Importance of stratifying acute kidney injury in cardiogenic shock resuscitated with mechanical circulatory support therapy

ACQUIRED: MECHANICAL CIRCULATORY SUPPORT ACQ

Importance of stratifying acute kidney injury in cardiogenic shock resuscitated with mechanical circulatory support therapy Andrew I. Abadeer, MEng,a Paul Kurlansky, MD,a Codruta Chiuzan, PhD,b Lauren Truby, MD,a Jai Radhakrishnan, MD,c Reshad Garan, MD,d Veli Topkara, MD,d Melana Yuzefpolskaya, MD,d Paolo Colombo, MD,d Koji Takeda, MD, PhD,a Yoshifumi Naka, MD, PhD,a and Hiroo Takayama, MD, PhDa ABSTRACT Objective: Although the outcomes of patients with cardiogenic shock remain poor, short-term mechanical circulatory support has become an increasingly popular modality for hemodynamic assistance and organ preservation. Because the kidney is exquisitely sensitive to poor perfusion, acute kidney injury is a common sequela of cardiogenic shock. This study examines the incidence and clinical impact of acute kidney injury in patients with short-term mechanical circulatory support for cardiogenic shock. Methods: Retrospective review was performed of 293 consecutive patients with cardiogenic shock who were treated with short-term mechanical circulatory support. The well-validated 2014 Kidney Disease Improving Global Outcomes criteria were used to stage acute kidney injury. Outcomes of interest were longterm mortality and renal recovery. Results: Acute kidney injury developed in 177 of 293 patients (60.4%), of whom 113 (38.6%) were classified with stage 3 (severe). Kaplan–Meier survival estimates indicated a 1-year survival of 49.2% in the nonsevere (stages 0-2) acute kidney injury cohort versus 27.3% in the severe acute kidney injury cohort (P<.001). Multivariable Cox regression demonstrated that severe acute kidney injury was a predictor of long-term mortality (hazard ratio, 1.54; confidence interval, 1.102.14; P ¼ .011). Among hospital survivors, renal recovery occurred more frequently (82.4% vs 63.2%, P ¼ .069) and more quickly (5.6 vs 24.5 days, P<.0001) in the nonsevere than in the severe acute kidney injury group. Conclusions: Acute kidney injury is common and frequently severe in patients in cardiogenic shock treated with short-term mechanical circulatory support. Milder acute kidney injury resolves with survival comparable to patients without acute kidney injury. Severe acute kidney injury is an independent predictor of longterm mortality. Nonetheless, many surviving patients with acute kidney injury do experience gradual renal recovery. (J Thorac Cardiovasc Surg 2017;154:856-64)

From the Divisions of aCardiothoracic Surgery, cNephrology, and dCardiology, and b Department of Biostatistics, Mailman School of Public Health, Columbia University Medical Center, New York, NY. This study was supported by National Institutes of Health T35 Grant DK 93430-3 Y.N. has received consulting fees from Thoratec Corp. Received for publication July 28, 2016; revisions received March 5, 2017; accepted for publication April 12, 2017; available ahead of print May 26, 2017. Address for reprints: Hiroo Takayama, MD, PhD, 177 Fort Washington Ave, New York, NY 10032 (E-mail: [email protected]). 0022-5223/$36.00 Copyright Ó 2017 by The American Association for Thoracic Surgery http://dx.doi.org/10.1016/j.jtcvs.2017.04.042

856

Long-term mortality by severity of AKI. Central Message Although milder AKI resolves with survival comparable to patients without AKI, severe AKI is a predictor of long-term mortality. Still, many surviving patients experience gradual renal recovery. Perspective The impact of AKI in the extremely high-risk hemodynamically compromised population of patients with CS treated with ST-MCS remains poorly defined. This retrospective review of a robust clinical experience demonstrates the negative impact of AKI on long-term survival, yet also defines the likelihood and time course of renal recovery.

See Editorial Commentary page 865. See Editorial page 855.

Patients with cardiogenic shock (CS) continue to have a poor prognosis. The mortality of those with acute myocardial infarction (AMI) complicated by CS, the most common Scanning this QR code will take you to supplemental figures and tables for this article.

The Journal of Thoracic and Cardiovascular Surgery c September 2017

Acquired: Mechanical Circulatory Support

ACQ

Abadeer et al

Abbreviations and Acronyms AKI ¼ acute kidney injury AMI ¼ acute myocardial infarction CI ¼ confidence interval CS ¼ cardiogenic shock CVA ¼ cerebrovascular accident ECMO ¼ extracorporeal membrane oxygenation IQR ¼ interquartile range KDIGO ¼ Kidney Disease Improving Global Outcomes LVAD ¼ left ventricular assist device PCS ¼ postcardiotomy shock RRT ¼ renal replacement therapy ST-MCS ¼ short-term mechanical circulatory support VAD ¼ ventricular assist device

cause of CS, remains up to 50% despite advances in early revascularization, hemodynamic device therapy, and pharmacotherapy.1-4 Short-term mechanical circulatory support (ST-MCS) devices such as venoarterial extracorporeal membrane oxygenation (ECMO) and external ventricular assist devices (VADs) (surgical or percutaneous) are increasingly being used in this setting to improve systemic perfusion and protect vital organ function.5,6 Multisystem organ dysfunction is a hallmark of recent or persistent systemic hypoperfusion and has been demonstrated to occur in 37% to 56% of patients with CS.7-10 As a consequence of the kidney’s exquisite sensitivity to poor perfusion, it is one of the first organs to suffer in CS, and its injury often persists even after restoration of perfusion with ST-MCS. Knowledge of the prognostic implication of renal injury is becoming more clinically relevant as more patients with CS and nonrecoverable myocardial injury are being sustained and evaluated for heart replacement therapy with durable left ventricular assist device (LVAD). Although acute kidney injury (AKI) is believed to develop frequently and be associated with significant morbidity after MCS resuscitation, available information on this entity is extremely limited. AKI often is reported only as part of a case series or meta-analysis without detailed analyses or stringent definitions, and the few studies that focus on AKI suffer from a small sample size and include only patients treated with ECMO.11 We evaluated the incidence of AKI in patients on STMCS and the effect of such injury on mortality. To better understand this clinically important entity, we investigated AKI both holistically and inclusively using granular variables, including detailed medical history, preoperative state, postoperative outcomes, and long-term follow-up.

