylactic cefazolin (3 doses) or vancomycin (2 doses), if there was a penicillin allergy; sulfamethoxazole 400 mg/trimethoprim 80 mg daily; and fluconazole daily and gancyclovir, if either recipient or donor

1029

cytomegalovirus serology was positive. This regimen was individualized according to perioperative tissue cultures. 2.4. Operational definitions Lung allocation score (LAS) was calculated for all patients according to the method described by Egan et al [14]. Acute kidney injury was defined and severity was staged according to the serum creatinine criteria (sCr) of the KDIGO classification scheme [15]. We defined the presence of AKI by an absolute increase in sCr ≥26.5 μmol/L or 1.5fold relative change from baseline sCr in the first 7 postoperative days. Severity of AKI was classified as follows: stage I, an increase in sCr ≥ 26.5 μmol/L or 1.5 to 1.9 times baseline; stage II, an increase in sCr 2.0 to 2.9 times baseline; and stage III, an increase in sCr 3.0 times baseline, or an increase to sCr ≥354 μmol/L or initiation of RRT. Data on urine output were unavailable, and the urine output criteria of the KDIGO classification were omitted. Baseline sCr was defined as the lowest value available prior to day of surgery and estimated glomerular filtration rate (eGFR) was calculated according to the CKD Epidemiology Collaboration formula [16]. Chronic kidney disease was defined by an eGFR b 60 mL/min per 1.73 m 2 and further classified according to the KDIGO guidelines [13]. Transient AKI was defined as sCr values returning to the no-AKI range within the first 7 days after LTx in those not receiving RRT [17]. Patients who received RRT at any time during the first 7 days after LTx or those with incomplete recovery (reduction of sCr but still classified as AKI, irrespective of stage) during the 7 days after LTx were classified as “persistent AKI.” The causes of death were categorized according to the groups reported in the International Society for Heart and Lung Transplantation registry: bronchiolitis, acute rejection, malignancy, infection, graft failure technical, and other [18]. 2.5. Outcomes Primary outcome was incidence of transient AKI. Secondary outcomes included the following: (1) mortality in intensive care unit (ICU), hospital, and at 1 year; (2) duration of mechanical ventilation, reintubation rate, and tracheostomy rate; (3) kidney function at 3 months and 1 year; and (4) duration of posttransplant stay in the ICU and hospital and rates of ICU readmission. 2.6. Data collection Data were obtained from 2 sources: the University of Alberta LTx Program database and medical record review. All patients receiving either LTx or combined heart and LTx have been prospectively entered into this database since 1997, and prior to this, data were retrospectively entered. The LTx database routinely captures demographic (eg, age, sex, and weight), preoperative (eg, pulmonary diagnosis, pulmonary function, and comorbid disease), intraoperative (eg, procedure type, donor, and recipient-specific data), and postoperative data (eg, acute physiology, organ function, adverse events, immunosuppression regimen, survival, and graft function) on standardized case report forms. Patient medical records were reviewed to capture supplementary information not routinely captured by the LTx database (eg, details of RRT, kidney recovery, and long-term kidney function). 2.7. Statistical analysis Descriptive statistics were calculated for the entire cohort and expressed as mean (SD) or median (25th, 75th percentiles [interquartile range, or IQR]) for parametric and nonparametric continuous variables, respectively, and number (%) for categorical variables. We compared variables between individuals with transient and persistent AKI by using Student t test and Mann-Whitney test, or χ 2 and Fisher exact test, where appropriate. For multiple comparisons with 3 groups

1030

P. Fidalgo et al. / Journal of Critical Care 29 (2014) 1028–1034

Fig. 1. Summary of patient flow.