MATERIALS AND METHODS Management of Cardiogenic Shock At the Columbia University Medical Center, both vasopressors and inotropes with or without an intra-aortic balloon pump are the first-line treatment for patients with signs of CS. A systolic blood pressure that was less than 90 mm Hg, a cardiac index of less than 2.0 L/min/m2, and evidence of end-organ dysfunction despite pharmacologic and intra-aortic balloon pump augmentation defines refractory CS. At this juncture, patients were immediately evaluated for ST-MCS candidacy specifically considering the patient’s and family’s goals of care, desire to pursue further treatment, length of ongoing cardiopulmonary resuscitation, alternative causes of shock such as sepsis, and relevant comorbidities.1 Once the decision has been made to use ST-MCS, ECMO or ST-VAD is placed. An external VAD with a magnetically levitated centrifugal pump (CentriMag, Abbott, Abbott Park, Ill) is used as a ST-VAD in our program.12 ECMO is used primarily in patients in whom neurologic status is unclear, in patients who are too hemodynamically unstable for transport to an operating room, and in patients with severe coagulopathy.1 Because of the complex nature of cardiorenal syndrome, there is no clear-cut decision tree or consensus detailing the initiation of renal replacement therapy (RRT). In our program, AKI, which often manifested with oliguria and further complicated by refractory acidosis, volume overload, or electrolyte anomaly, results in nephrology consultation. The decision to initiate continuous RRT is made on an individual case-by-case basis. Once ST-MCS is begun, patients are monitored and managed by a multidisciplinary heart team and are evaluated for heart transplantation or destination LVAD therapy. On clinical improvement of the general status of the patient on ST-MCS, myocardial function is evaluated by weaning device support under both echocardiographic and hemodynamic monitoring. With successful weaning, the device is explanted. Myocardial recovery is defined as survival to hospital discharge or more than 30 days after the device explant. Otherwise, if appropriate, the patient may undergo a device exchange to an implantable LVAD or total artificial heart, or undergo heart transplantation. If a patient neither achieves myocardial recovery nor meets the inclusion criteria of heart replacement therapy, comfort care is established.

Data Collection This study was approved by the institutional review board of Columbia University Medical Center with waiver of consent. This was a retrospective single-center study focusing on 293 consecutive patients who received STMCS for refractory CS between 2007 and 2013, either with ST-VAD or ECMO. Excluded were those patients who had preexisting end-stage renal disease (n ¼ 11), who had inadequate/incomplete data to stage AKI (n ¼ 2), or for whom RRT was initiated more than 1 day before device insertion (n ¼ 14). Laboratory values were collected from the patients’ charts immediately before device insertion and periodically thereafter throughout the hospital stay. All patients were followed at least until hospital discharge. Outpatient charts were reviewed for postdischarge follow-up, and patients were censored from survival analysis at the time of last known follow-up. Baseline renal status and degree of AKI were classified on the basis of the well-validated and widely accepted 2014 Kidney Disease Improving Global Outcomes (KDIGO) criteria.13,14 A creatinine level just before device insertion was used as a baseline and compared with the highest creatinine level in the first 7 days of device support. The severity of kidney injury was staged according to this ratio (Table 1). Patients placed on RRT within the first 7 postoperative days were recorded as stage 3. Although KDIGO specifies criteria for urine output that accompany creatinine levels, these were not used because of variability in output recording, and therefore strictly speaking, the current criteria are modified KDIGO. Renal recovery was assessed at discharge for patients who experienced AKI and was defined as freedom from RRT and, on the basis of KDIGO definitions, normalization of creatinine to within 150% of baseline. Date

The Journal of Thoracic and Cardiovascular Surgery c Volume 154, Number 3

857

Acquired: Mechanical Circulatory Support

Abadeer et al

ACQ TABLE 1. Defining acute kidney injury Category

Stage*

n (%)

Serum creatinine

Nonsevere

No AKI I II III

116 (39.6) 40 (13.7) 24 (8.2) 113 (38.6)

<1.5 3 baseline Cr 1.5-< 2.0 3 baseline Cr 2.0-< 3.0 3 baseline Cr 3.0 3 baseline Cr OR Initiation of CVVH

Severe

AKI, Acute kidney injury; Cr, creatinine; OR, odds ratio; CVVH, continuous venovenous hemofiltration. *Only the first 7 postoperative days were used in the staging of AKI.

of renal recovery was recorded as the date in which 3 consecutive creatinine levels were within 150% of baseline.

Outcomes of Interest The primary outcome of interest was postoperative AKI, whereas longterm mortality was a secondary outcome. To address the impact of severe AKI on these outcomes, patients were divided into 2 groups on the basis of KDIGO AKI stage: The ‘‘severe’’ AKI group included those who had stage 3 AKI or were placed on RRT within the first 7 postoperative days. Because clinical behavior of those in stages 0, 1, and 2 were not distinguishable within the numeric limitations of this study, these patients were combined into the ‘‘nonsevere’’ AKI group (Table 1).

Statistical Analysis Continuous variables were summarized as means and standard deviations for normally distributed data, or medians and interquartile range (IQR) for non-normal distributions and compared using 2-sample independent t tests or Wilcoxon rank-sum tests as appropriate. Categoric variables were described using frequency and percentages and compared using chisquare tests. Logistic and Cox regression models were used for identifying predictors of severe AKI and long-term mortality, respectively. All variables (demographics, medical history, and preoperative characteristics) that showed potential significant difference between the severe and nonsevere AKI groups were considered in univariable analyses. Further, variables with a P value .20 or less in the univariable models were included in the multivariable analyses, with the Hosmer–Lemeshow test used to assess the goodness-of-fit of the final model. There were no missing data regarding the outcome of severity of AKI. Multiple imputation was used for missing preoperative data for clinically relevant continuous variables. Twenty complete datasets were generated using the multivariate normal algorithm and used in Cox or logistic analyses, the results of which were pooled among these datasets into a single imputation result. Imputed variables were considered missing at random, and the completeness of these variables included alanine aminotransferase (20.5% missing), hemoglobin (10.9% missing), and mean arterial pressure (20.1% missing). Variables such as lactate, with more than 25% missing data, were reported but excluded from analysis. Kaplan–Meier method was used to estimate survival from the date of device implant to death or last follow-up. Survival estimates between the 2 groups were compared with a 2-sided log-rank test. In addition, univariable and multivariable Cox regressions were used to obtain the hazard ratio estimates and 95% confidence intervals (CIs) for the selected variables. The proportional hazards assumption was tested for the Cox models by including an interaction with ln(time) and assuring nonsignificance of the interaction, as well as by plotting and comparing the ln {ln(survival)} of the survival curves for each category. The proportional hazards assumption was not violated. All analyses were performed in Stata IC 12.0 (StataCorp LP, College Station, Tex), with a type I error set at 0.05.

858

RESULTS Patient Characteristics The baseline characteristics of patients who were placed on ST-MCS for CS are compared in Table 2. A total of 293 patients were eligible for inclusion in the study and received MCS for CS therapy during the study period. Of these, 163 were placed on ECMO and 130 were placed on VAD (Table E1). The causes of CS included postcardiotomy shock (PCS) in 90 patients (30.7%), AMI in 79 patients (27.0%), acute decompensated heart failure in 51 patients (17.4%), and transplant graft dysfunction in 36 patients (12.3%). Of the entire cohort of 293 patients, 177 (60.4%) experienced some stage of AKI. Of these, 40 (13.7%) experienced stage 1 AKI, 24 (8.2%) experienced stage 2 AKI, and 113 (38.6%) were classified as stage 3 or ‘‘severe’’ AKI. Only 116 patients (39.6%) experienced no AKI. Comparing the severe AKI group and the nonsevere group, there was a higher proportion of chronic kidney disease (30.1% vs 18.3%, P ¼ .02) and prior cerebrovascular accident (CVA) (14.2% vs 3.9%, P ¼ .001). The proportion of PCS as the cause of CS was higher in the severe AKI group (38.9% vs 25.6%, P ¼ .016), whereas AMI had a lower representation in the severe AKI group (Table 2). Preoperative hemodynamics and laboratory data are shown in Table 3 with significant differences in hemoglobin, creatinine, and aspartate aminotransferase between groups. Preoperative mean arterial pressure was 65.6  19.6 mm Hg for patients with PCS, 63.6  24.7 mm Hg for patients with AMI, 67.6  20.9 mm Hg for patients with graft dysfunction, 65.9  21.3 mm Hg for patients with acute decompensated heart failure, and 59.9  21.0 mm Hg for other causes of CS. The severe AKI group more frequently received ECMO as ST-MCS (P ¼ .027). During hospitalization, 131 patients (44.7%) required RRT in the form of continuous venovenous hemofiltration. The average age of patients requiring RRT was 56.2  15.4 years. Of these, 44 (33.6%) survived to hospital discharge with an average hospital stay of 84.8  58.1 days. Average length of RRT among those patients who survived to discharge was 16.0  11.6 days. Among survivors, 11 (8.4%) required hemodialysis on hospital discharge.

Renal Recovery Among the 177 patients who experienced some degree of AKI, 72 (40.7%) survived to hospital discharge. Among these, 52 patients (72.2%) experienced renal recovery with more patients experiencing renal recovery in the nonsevere than in the severe AKI group (82.4% vs 63.2%, P ¼ .069) (Table E2). Figure 1 displays the cumulative incidence of renal recovery over time, indicating that 85% of patients who experienced renal recovery recovered within

The Journal of Thoracic and Cardiovascular Surgery c September 2017

Acquired: Mechanical Circulatory Support

ACQ

Abadeer et al

TABLE 2. Demographics, medical history, and cause of cardiogenic shock Characteristics

Overall

Patients (n) 293 Patient demographics and medical history n (%) Age (y); mean  SD 55.8  15.4 Age >65 y 86 (29.4) Male 199 (67.9) 28.3  6.7 BMI (kg/m2) CAD 136 (46.4) HLD 118 (40.3) HTN 158 (53.9) COPD 24 (8.2) DM 90 (30.7) Prior CVA 23 (7.9) CKD 67 (22.9) Cause of CS (%) PCS 90 (30.7) AMI 79 (27.0) Graft 36 (12.3) ADHF 51 (17.4) Other 38 (13.0)

P value

Nonsevere

Severe

180

113

55.5  15.2 48 (26.7) 119 (66.1) 28.0  6.8 79 (43.9) 67 (37.2) 90 (50.0) 16 (8.9) 53 (29.4) 7 (3.9) 33 (18.3)

56.3  15.7 38 (33.6) 80 (70.8) 28.9  6.8 57 (50.4) 51 (45.1) 68 (60.2) 8 (7.1) 37 (32.7) 16 (14.2) 34 (30.1)

.653 .203 .403 .244 .274 .179 .089 .583 .551 .001 .020

46 (25.6) 56 (31.1) 21 (11.7) 34 (18.9) 24 (13.3)

44 (38.9) 23 (20.4) 15 (13.3) 17 (15.0) 14 (12.4)

.016 .043 .683 .398 .815

Data are presented as mean  standard deviation or n (%). Graft or primary graft dysfunction. P values represent a comparison of the severe with the nonsevere AKI group. Continuous variables were summarized as means and standard deviations, or medians and IQR for non-normal and compared using 2-sample independent t tests or Wilcoxon rank-sum tests. Categoric variables were described using frequency and percentages and compared using chi-square tests. SD, Standard deviation; BMI, body mass index; CAD, coronary artery disease; HLD, hyperlipidemia; HTN, hypertension; COPD, chronic obstructive pulmonary disease; DM, diabetes mellitus; CVA, cerebrovascular accident; CKD, chronic kidney disease; CS, cardiogenic shock; PCS, postcardiotomy shock; AMI, acute myocardial infarction; ADHF, acute decompensated heart failure. Bold values indicate a P value < .05.

14 days in the nonsevere group compared with only 33% of the severe AKI group (P <.000). The patients with severe AKI who survived to hospital discharge with renal recovery (n ¼ 24) had a mean renal recovery time of 24.5 days;

significantly longer than those with stage 1 or 2 AKI (n ¼ 28, 5.6 days, P < .000). Eleven patients (15.3%) required maintenance hemodialysis on hospital discharge, and 9 patients (n ¼ 12.5) had some degree of residual renal

TABLE 3. Operative characteristics Characteristics

n

Preoperative hemodynamics SBP (mm Hg) 229 DBP (mm Hg) 229 MAP (mm Hg) 235 Postoperative hemodynamics (24-h) SBP (mm Hg) 220 DBP (mm Hg) 220 MAP (mm Hg) 247 Laboratory values Blood Hgb (g/dL) 259 Creatinine (mg/dL) 290 AST (U/L) 233 ALT (U/L) 233 Lactate (mg/dL) 143 Other preoperative characteristics, n (%) IABP Active CPR ST-MCS modality, n (%) ECMO

Severe

P value

Overall

n

Nonsevere

n

92.6  25.9 51.2  21.2 64.7  21.6

138 138 140

92.8  24.7 53.1  21.1 65.9  21.7

91 91 94

92.3  27.9 48.3  21.1 63.0  21.3

.876 .092 .318

97.6  20.1 71.1  17.7 79.8  16.1

130 130 147

97.6  19.1 72.1  17.7 80.4  16.0

90 90 100

97.6  20.1 69.6  17.7 79.0  16.2

.974 .290 .517

10.7  2.4 1.4 (1-1.9) 76 (29-264) 46 (21-128) 4.2 (2.2-8.5)

159 180 143 143 80

11.1  2.3 1.3 (1.0-1.7) 65 (27-179) 45 (20-111) 3.7 (2.1-8.2)

101 111 90 90 63

10.1  2.4 1.6 (1.2-2.3) 104 (37-370) 51.5 (23-168) 5.2 (2.4-9.3)

.001 .000 .019 .274 .201

132 (45.1) 55 (18.8)

82 (45.6) 31 (17.2)

50 (44.3) 24 (21.2)

.827 .391

163 (55.6)

91 (50.6)

72 (63.7)

.027

Data are presented as mean  standard deviation or median (IQR, Q1-Q3), unless otherwise specified. P values represent a comparison of the severe with the nonsevere AKI group. SBP, Systolic blood pressure; DBP, diastolic blood pressure; MAP, mean arterial pressure; Hgb, hemoglobin; AST, aspartate aminotransferase; ALT, alanine aminotransferase; IABP, intra-aortic balloon pump; CPR, cardiopulmonary resuscitation; ST-MCS, short-term mechanical circulatory support; ECMO, extracorporeal membrane oxygenation. Bold values indicate a P value < .05.