(“no AKI,” “transient AKI,” and “persistent AKI”), 1-way analysis of variance was used to compare means among the groups, followed by a Tukey P value adjustment to compare continuous variables between groups post hoc for those significant in the omnibus test. If data were not normally distributed, a Kruskal-Wallis test was performed. Multiple comparisons using the Mann-Whitney test was carried out to determine differences between categorical variables. In the event of missing data values or lost to follow-up, data were not replaced or estimated. Multivariate logistic regression was used to calculate the adjusted odds ratios (ORs) with 95% confidence intervals (CIs) for the variables associated with transient AKI. In addition to variables associated with transient AKI (in comparison to persistent) on univariate analysis with a level of significance at P b .1, those with a priori determined clinical significance were incorporated into the multivariate model. Covariates were assessed for multicollinearity and excluded if the variance inflation factor was more than 2.5 or the bivariate correlation between independent variables more than 0.80. The model fit was assessed by the Hosmer-Lemeshow goodness-of-fit test and calibration by the area under the receiver operative characteristic (AUROC) curve. Survival at 1 and 5 years was estimated using the Kaplan-Meier method, and the Breslow test was used to compare survival curves between groups. Multivariate Cox proportional hazard models were constructed to investigate independent associations between transient AKI and long-term risk of death. Explanatory variables used in the model included the following: demographics (age, sex, race), pretransplant comorbidities (body mass index [BMI], diabetes mellitus, pulmonary artery pressure ≥25 mm Hg, coronary artery disease, ratio forced expiratory volume 1 second and forced vital capacity (FEV1/FVC), eGFR, LAS, diagnosis of lung failure, intraoperative [operative procedure, CPB time] and postoperative variables (extracorporeal membrane oxygenation [ECMO] use, AKI stages II-III, hospital length of stay). For multiple comparisons with 3 groups, the Bonferroni correction was applied to account for the increased possibility of type I error, and P b .02 was considered significant. P values less than .05 were considered

statistically significant for all other comparisons. Statistical analysis was performed using SPSS version 21.0 (IBM Corp, Armonk, NY). 3. Results Of 484 patients receiving LTx during the study period, 445 patients fulfilled the eligibility criteria (Fig. 1). At LTx, median (IQR) age was 54 (41-60) years, 38% were female, and the most common etiologies of lung disease were chronic obstructive pulmonary disease (COPD; 34% [n = 149]), idiopathic pulmonary fibrosis (IPF; 22% [n = 99]), and cystic fibrosis (CF; 16% [n = 71]). Most patients received sequential double LTx (79.6% [n = 364]), with fewer receiving single LTx (16% [n = 71]) or heart-lung transplant (4.4% [n = 20]). The median (IQR) duration of follow-up was 3.9 (1.5-7.1) years. Acute kidney injury occurred in 68.8% (n = 306) within a median (IQR) of 1 (1-2) day after LTx. Of these, stage I occurred in 38.9% (n = 173), stage II in 17.5% (n = 78), and stage III in 12.4% (n = 55). Renal replacement therapy was received by 8.1 % (n = 36), with the majority (n = 30 [83.3%]) during the first week postoperatively. Of patients with AKI, transient AKI occurred in 51.3% (n = 157 [35.3%] overall). The median (IQR) time to recovery after AKI was 2 (1-3) days. 3.1. Factors associated with transient AKI Baseline and perioperative characteristics stratified by recovery status are shown in Tables 1 and 2. In univariate analysis, transient AKI was associated with younger age (OR, 0.98; 95% CI, 0.96-0.99; P = .01), lower BMI (per unit kg/m 2; OR, 0.92; 95% CI, 0.88-0.96; P b .001), IPF (OR, 0.51; 95% CI, 0.30-0.86; P = .01), previous sternotomy (OR, 0.48; 95% CI, 0.28-0.83; P = .01), and higher pre-LTx PaO2 (OR, 1.02; 95% CI, 1.01-1.03; P = .01). Postoperative use of CyA (OR, 0.40; 95% CI, 0.250.65; P b .001) was associated with a lower likelihood of transient AKI, whereas the use of Tac (OR, 2.55; 95% CI, 1.59-4.10; P b .001) was associated with a higher likelihood of transient AKI. Finally, transient

P. Fidalgo et al. / Journal of Critical Care 29 (2014) 1028–1034

1031

Table 1 Baseline characteristics of LTx recipients stratified by recovery status Variable

No AKI (n = 139; 31.2%)

Transient AKI (n = 157; 35.3%)

Persistent AKI (n = 149; 33.5%)

P

Age (y), median med (IQR) Male sex, n (%) White, n (%) BMI (kg/m2), median (IQR) Primary diagnosis COPD, n (%) IPF, n (%) CF, n (%) α-1 Antitrypsin, n (%) Talcosis Other diagnosisa, n (%) Comorbidities PAM N25 mm Hg, n (%) Diabetes mellitus, n (%) Coronary artery disease, n (%) eGFR (mL/min per 1.73 m2), mean (SD) CKD, n (%) Previous sternotomy, n (%) LAS score, median (IQR) 6-min walk (b150 m), n (%) FEV1/FVC (%), median (IQR) pCO2 (mm Hg) median (IQR; n = 432) paO2 (mm Hg) median (IQR; n = 368) Preoperative ECMO, n (%) Preoperative IMV, n (%) Preoperative NIV, n (%)