The Journal of Thoracic and Cardiovascular Surgery c Volume 154, Number 3

859

Acquired: Mechanical Circulatory Support

Abadeer et al

ACQ FIGURE 1. Cumulative incidence of renal recovery by severity of AKI. AKI, Acute kidney injury.

injury in the form of a persistently elevated creatinine from preoperative baseline (Table E2). Outcomes Patient outcomes are reported in Table 4. Median length of device support was 6 days (IQR, 3-14): 7 days (IQR, 316) in the nonsevere group and 5 days (IQR, 2-12) in the severe group (P ¼ .073). In-hospital mortality of the overall cohort was 53.9%: 66.4% in the severe AKI group, and 46.1% in the nonsevere AKI group (P <.001). Only 5 patients (4.5%) in the severe AKI group received durable VAD compared with 30 patients (17.1%, P ¼ .002) in the nonsevere AKI group. Transition to heart transplantation was rare in the severe AKI group (2.7% vs 11.7% in nonsevere group, P ¼ .006). The 1-year Kaplan-Meier survival estimates were 49.2% (CI, 41-57) in the nonsevere cohort and 27.3% (CI, 19-37) in the severe AKI cohort (P <.000) (Figure 2). Multivariable Cox regression revealed that severe AKI was a significant predictor of long-term mortality (hazard ratio, 1.54; CI, 1.10-2.14; P ¼ .011) along with age and device type (Table 5), and the effects of severe AKI on the survival function when adjusted for other variables are displayed in

Figure E1. To clarify the effect of early mortality, Cox regression was performed excluding those patients with mortality within 7 and 30 days. In excluding patients with mortality within 7 days, the multivariable hazard ratio for subsequent overall mortality of those with severe AKI versus those with nonsevere AKI was 2.07 (1.28-3.37, P ¼ .003). Likewise, when excluding patients with mortality within 30 days, multivariable Cox regression yielded a hazard ratio of 3.50 (1.58-7.77, P ¼ .002). As such, the impact of severe AKI was significant for the duration of the follow-up period. Independent predictors of severe AKI included prior CVA (odds ratio, 3.10; CI, 1.14-8.42; P ¼ .026) and baseline creatinine (odds ratio, 1.42; CI, 1.10-2.13; P ¼ .012) (Table 6). To investigate the impact of less severe AKI on survival, the nonsevere AKI group was subdivided into stage 0, 1, and 2 AKI groups. Neither AKI stage 1 nor 2 was found to be a predictor of longterm mortality. Subgroup Analyses The renal function of the patients with PCS and transplant graft dysfunction also could have suffered from surgical insult and cardiopulmonary bypass. To mitigate these alternate mechanisms of renal injury, subgroup analyses were performed focusing on the patients with AMI (n ¼ 79) and acute decompensated heart failure (n ¼ 51). Among these 130 patients, severe AKI developed in 40 (30.7%). The median length of device support was 7 days (IQR, 3-20) with significantly longer support time in the nonsevere group (8.5 days: IQR, 4-21 vs 5 days: IQR, 216 in severe group, P ¼ .039) (Table E3). Both transition to VAD and heart transplantation were less common in the severe AKI group (12.5% and 5.0%, respectively) compared with the nonsevere group (21.8% and 17.8%, respectively) (Table E3). The 1-year survival estimate was 26.9% in the severe AKI group versus 50.4% in the nonsevere AKI group (P ¼ .004, Figure E2). Just as in the full cohort, severe AKI was a significant predictor for long-term mortality

TABLE 4. Outcomes Characteristics

Overall (n ¼ 293)

Nonsevere (n ¼ 180)

Severe (n ¼ 113)

P value

Length of support Hospital stay ICU stay In-hospital mortality n (%) 30-d mortality Destinations, n (%) Myocardial recovery Transition to durable VAD OHT

6 (3-14) 36 (15-62) 16 (5-32) 158 (53.9) 126 (43.0)

7 (3-16) 38 (18.5-56.5) 17 (7-33) 83 (46.1) 64 (35.6)

5 (2-12) 32 (10-68) 13 (4-30) 75 (66.4) 62 (54.9)

.073 .512 .608 .001 .001

60 (33.9) 30 (17.1) 21 (11.7)

32 (28.6) 5 (4.5) 3 (2.7)

.344 .002 .006

92 (31.8) 35 (12.2) 24 (8.2)

Data are presented as median (IQR, Q1-Q3) or n (%). P values represent a comparison of the severe with the nonsevere group. ICU, Intensive care unit; VAD, ventricular assist device; OHT, orthotopic heart transplant. Bold values indicate a P value < .05.

860

The Journal of Thoracic and Cardiovascular Surgery c September 2017

Acquired: Mechanical Circulatory Support

ACQ

Abadeer et al

FIGURE 2. Long-term mortality by severity of AKI. AKI, Acute kidney injury.

(hazard ratio, 2.26; CI, 1.23-4.17; P ¼ .009) (Table E4). Baseline creatinine also remained an independent predictor of severe AKI (hazard ratio, 1.67; CI, 1.05-2.69; P ¼ .032) (Table E5). Overall, the subanalyses showed results similar to those of the main cohort analyses in terms of the primary outcome of interest.

DISCUSSION This is one of the largest studies describing AKI and its impact on outcomes in patients with CS resuscitated with ST-MCS. The major findings are that (1) only 40% of the patients were free from any AKI; 40% had severe AKI, 14% had mild AKI, and 8% had moderate AKI; (2) history of CVA and preoperative creatinine were independent predictors for developing severe AKI; (3) time until recovery of renal function differed significantly between severe and nonsevere renal injury patients; (4) severe AKI, but not nonsevere AKI, resulted in significantly higher in-hospital mortality (>60%) and lower 1-year survival (27%); and (5) severe AKI, but not nonsevere AKI, was an independent predictor of late mortality. AKI complicating CS is a well-described phenomenon. The mechanism of AKI in refractory CS likely is due to a combination of poor renal perfusion and severe renal congestion. Although these mechanisms may be reversible with prompt institution of MCS, the present study suggests that more than half of patients with CS develop AKI despite restoration of systemic perfusion with ST-MCS and that in this group of patients, AKI is most frequently of the ‘‘severe’’ subtype. In addition to renal injury due to impaired

TABLE 5. Predictors of long-term mortality (Cox regression) Characteristics

HR (95% CI)

P value

Multivariable HR (95% CI)