53 83 138 23

51 (38-60)⁎ 100 (63.7) 155 (98.7) 22 (19-27)⁎

56 93 149 26

.05 .78 .40 b.001

(42-60) (59.7) (99.3) (20-27)

58 (41.7) 24 (17.3) 17 (12.2) 18 (12.9) 3 (2.2) 7 (5) 42 (30.2) 7 (5.0) 34 (24.5) 96.1 (19.0) 4 (2.9) 34 (24.5) 33.9 (32.4-37.6) 12 (8.6) 36 (28-59) 45 (39-50) 63 (54-74) 0 (0) 1 (0.7) 2 (1.4)

48 (30.6) 29 (18.5)⁎ 34 (21.7) 9 (5.7) 12 (7.6) 15 (9.6) 56 (35.7) 12 (7.6) 39 (24.8) 98.6 (24.3) 10 (6.4) 27 (17.2)⁎ 35.4 (33.2-39.6)⁎⁎ 20 (12.7) 42 (29-73) 44 (39-50) 63 (55-76)⁎ 4 (2.5) 3 (1.9) 5 (3.2)

(44-61) (62.4) (100) (21-30)⁎⁎

43 (28.9) 46 (30.9)⁎⁎ 20 (13.4) 12 (8.1) 5 (3.4) 16 (10.7)

.04 .01 .05 .08 .05 .19

59 (39.6) 15 (10.1) 53 (35.6) 92.4 (23.0) 13 (8.7) 45 (30.2) 35.4 (33.0-42.5)⁎⁎ 15 (10.1) 50 (31-77)⁎⁎ 44 (37-50) 58 (50-68) 2 (1.3) 2 (1.3) 1 (0.7)

.25 .28 .06 .15 .11 .03 .01 .50 .01 .56 .01 .17 .67 .24

PAM indicates pulmonary artery pressure; IMV, invasive mechanical ventilation; NIV, noninvasive mechanical ventilation. P value—for trend comparisons across the 3 groups. a Included secondary pulmonary hypertension (n = 14), primary pulmonary hypertension (n = 11), sarcoidosis (n = 10), and interstitial lung disease (n = 3). ⁎ P b .02 (Bonferonni correction) for comparison between transient and persistent AKI. ⁎⁎ P b .02 (Bonferonni correction) for comparison between transient AKI and no AKI, or persistent AKI and no AKI.

AKI was associated with shorter duration of mechanical ventilation (per hour; OR, 0.998; 95% CI, 0.997-0.999; P b .001) and lower severity of AKI (stages II-III; OR, 0.11; 95% CI, 0.07-0.19; P b .001). By multivariate analysis, factors independently associated with transient AKI included lower BMI (OR, 0.91; 95% CI 0.85-0.98; P = .014), CyA use (OR, 0.27; 95% CI, 0.12-0.64; P = .003), shorter duration of mechanical ventilation (per natural log hour; OR, 0.46; 95% CI, 0.24-0.91; P = .03), and lower severity of AKI (stages II-III; OR, 0.15; 95% CI, 0.08-0.28; P b .001; Table 3).

[19-43] vs 39 [22-74], P b .001), compared with the persistent AKI group, respectively (Table 4; Figure S1). 3.3. Impact of transient AKI on short- and long-term survival Crude ICU, in-hospital, 1-year, and 5-year survival after LT were 96.9%, 95.1%, 88.5%, and 59.9%, respectively. Survival was significantly higher for those with transient compared with persistent AKI (Table 4; Fig. 2). In multivariable analysis, persistent AKI remained associated with higher mortality (hazard ratio [HR], 1.77; 95% CI, 1.08-2.93; P = .02; Table 5).

3.2. Impact of transient AKI on secondary outcomes 3.4. Impact of transient AKI on long-term kidney function Transient AKI was associated with lower rates of reintubation (OR, 0.27; 95% CI, 0.15-0.49; P b .001) and tracheostomy (OR, 0.34; 95% CI, 0.20-0.57; P b .001) as well as shorter postoperative stays (median [IQR], in days) in ICU and hospital (7 [4-10] vs 11 [5-24], P b .001; 26

At 3 months and 1 year after LTx, transient AKI patients had significantly higher eGFR and lower relative declines in eGFR compared with those with persistent AKI (Fig. 3; Table 4).