P value

Age BMI Medical history CAD HLD HTN DM COPD Prior CVA CKD Preoperative status Cause 1-PCS 2-AMI 3-Graft 4-ADHF 5-Other MAP Hemoglobin Baseline creatinine ALT IABP Active CPR Device (CentriMag [Thoratec, Pleasanton, Calif] ¼ 0, ECMO ¼ 1) Postoperative status Severe AKI

1.02 (1.00-1.03) 1.00 (0.98-1.03)

.000 .674

1.02 (1.00-1.03)

.004

1.25 (0.92-1.69) 0.99 (0.73-1.36) 1.05 (0.78-1.43) 0.93 (0.67-1.29) 0.84 (0.45-1.54) 1.10 (0.64-1.90) 1.11 (0.79-1.58)

.153 .968 .738 .647 .564 .732 .541

1.18 (0.81-1.71)

.394

Reference 0.66 (0.44-0.98) 0.47 (0.27-0.82) 0.72 (0.47-1.12) 0.82 (0.49-1.37) 0.99 (0.98-0.99) 0.94 (0.88-1.02) 1.07 (0.98-1.18) 1.00 (1.00-1.00) 0.76 (0.55-1.03) 1.97 (1.38-2.79) 1.93 (1.41-2.65)

.068 .040 .008 .147 .446 .000 .140 .139 .940 .077 .000 .000

Reference 0.69 (0.43-1.08) 0.64 (0.35-1.17) 0.99 (0.61-1.60) 1.02 (0.60-1.72) 0.99 (0.98-1.00) 0.99 (0.92-1.07) 1.08 (0.97-1.21)

.20 .106 .145 .959 .938 .235 .778 .179

0.96 (0.67-1.36) 1.35 (0.78-2.34) 1.45 (1.00-2.11)

.802 .291 .048

1.73 (1.28-2.36)

.000

1.54 (1.10-2.14)

.011

HR, Hazard ratio; CI, confidence interval; BMI, body mass index; CAD, coronary artery disease; HLD, hyperlipidemia; HTN, hypertension; DM, diabetes mellitus; COPD, chronic obstructive pulmonary disease; CVA, cerebrovascular accident; CKD, chronic kidney disease; PCS, postcardiotomy shock; AMI, acute myocardial infarction; ADHF, acute decompensated heart failure; MAP, mean arterial pressure; ALT, alanine aminotransferase; IABP, intra-aortic balloon pump; CPR, cardiopulmonary resuscitation; ECMO, extracorporeal membrane oxygenation; AKI, acute kidney injury. Bold values indicate a P value < .05.

The Journal of Thoracic and Cardiovascular Surgery c Volume 154, Number 3

861

Acquired: Mechanical Circulatory Support

Abadeer et al

ACQ TABLE 6. Predictors of severe acute kidney injury (logistic regression) Characteristics

OR (95% CI)

P value

Age BMI Medical history CAD HLD HTN DM COPD Prior CVA CKD Preoperative status Cause 1-PCS 2-AMI* 3-Graft* 4-ADHF* 5-Other* MAP Hemoglobin Baseline creatinine ALT IABP Active CPR Device (CentriMag ¼ 0, ECMO ¼ 1)

1.00 (0.99-1.02) 1.02 (0.99-1.06)

.652 .241

1.30 (0.81-2.09) 1.39 (0.86-2.24) 1.51 (0.94-2.43) 1.17 (0.70-1.94) 0.78 (0.32-1.89) 4.08 (1.62-10.25) 1.92 (1.10-3.33)

.274 .180 .090 .551 .583 .003 .021

Reference 0.42 (0.22-0.80) 0.73 (0.33-1.60) 0.51 (0.25-1.05) 0.60 (0.27-1.30) 0.99 (0.98-1.00) 0.83 (0.74-0.93) 1.74 (1.29-2.35) 1.00 (1.00-1.00) 0.95 (0.59-1.52) 1.29 (0.72-2.34) 1.72 (1.06-2.78)

.089 .008 .431 .066 .194 .240 .001 .000 .037 .827 .392 .028

Multivariable OR (95% CI)

P value

1.25 (0.68-2.27) 1.04 (0.57-1.92)

.474 .893

3.10 (1.14-8.42) 1.42 (0.74-2.73)

.026 .290

Reference 0.45 (0.22-0.92) 0.55 (0.23-1.35) 0.44 (0.19-1.02) 0.55 (0.23-1.30)

.367 .028 .195 .056 .172

0.90 (0.80-1.03) 1.53 (1.10-2.13) 1.00 (1.00-1.00)

.130 .012 .067

1.60 (0.93-2.75)

.087

OR, Odds ratio; CI, confidence interval; BMI, body mass index; CAD, coronary artery disease; HLD, hyperlipidemia; HTN, hypertension; DM, diabetes mellitus; COPD, chronic obstructive pulmonary disease; CVA, cerebrovascular accident; CKD, chronic kidney disease; PCS, postcardiotomy shock; AMI, acute myocardial infarction; ADHF, acute decompensated heart failure; MAP, mean arterial pressure; ALT, alanine aminotransferase; IABP, intra-aortic balloon pump; CPR, cardiopulmonary resuscitation; ECMO, extracorporeal membrane oxygenation. *Compared with postcardiotomy shock. Bold values indicate a P value < .05.

perfusion before MCS, it is possible that ST-MCS with ECMO or VAD itself predisposes the damaged kidneys to further insult. Blood exposure to artificial surfaces causes systemic inflammation and hemoglobinuria-induced renal injury due to hemolysis in the extracorporeal circuit.15 Furthermore, complications associated with ST-MCS, such as bleeding and thromboembolism, are not uncommon and could have a distinct relationship with the development of AKI. Development of AKI in patients treated with ST-MCS has several critical implications in current clinical decision-making. First, when AKI develops immediately after MCS support initiation, there can be a discouraging clinical impression that such injury is associated with unacceptably high short-term mortality, which may negatively influence decision making (eg, patients with AKI might be offered palliative care earlier in the hospital course than patients without AKI). In addition, the uncertainty of renal recovery from AKI might influence the midterm patient care strategy (eg, patients with AKI might be considered ineligible for durable LVAD or heart transplantation). In fact, our study has confirmed that some of these shared clinical impressions have validity: Once patients have had severe AKI, the in-hospital mortality is high (>60%); proceeding with heart replacement therapy is rare (2.7% to 862

heart transplant and 4.5% to durable LVAD); therefore, myocardial recovery to allow successful ST-MCS explantation is often the only available survival option for these patients (27%). However, the current study also suggests that mild to moderate AKI could be temporary with no significant detrimental effect on patient survival. The evaluation of renal recovery in this study is of particular interest, because it is often unclear what chances, if any, there are for renal recovery after AKI (especially in patients requiring the initiation of RRT). In addition, the time frame for renal recovery in a cohort with such a tenuous hemodynamic profile has been a matter of significant debate. In addressing this issue, this study demonstrated that full renal recovery occurred in approximately 30% of patients who experienced some degree of AKI. In addition, we have found that recovery time differs widely by severity of AKI. Hospital survivors with nonsevere AKI can expect approximately 80% chance of renal recovery, which takes place within several days, whereas hospital survivors with severe AKI can expect approximately 60% chance of renal recovery, which occurs in a 3-week period. The current literature of patients treated with ECMO has shown that AKI occurs in approximately 50% of patients.16,17 In a recent study examining 78 adult patients treated with ECMO, Chen and colleagues18 found that