Table 2 Intraoperative and postoperative characteristics of LTx recipients stratified by recovery status Variable

No AKI (n = 139; 31.2%)

Transient AKI (n = 157; 35.3%)

Persistent AKI (n = 149; 33.5%)

P

CPB, n (%) CPB (min), median (IQR) Ischemia first lung (min; n = 433), median (IQR) Ischemia second lung (min; n = 432), median (IQR) Heart ischemia (min; n = 18), median (IQR) Double-lung transplant, n (%) Single-Lung transplant, n (%) Heart-lung transplant, n (%) CyA, n (%) Tac, n (%) Fluconazole, n (%) Voriconazole, n (%) ECMO post-Tx, n (%)

95 (68.3) 170 (0-235) 273 (143-344) 352 (233-422) 283 (243-323) 104 (74.8) 31 (22.3) 4 (2.9) 84 (60.4) 58 (41.7) 84 (60.4) 44 (31.7) 0 (0)

125 (79.6) 218 (0-272)⁎ 283 (172-348) 361 (256-425) 265 (156-345) 130 (82.8) 18 (11.5) ⁎ 9 (5.7) 82 (52.2)⁎⁎ 78 (50.0)⁎⁎ 86 (54.8)⁎⁎ 58 (36.9)⁎⁎ 1 (0.6)⁎⁎

118 219 285 370 266 120 22 7 109 42 111 25 10

.04 .001 .31 .24 .85 .22 .04 .49 .001 b.001 .001 b.001 b.001

P value—for trend comparisons across the 3 groups. ⁎ P b .02 (Bonferonni correction) for comparison between transient AKI and no AKI, or persistent AKI and no AKI. ⁎⁎ P b .02 (Bonferonni correction) for comparison between transient and persistent AKI.

(79.2) (0-273) ⁎ (170-380) (265-448) (166-389) (80.5) (14.8) (4.7) (73.2) (28.2)⁎ (74.5) ⁎ (16.8) ⁎ (6.7)⁎

1032

P. Fidalgo et al. / Journal of Critical Care 29 (2014) 1028–1034

Table 3 Risk factors of transient AKI after LTx Variable

Age (per year) BMI (per unit kg/m2) IPF Previous sternotomy Coronary artery disease eGFR (per mL/min per 1.73 m2) Transplant era (b2000 vs N2000) CyA Fluconazole ECMO Duration MV (natural log) AKI II-III vs AKI I

Univariate analysis, OR (95% CI)

Multivariate analysis Model 1, OR (95% CI)

Model 2, OR (95% CI)

0.98 0.92 0.51 0.48 0.60 1.01 1.98 0.40 0.42 0.09 0.26 0.15

0.99 0.94 1.18 0.48 0.90 1.00 1.68

0.98 0.91 1.03 0.62 0.86 0.99 1.88 0.29 0.68

(0.96-0.99) (0.88-0.96) (0.30-0.87) (0.28-0.83) (0.37-0.98) (1.00-1.02) (0.97-4.04) (0.25-0.65) (0.26-0.67) (0.01-0.71) (0.16-0.42) (0.09-0.26)

(0.97-1.01) (0.89-0.99) (0.59-2.36) (0.27-0.88) (0.50-1.61) (0.99-1.02) (0.79-3.56)

(0.95-1.01) (0.85-0.98) (0.41-2.59) (0.29-1.33) (0.41-1.82) (0.97-1.01) (0.68-5.21) (0.12-0.67) (0.30-1.53)

0.42 (0.21-0.81) 0.16 (0.08-0.29)

MV indicates mechanical ventilation. AKI II-III—stages II and III were collapsed into a categorical variable and those treated with RRT were excluded because they were included into the definition of persistent AKI. Hosmer-Lemeshow goodness-of-fit performed well (P N .3) for both models. Both models were adjusted for sex and race (P N .05 in univariate and multivariate analyses; data not shown). Model 1: n = 306, χ2 (df) = 24.20 (7), AUROC = 0.66. Model 2: n = 268, χ2 (df) = 96.45 (12), AUROC = 0.82; ECMO not included in the final model (n = 11).

greater incident CKD, and relative declines in kidney function during the first year after LTx.

4. Discussion We performed a retrospective cohort study of all patients receiving LTx from a large Canadian population base to describe the incidence, factors, and outcomes associated with transient compared with persistent AKI after LTx.