The Journal of Thoracic and Cardiovascular Surgery c September 2017

Acquired: Mechanical Circulatory Support

ACQ

Abadeer et al

according to RIFLE criteria, approximately 78% of patients experienced AKI. A meta-analysis of the complications of ECMO analyzed studies including 1866 patients and found that despite variations in criteria and staging, AKI occurred in 55.6% (35.5%-74.0%) of patients.16 The current study found the incidence of AKI to be consistent with that reported in the literature (60.5%). The poor outcomes of patients experiencing AKI have been reported in a number of studies outside the use of ST-MCS. In these studies, the acute increases in creatinine classifying AKI are associated with deleterious short- and long-term effects on survival to discharge and survival to 1 year.17,19 A few small studies have described the impact that renal injury plays in outcomes of ST-MCS, with the majority of the literature focusing on the outcomes of ECMO. Our study demonstrates confirmatory information that severe AKI within the first 7 postoperative days of ST-MCS for CS is associated with poor outcomes and is an independent predictor of late mortality. However, our study also suggests that milder forms of AKI do not result in worse outcomes. As such, it is possible that not all AKI should be treated equally when clinical decisions are made. Contrary to this, studies of patients undergoing general cardiac surgery have found that even small elevations in postoperative creatinine were associated with significantly increased rates of mortality.20-23 This might be explained by more profound acute illness, far higher inhospital mortality, and the limited sample size of the present study, all of which might have masked the potential detrimental effect of less severe forms of AKI. Two different types of ST-MCS, ECMO and ST-VAD, are included in our study, whereas previous studies have included only patients on ECMO. Each modality differs in the device characteristics, insertion technique, associated surgical insult, postoperative management, and the indications, and the analyses of the individual devices were previously reported.24 Our ST-VAD was inserted surgically, usually through a midline sternotomy with the use of cardiopulmonary bypass, whereas percutaneous ECMO is less invasive and used in sicker and more unstable patients. CS includes a certain spectrum of clinical presentation, and our strategy is to choose a device that seems most appropriate to the need of the individual patient. Focusing on a single device would miss the information on the other important arm of patients with CS. ECMO use was more frequent in the severe AKI group and was found to be an independent predictor of long-term mortality. Whether this is due to detrimental effect of the ECMO support or manifestation of the severity of illness of the ECMO-treated patients is beyond the scope of the current study. Baseline creatinine, but not chronic kidney disease, was found to be an independent predictor of severe AKI. Unlike the dichotomous chronic kidney disease variable, baseline

creatinine provides a continuous measure of renal function unrelated to acuity or chronicity of injury. In this study, every 1 mg/dL increase in preoperative creatinine increased the odds of severe AKI by 53%, likely indicating a certain postinsult sensitivity of the kidney to further injury. Prior CVA was the only other independently significant predictor in the development of severe AKI. The relationship between prior CVA and AKI is not readily apparent, but may represent a sensitivity of microvascular beds to ischemic injury. However, despite the microvascular changes associated with coronary artery disease, hyperlipidemia, hypertension, and diabetes, none of these were associated with the incidence of severe AKI. Also of note, none of the presupport hemodynamic or laboratory values predicted the severity of AKI. The acuity and dynamic nature of CS make it difficult for 1 variable to represent the composite insult to the kidney and the patient. Study Limitations Although this is one of the largest studies on AKI in STMCS for CS, there are a number of limitations in addition to the inherent limitations of a retrospective study. The incomplete availability of outpatient outcomes after hospital discharge including cause of death limits the temporal scope of this study. This study also is limited by the definitions and staging of AKI because AKI occurring before device insertion was not captured. Moreover, creatinine value determined while a patient is in CS may not be stable and may not represent a patient’s true ‘‘baseline’’ creatinine. Nonetheless, the classifications derived from the use of this value permitted meaningful discrimination of postimplant renal dysfunction. The creatinine cutoff values of the KDIGO criteria, although seemingly arbitrary, nonetheless appropriately stratify the degree of renal insult in an easily quantifiable manner and have been validated.13 Despite this limitation, the present study attempted to follow a clear definition of AKI that is well established and validated. The mixed cause of CS in the study cohort is another potential source of bias. In particular, cardiopulmonary bypass preceding CS in those with PCS and transplant graft dysfunction is an additive burden to kidney function, which could introduce a bias into our analysis. To mitigate these mechanisms of renal injury, subgroup analysis was performed focusing on the patients with AMI and acute decompensated heart failure. The incidence, outcomes, and risk factors were similar in the subgroups. Last, as stated earlier, the design of the current study does not address the possible deleterious effect of ST-MCS on renal function. Although the flow restoration by such devices should only contribute to improvement in renal perfusion and function, known or unknown characteristics related to the device-induced environment may potentiate renal impairment.

The Journal of Thoracic and Cardiovascular Surgery c Volume 154, Number 3

863

Acquired: Mechanical Circulatory Support

Abadeer et al

ACQ CONCLUSIONS AKI is common in patients with CS treated with STMCS. Severe AKI, but not less severe AKI, is a significant independent predictor of long-term mortality. Although the process may be slow, renal recovery is common, even after severe AKI. Conflict of Interest Statement Authors have nothing to disclose with regard to commercial support. References 1. Takayama H, Truby L, Koekort M, Uriel N, Colombo P, Mancini DM, et al. Clinical outcome of mechanical circulatory support for refractory cardiogenic shock in the current era. J Heart Lung Transplant. 2013;32:106-11. 2. Webb JG, Sanborn TA, Sleeper LA, Carere RG, Buller CE, Slater JN, et al. Percutaneous coronary intervention for cardiogenic shock in the SHOCK Trial Registry. Am Heart J. 2001;141:964-70. 3. Hochman JS, Buller CE, Sleeper LA, Boland J, Dzavik V, Sanborn TA, et al. Cardiogenic shock complicating acute myocardial infarction–etiologies, management and outcome: a report from the SHOCK Trial Registry. SHould we emergently revascularize Occluded Coronaries for cardiogenic shocK? J Am Coll Cardiol. 2000;36:1063-70. 4. Dauerman HL, Goldberg RJ, White K, Gore JM, Sadiq I, Gurfinkel E, et al. Revascularization, stenting, and outcomes of patients with acute myocardial infarction complicated by cardiogenic shock. Am J Cardiol. 2002;90:838-42. 5. Takayama H, Soni L, Kalesan B, Truby LK, Ota T, Cedola S, et al. Bridge-to-decision therapy with a continuous-flow external ventricular assist device in refractory cardiogenic shock of various causes. Circ Heart Fail. 2014;7:799-806. 6. O’Gara PT, Kushner FG, Ascheim DD, Casey DE Jr, Chung MK, de Lemos JA, et al. 2013 ACCF/AHA guideline for the management of ST-elevation myocardial infarction: a report of the American College of Cardiology Foundation/ American Heart Association Task Force on Practice Guidelines. Circulation. 2013;127:e362-425. 7. Mohite PN, Kaul S, Sabashnikov A, Rashid N, Fatullayev J, Zych B, et al. Extracorporeal life support in patients with refractory cardiogenic shock: keep them awake. Interact Cardiovasc Thorac Surg. 2015;20:755-60. 8. Schmidt M, Burrell A, Roberts L, Bailey M, Sheldrake J, Rycus PT, et al. Predicting survival after ECMO for refractory cardiogenic shock: the survival after veno-arterial-ECMO (SAVE)-score. Eur Heart J. 2015;36:2246-56. 9. Cooper HA, Panza JA. Cardiogenic shock. Cardiol Clin. 2013;31:567-80. viii. 10. Yancy CW, Jessup M, Bozkurt B, Butler J, Casey DE Jr, Drazner MH, et al. 2013 ACCF/AHA guideline for the management of heart failure: a report of the American College of Cardiology Foundation/American Heart Association Task Force on practice guidelines. Circulation. 2013;128:e240-327.

864

11. Chen YC, Tsai FC, Chang CH, Lin CY, Jenq CC, Juan KC, et al. Prognosis of patients on extracorporeal membrane oxygenation: the impact of acute kidney injury on mortality. Ann Thorac Surg. 2011;91:137-42. 12. Kiernan MS, Joseph SM, Katz JN, Kilic A, Rich JD, Tallman MP, et al. Sharing the care of mechanical circulatory support: collaborative efforts of patients/caregivers, shared-care sites, and left ventricular assist device implanting centers. Circ Heart Fail. 2015;8:629-35. 13. Levin A, Stevens PE. Summary of KDIGO 2012 CKD Guideline: behind the scenes, need for guidance, and a framework for moving forward. Kidney Int. 2014;85:49-61. 14. Thomas ME, Blaine C, Dawnay A, Devonald MA, Ftouh S, Laing C, et al. The definition of acute kidney injury and its use in practice. Kidney Int. 2015;87: 62-73. 15. Askenazi DJ, Selewski DT, Paden ML, Cooper DS, Bridges BC, Zappitelli M, et al. Renal replacement therapy in critically ill patients receiving extracorporeal membrane oxygenation. Clin J Am Soc Nephrol. 2012;7:1328-36. 16. Cheng R, Hachamovitch R, Kittleson M, Patel J, Arabia F, Moriguchi J, et al. Complications of extracorporeal membrane oxygenation for treatment of cardiogenic shock and cardiac arrest: a meta-analysis of 1,866 adult patients. Ann Thorac Surg. 2014;97:610-6. 17. Haft JW, Pagani FD, Romano MA, Leventhal CL, Dyke DB, Matthews JC. Shortand long-term survival of patients transferred to a tertiary care center on temporary extracorporeal circulatory support. Ann Thorac Surg. 2009;88:711-8. 18. Chen YC, Tsai FC, Fang JT, Yang CW. Acute kidney injury in adults receiving extracorporeal membrane oxygenation. J Formos Med Assoc. 2014;113:778-85. 19. Kandler K, Jensen ME, Nilsson JC, Moller CH, Steinbruchel DA. Acute kidney injury is independently associated with higher mortality after cardiac surgery. J Cardiothorac Vasc Anesth. 2014;28:1448-52. 20. Mangano CM, Diamondstone LS, Ramsay JG, Aggarwal A, Herskowitz A, Mangano DT. Renal dysfunction after myocardial revascularization: risk factors, adverse outcomes, and hospital resource utilization. The Multicenter Study of Perioperative Ischemia Research Group. Ann Intern Med. 1998;128:194-203. 21. Loef BG, Epema AH, Smilde TD, Henning RH, Ebels T, Navis G, et al. Immediate postoperative renal function deterioration in cardiac surgical patients predicts in-hospital mortality and long-term survival. J Am Soc Nephrol. 2005;16: 195-200. 22. Lassnigg A, Schmid ER, Hiesmayr M, Falk C, Druml W, Bauer P, et al. Impact of minimal increases in serum creatinine on outcome in patients after cardiothoracic surgery: do we have to revise current definitions of acute renal failure? Crit Care Med. 2008;36:1129-37. 23. Ho J, Reslerova M, Gali B, Nickerson PW, Rush DN, Sood MM, et al. Serum creatinine measurement immediately after cardiac surgery and prediction of acute kidney injury. Am J Kidney Dis. 2012;59:196-201. 24. Truby L, Mundy L, Kalesan B, Kirtane A, Colombo PC, Takeda K, et al. Contemporary outcomes of venoarterial extracorporeal membrane oxygenation for refractory cardiogenic shock at a large tertiary care center. ASAIO J. 2015;61: 403-9.

Key Words: kidney, mechanical circulatory support, renal function, renal insufficiency

The Journal of Thoracic and Cardiovascular Surgery c September 2017

Acquired: Mechanical Circulatory Support

ACQ

Abadeer et al

FIGURE E1. Adjusted survival function based on Cox estimates. AKI, Acute kidney injury; CAD, coronary artery disease; MAP, mean arterial pressure; IABP, intra-aortic balloon pump; CPR, cardiopulmonary resuscitation.

FIGURE E2. Subanalysis: long-term survival by severity of AKI. AKI, Acute kidney injury.

The Journal of Thoracic and Cardiovascular Surgery c Volume 154, Number 3

864.e1

Acquired: Mechanical Circulatory Support

Abadeer et al

ACQ TABLE E1. Device characteristics Characteristics

Overall

Patients (n) 293 Patient demographics and medical history n (%) Age, y 55.8  15.4 Age >65, y 86 (29.4) Male 199 (67.9) 28.3  6.8 BMI (kg/m2) CAD 136 (46.4) HLD 118 (40.3) HTN 159 (54.3) COPD 24 (8.2) DM 90 (30.7) Prior CVA 23 (7.9) CKD 68 (23.2) Cause of CS (%) PCS 91 (31.1) AMI 79 (27.0) Graft 36 (12.3) ADHF 51 (17.4) Other 37 (12.6) Preoperative hemodynamics SBP (mm Hg) 92.6  25.9 DBP (mm Hg) 51.2  21.1 MAP (mm Hg) 64.7  21.5 24-h postoperative hemodynamics SBP (mm Hg) 97.5  19.4 DBP (mm Hg) 71.0  17.5 MAP (mm Hg) 79.7  16.0 CVP (mm Hg) 11.3  4.4