4.1. Key findings Transient AKI or early recovery from AKI occurring within the first 7 days after transplant was evident in approximately half of those with AKI. We found that lower BMI, no CyA exposure, shorter duration of mechanical ventilation, and less severe AKI (stage I compared with stages II-III) were independently associated with a higher likelihood of transient AKI and early postoperative recovery. Transient AKI was associated with a lower rate of postoperative complications including a shorter duration of mechanical ventilation and reduced reintubation and tracheostomy rates, along with shorter stays in both ICU and hospital. Importantly, we found that those with persistent AKI had higher adjusted risk of long-term mortality and among survivors,

4.2. Comparison to previous studies and interpretation of results Prior studies have largely defined renal recovery after AKI as dialysis independence; however, the association between less severe stages of AKI and recovery remains poorly characterized. We defined recovery as a return of sCr to below the diagnostic criteria for AKI stage I (SCr b1.5× baseline), as similarly described [17]. We further characterized transient AKI as recovery occurring within the first 7 postoperative days, based on prior data in cardiac surgical patients, having suggested that peak sCr or worst AKI stage generally occurs within the first 48 hours after surgery and rapid recovery is associated with improved outcome [6,19]. Indeed, we observed a peak in sCr at a median (IQR) of 3 (2-5) days. Although we recognize that recovery might have happened later than the seventh postoperative day in some patients, we hypothesized that persistent AKI (or failure to recover quickly) would be a strong modifying contributor to postoperative complications and less favorable outcomes.

Table 4 Summary of outcomes after LTx stratified by AKI and recovery status Variable MV (h), median (IQR) Tracheostomy, n (%) Reintubation, n (%) ICU LOS (d), median (IQR) Hospital LOS (d), median (IQR) Readmission to ICU, n (%) AKI 1a AKI 2a AKI 3a eGFR at 3 mo (mL/min per 1.73 m2; n = 422), mean (SD) Relative change eGFR at 3 mo (%), mean (SD)b eGFR at 1 y (mL/min per 1.73 m2; n = 392), mean (SD) Relative change eGFR at 1 y (%), mean (SD)b CKD at 3 mo (n = 422), n (%) CKD at 1 y (n = 392), n (%) Death ICU discharge, n (%) Death hospital discharge, n (%) Death at 1 y, n (%)

No AKI (n = 139; 31.2%) 42 (24-96) 13 (9.4) 20 (17.1) 4 (3-6) 21 (17-30) 5 (3.6)

75.3 (22.1) −19.5 (26.3) 59.1 (23.5) −37.5 (22.9) 38 (27.5) 76 (58) 0 (0) 0 (0) 7 (5)

Transient AKI (n = 157; 35.3%) 69 27 22 7 26 5 126 27 4 76.6 −20.6 63.0 −36.3 43 70 2 4 15

(41-142) ⁎ (17.2)⁎⁎ (16.4)⁎⁎ (4-10)⁎ (19-43)⁎ (3.6) (72.8) (34.6) (7.3) (25.4)⁎⁎ (25.6)⁎⁎ (25.8)⁎⁎ (20.5)⁎⁎ (28.5)⁎⁎ (49.6)⁎⁎ (1.3)⁎⁎ (2.5)⁎⁎ (9.6)⁎⁎

MV indicates mechanical ventilation; LOS, length of stay. P value for trend comparisons across the 3 groups. a % reported for the row. b Relative change = [(eGFR baseline − eGFR 1 year)/eGFR baseline] × 100. ⁎ P b .02 (Bonferonni correction) for comparison between transient AKI and no AKI, or persistent AKI and no AKI. ⁎⁎ P b .02 (Bonferonni correction) for comparison between transient and persistent AKI.

Persistent AKI (n = 149; 33.5%)

P

189 (63-403) ⁎,⁎⁎ 57 (38.3)⁎ 54 (41.9)⁎ 11 (5-24) ⁎,⁎⁎ 39 (22-74)⁎,⁎⁎ 6 (4) 47 (27.2) ⁎⁎ 51 (65.4)⁎⁎ 51 (92.7)⁎⁎ 60.2 (23.3)⁎ −34.4 (24.9)⁎ 51.7 (21.7)⁎

b.001 b.001 b.001 b.001 b.001 .9 b.001 b.001 b.001 b.001 b.001 .002 .01 b.001 .02 .001 b.001 b.001

−44.2 68 80 10 18 29

(24.3) (51.1)⁎ (66.7) (6.7)⁎ (12.1)⁎ (19.5)⁎

P. Fidalgo et al. / Journal of Critical Care 29 (2014) 1028–1034

Fig. 2. Kaplan-Meier survival curves stratified by transient AKI, persistent AKI, and no AKI.

Table 5 Risk of death after LTx stratified by recovery status from AKI

No AKI Transient AKI Persistent AKI

Unadjusted

Model 1

HR

HR

95% CI

Reference 1.19 0.81-1.76 1.72 1.19-2.47

Model 2 95% CI

Reference 1.43 0.95-2.16 1.88 1.27-2.76

HR

95% CI

Reference 1.43 0.93-2.19 1.77 1.08-2.93

Model 1 (n = 437, χ2 [df] = 41.50 [19]): age, sex, race, BMI (in kg/m2), pretransplant comorbidities (diabetes mellitus, pulmonary artery pressure N25 mm Hg, coronary artery disease, FEV1/FVC, eGFR [in mL/min per 1.73 m2], LAS score), type of Lung Tx (single LTx vs other), and lung disease diagnosis (COPD [reference category] vs IPF vs CF vs α-1 antitrypsin vs talcosis vs others). Model 2 (n = 430, χ2 [df] = 82.34 [23]) was adjusted for model 1 plus CPB time (in minutes), ECMO use, AKI stages II to III (vs no AKI stage 1), and length of hospital stay (in days). Persistent AKI, unadjusted, P = .004: model 1, P = .001; model 2, P = .024.

Few prior studies have described the impact of transient AKI on outcome. Among the 11% of patients developing AKI after acute myocardial infarction, transient AKI occurred in 65% of those with AKI, defined as normalization of serum creatinine within the index hospitalization. In this study, transient AKI was independently associated with increasing long-term mortality [20]. In a large hospital-based cohort study, transient AKI represented one third of all AKI, defined as recovery of kidney function to no-AKI RIFLE

1033

category within 72 hours of onset, and was similarly shown to be independently associated with increased hospital mortality [17]. In our cohort, transient AKI occurred in approximately half of all patients with AKI and was not independently associated with longterm risk of death. This was considerably higher than the early recovery described by Wehbe et al [10] in their cohort of LTx patients, despite a similar incidence of AKI. There are, however, some differences that should be considered. First, this study considered a period of 2 weeks after LTx to define AKI. Second, they used the return to baseline sCr during the hospital stay to define recovery, whereas we opted to define recovery as a return of sCr to the no-AKI range. Finally, they reported a lower rate of CyA use (15.5%) compared with our study, which we found to be independently associated with persistent AKI. In multivariate analysis, only higher BMI, CyA use, longer duration of mechanical ventilation, and AKI stages II to III were independently associated with higher likelihood of persistent AKI. Longer duration of postoperative mechanical ventilation has been shown to be associated with development of AKI [2] but has not been described as a potential contributor to persistent AKI after LTx [7,10]. Because our data are derived from an observational study, we are unable to attribute causality between mechanical ventilation and AKI. However, a number of lung-kidney factors may contribute to persistent AKI after LTx including pathologic organ crosstalk, in particular in lungs that have suffered ischemia/reperfusion injury and risk of ventilator-induced lung injury [21,22]. Similarly, persistent AKI could also contribute to weaning failure and prolonged respiratory failure by disrupting acid-base and fluid balance homeostasis in the context of impaired pulmonary lymphatic drainage [23]. On the other hand, mechanical ventilation itself may worsen kidney function by stimulating fluid retention (ie, raised increased intrathoracic pressure). Similar to Wehbe et al [10], we also found severity of AKI to be an important predictor of early recovery [24]. Calcineurin inhibitor therapy has long been associated with a decline in kidney function and incident CKD in solid-organ transplantation [25–27]. We found that CyA use was an independent risk factor for persistent AKI, with its mechanism being potentially related not only to systemic exposure but also dependent on individual variability and local kidney factors [28]. Finally, we found higher BMI to be an independent risk for persistent AKI after LTx. There is a growing body of evidence showing an independent association between obesity and risk of postoperative AKI in cardiac surgery. A number of plausible mechanisms have been postulated, including higher intra-abdominal pressure, longer operative time, prolonged weaning from mechanical ventilation, and greater basal measures of oxidative stress [29]. Whether these mechanisms contribute to the higher risk of persistent AKI among obese patients remains to be definitively proven.

Fig. 3. Summary of long-term kidney function by AKI status: (A) at 3 months and (B) at 1 year.

1034

P. Fidalgo et al. / Journal of Critical Care 29 (2014) 1028–1034

In contrast to Wehbe et al [10], we found that persistent AKI was independently associated with worse long-term survival, whereas transient AKI showed similar long-term survival to those without AKI. Interestingly, the same association between failure to recover from AKI and worse survival has been previously reported among other cohorts of critically ill surgical patients [7,8,30]. Persistent AKI in our study was associated with a more pronounced loss of kidney function at 3 months and 1 year after LTx compared with those with transient AKI. Interestingly, even among those with transient AKI, a substantial loss of kidney function from preoperative baseline was evident at 3 months and 1 year after LTx. These findings imply that even in patients with transient AKI after LTx who recover to baseline function, there is a heightened risk of more rapid long-term loss of kidney function and a greater need for monitoring and risk modification. Our study has a number of limitations worthy of discussion. First, the findings of our study are derived from a single modestly sized LTx referral center. Second, we retrospectively interrogated a prospectively captured database of all LTx performed at our center. Third, our study included data on LTx performed over a 20-year era where advances and modifications in patient selection, immunosuppression, operative technique, and postoperative support have evolved. Accordingly, despite our large, robust, and complete data set, these limitations may have predisposed our study to bias and confounding, along with limited generalizability. Finally, we did not have available data available on urine output to integrate into the KDIGO classification of AKI or data on fluid balance. Patients with fluid overload have been reported to be significantly less likely to recover kidney function [31], and it seems that a higher fluid overload at RRT initiation predicts worse kidney recovery [32]. We believe that the specific impact of postoperative fluid balance on AKI, recovery, and survival after LTx warrants further investigation. 5. Conclusions Acute kidney injury is common after LTx, and transient AKI, compared with persistent AKI, is associated with fewer complications, a shorter course in hospital, improved survival, and better preservation of long-term kidney function among survivors. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.jcrc.2014.07.024. Acknowledgments The authors acknowledge Matthew Hubert's contribution to the collection of data. Dr Bagshaw holds a Canada Research Chair in Critical Care Nephrology and is a Clinical Investigator supported by Alberta Innovates–Health Solutions (AI-HS). Dr Fidalgo and Dr Cardoso are supported by an unrestricted educational grant donated by Gambro Inc. References [1] Arnaoutakis GJ, George TJ, Robinson CW, et al. Severe acute kidney injury according to the RIFLE (risk, injury, failure, loss, end stage) criteria affects mortality in lung transplantation. J Heart Lung Transplant 2011;30(10):1161–8. [2] Rocha PN, Rocha AT, Palmer SM, Davis RD, Smith SR. Acute renal failure after lung transplantation: incidence, predictors and impact on perioperative morbidity and mortality. Am J Transplant 2005;5(6):1469–76. [3] Wehbe E, Brock R, Budev M, et al. Short-term and long-term outcomes of acute kidney injury after lung transplantation. J Heart Lung Transplant 2012;31(3):244–51.

[4] Coca SG, Yusuf B, Shlipak MG, Garg AX, Parikh CR. Long-term risk of mortality and other adverse outcomes after acute kidney injury: a systematic review and metaanalysis. Am J Kidney Dis 2009;53(6):961–73. [5] Macedo E, Bouchard J, Mehta RL. Renal recovery following acute kidney injury. Curr Opin Crit Care 2008;14(6):660–5. [6] Swaminathan M, Hudson CC, Phillips-Bute BG, et al. Impact of early renal recovery on survival after cardiac surgery-associated acute kidney injury. Ann Thorac Surg 2010;89(4):1098–104. [7] Mehta RH, Honeycutt E, Patel UD, et al. Impact of recovery of renal function on long-term mortality after coronary artery bypass grafting. Am J Cardiol 2010;106 (12):1728–34. [8] Hobson CE, Yavas S, Segal MS, et al. Acute kidney injury is associated with increased long-term mortality after cardiothoracic surgery. Circulation 2009;119 (18):2444–53. [9] Pham PT, Slavov C, Pham PC. Acute kidney injury after liver, heart, and lung transplants: dialysis modality, predictors of renal function recovery, and impact on survival. Adv Chronic Kidney Dis 2009;16(4):256–67. [10] Wehbe E, Duncan AE, Dar G, Budev M, Stephany B. Recovery from AKI and shortand long-term outcomes after lung transplantation. Clin J Am Soc Nephrol 2013;8 (1):19–25. [11] von Elm E, Altman DG, Egger M, Pocock SJ, Gotzsche PC, Vandenbroucke JP. Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. BMJ 2007;335(7624): 806–8. [12] Vargas Hein O, Staegemann M, Wagner D, et al. Torsemide versus furosemide after continuous renal replacement therapy due to acute renal failure in cardiac surgery patients. Ren Fail 2005;27(4):385–92. [13] Ostermann M, Chang RW. Acute kidney injury in the intensive care unit according to RIFLE. Crit Care Med 2007;35(8):1837–43. [14] Egan TM, Murray S, Bustami RT, et al. Development of the new lung allocation system in the United States. Am J Transplant 2006;6(5 Pt 2):1212–27. [15] Kidney Disease: Improving Global Outcomes (KDIGO) Acute Kidney Injury Work Group. KDIGO clinical practice guideline for acute kidney injury. Kidney Int Suppl 2012;2:1–138. [16] Levey AS, Stevens LA, Schmid CH, et al. A new equation to estimate glomerular filtration rate. Ann Intern Med 2009;150(9):604–12. [17] Uchino S, Bellomo R, Bagshaw SM, Goldsmith D. Transient azotaemia is associated with a high risk of death in hospitalized patients. Nephrol Dial Transplant 2010;25 (6):1833–9. [18] Christie JD, Edwards LB, Kucheryavaya AY, et al. The Registry of the International Society for Heart and Lung Transplantation: 29th adult lung and heart-lung transplant report—2012. J Heart Lung Transplant 2012;31(10):1073–86. [19] Yehia M, Collins JF, Beca J. Acute renal failure in patients with pre-existing renal dysfunction following coronary artery bypass grafting. Nephrology (Carlton) 2005;10(6):541–3. [20] Choi JS, Kim YA, Kim MJ, et al. Relation between transient or persistent acute kidney injury and long-term mortality in patients with myocardial infarction. Am J Cardiol 2013;112(1):41–5. [21] Shilliday IR, Quinn KJ, Allison ME. Loop diuretics in the management of acute renal failure: a prospective, double-blind, placebo-controlled, randomized study. Nephrol Dial Transplant 1997;12(12):2592–6. [22] Imai Y, Parodo J, Kajikawa O, et al. Injurious mechanical ventilation and end-organ epithelial cell apoptosis and organ dysfunction in an experimental model of acute respiratory distress syndrome. JAMA 2003;289(16):2104–12. [23] Koyner JL, Murray PT. Mechanical ventilation and the kidney. Blood Purif 2010;29 (1):52–68. [24] Devarajan P. Update on mechanisms of ischemic acute kidney injury. J Am Soc Nephrol 2006;17(6):1503–20. [25] Esposito C, De Mauri A, Vitulo P, et al. Risk factors for chronic renal dysfunction in lung transplant recipients. Transplantation 2007;84(12):1701–3. [26] Bloom RD, Reese PP. Chronic kidney disease after nonrenal solid-organ transplantation. J Am Soc Nephrol 2007;18(12):3031–41. [27] Barraclough K, Menahem SA, Bailey M, Thomson NM. Predictors of decline in renal function after lung transplantation. J Heart Lung Transplant 2006;25(12):1431–5. [28] Naesens M, Kuypers DR, Sarwal M. Calcineurin inhibitor nephrotoxicity. Clin J Am Soc Nephrol 2009;4(2):481–508. [29] Billings IV FT, Pretorius M, Schildcrout JS, et al. Obesity and oxidative stress predict AKI after cardiac surgery. J Am Soc Nephrol 2012;23(7):1221–8. [30] Bihorac A, Yavas S, Subbiah S, et al. Long-term risk of mortality and acute kidney injury during hospitalization after major surgery. Ann Surg 2009;249(5):851–8. [31] Bouchard J, Soroko SB, Chertow GM, et al. Fluid accumulation, survival and recovery of kidney function in critically ill patients with acute kidney injury. Kidney Int 2009;76(4):422–7. [32] Heung M, Wolfgram DF, Kommareddi M, Hu Y, Song PX, Ojo AO. Fluid overload at initiation of renal replacement therapy is associated with lack of renal recovery in patients with acute kidney injury. Nephrol Dial Transplant 2012;27(3):956–61.