P value

ECMO

VAD

163

130

57.0  16.3 54 (33.1) 110 (67.5) 28.9  6.9 76 (46.6) 64 (39.3) 92 (56.4) 15 (9.2) 47 (28.8) 13 (8.0) 32 (19.6)

54.1  14.0 32 (24.6) 89 (68.5) 27.5  6.6 60 (46.2) 54 (41.5) 67 (51.5) 9 (6.9) 43 (33.1) 10 (7.7) 36 (27.7)

.114 .112 .859 .081 .94 .693 .403 .480 .434 .929 .104

61 (37.4) 45 (27.6) 16 (9.8) 20 (12.3) 21 (12.9)

30 (23.1) 34 (26.2) 20 (15.4) 31 (23.9) 16 (12.3)

.008 .781 .149 .009 .883

87.3  27.8 44.7  21.6 59.1  22.5

99.9  21.0 60.3  16.6 72.6  17.2

.0003 .0000 .0000

98.8  20.8 67.3  16.4 77.4  16.3 10.9  4.2

95.5  17.0 76.8  17.6 82.7  15.0 11.9  4.5

.218 .0001 .0092 .080

Data are presented as mean  standard deviation or median (IQR, Q1-Q3), unless otherwise specified. P values represent a comparison of the severe with the nonsevere AKI group. ECMO, Extracorporeal membrane oxygenation; VAD, ventricular assist device; BMI, body mass index; CAD, coronary artery disease; HLD, hyperlipidemia; HTN, hypertension; COPD, chronic obstructive pulmonary disease; DM, diabetes mellitus; CVA, cerebrovascular accident; CKD, chronic kidney disease; CS, cardiogenic shock; PCS, postcardiotomy shock; AMI, acute myocardial infarction; ADHF, acute decompensated heart failure; SBP, systolic blood pressure; DBP, diastolic blood pressure; MAP, mean arterial pressure; CVP, central venous pressure. Bold values indicate a P value < .05.

TABLE E2. Comparison of discharge creatinine with initial severity of acute kidney injury among survivors* Characteristics

Total n ¼ 72

Nonsevere n ¼ 34

Severe n ¼ 38

P value

<1.5 3 baseline Cr (renal recovery) 1.5-< 2.0 3 baseline Cr 2.0-< 3.0 3 baseline Cr >3 3 baseline Cr Discharged to dialysis

52 (72.2) 5 (6.9) 2 (2.8) 2 (2.8) 11 (15.3)

28 (82.4) 3 (8.8) 0 (0.0) 0 (0.0) 3 (8.8)

24 (63.2) 2 (5.3) 2 (5.3) 2 (5.3) 8 (21.1)

.069 .55 .18 .18 .15

Cr, Creatinine. *Among patients who experienced AKI and survived to discharge. Bold values indicate a P value < .05.

864.e2

The Journal of Thoracic and Cardiovascular Surgery c September 2017

Acquired: Mechanical Circulatory Support

ACQ

Abadeer et al

TABLE E3. Subanalysis: Outcomes Characteristics Length of support Hospital stay ICU stay In-hospital mortality 30-d mortality Destinations, n % Myocardial recovery Transition to VAD (short-term and durable) OHT

Overall n ¼ 130

Nonsevere n ¼ 90

Severe n ¼ 40

P value

7 (3-20) 38.5 (15-70) 15 (5-34) 66 (50.8) 47 (36.2)

8.5 (4-21) 41 (24-69) 22 (7-35) 39 (43.3) 24 (26.7)

5 (2.2-16.6) 30.5 (8-78.5) 8.5 (4-27) 27 (67.5) 23 (57.5)

.039 .24 .22 .011 .001

29 (22.7) 24 (18.9) 18 (13.9)

22 (25.0) 19 (21.8) 16 (17.8)

7 (17.5) 5 (12.5) 2 (5.0)

.347 .212 .052

Data are presented as median (IQR) or n (%). P values represent a comparison of the severe with the nonsevere group. ICU, Intensive care unit; VAD, ventricular assist device; OHT, orthotopic heart transplant. Bold values indicate a P value < .05.

TABLE E4. Subanalysis: Predictors of long-term mortality (Cox) Characteristics

OR (95% CI)

P value

Multivariable HR (95% CI)

P value

Age BMI Medical history CAD HLD HTN DM COPD Prior CVA CKD Preoperative status MAP Hemoglobin Baseline creatinine ALT IABP Active CPR Device (CentriMag ¼ 0, ECMO ¼ 1) Postoperative status Severe AKI

1.02 (1.00-1.04) 1.00 (0.97-1.04)

.04 .77

1.00 (0.98-1.03)

.51

1.19 (0.74-1.91) 0.83 (0.51-1.36) 1.20 (0.75-1.91) 1.12 (0.70-1.81) 1.11 (0.45-2.77) 2.09 (0.99-4.39) 0.99 (0.56-1.75)

.47 .46 .44 .63 .81 .052 .97

0.95 (0.38-2.41)

.92

0.99 (0.98-1.00) 0.96 (0.87-1.06) 1.10 (1.00-1.21) 1.00 (1.00-1.00) 0.70 (0.44-1.12) 1.65 (0.97-2.83) 2.24 (1.39-3.61)

.080 .43 .036 .027 .138 .065 .001

0.98 (0.96-1.00)

.066

1.08 (0.95-1.23) 1.00 (0.99-1.00) .52 (0.28-0.95) .46 (0.14-1.45) 1.75 (0.91-3.36)

.26 .36 .035 .18 .10

2.02 (1.25-3.27)

.004

2.26 (1.23-4.17)

.009

OR, Odds ratio; CI, confidence interval; HR, hazard ratio; BMI, body mass index; CAD, coronary artery disease; HLD, hyperlipidemia; HTN, hypertension; DM, diabetes mellitus; COPD, chronic obstructive pulmonary disease; CVA, cerebrovascular accident; CKD, chronic kidney disease; MAP, mean arterial pressure; ALT, alanine aminotransferase; IABP, intra-aortic balloon pump; CPR, cardiopulmonary resuscitation; ECMO, extracorporeal membrane oxygenation; AKI, acute kidney injury. Bold values indicate a P value<.05.

The Journal of Thoracic and Cardiovascular Surgery c Volume 154, Number 3

864.e3

Acquired: Mechanical Circulatory Support

Abadeer et al

ACQ TABLE E5. Subanalysis: Predictors of severe acute kidney injury (logistic) Characteristics

OR (95% CI)

P value

Age BMI Medical history CAD HLD HTN DM COPD Prior CVA CKD Preoperative status MAP Hemoglobin Baseline creatinine ALT IABP Active CPR Device (CentriMag ¼ 0, ECMO ¼ 1)

1.00 (0.97-1.03) 1.03 (0.98-1.08)

.93 .24

2.04 (0.92-4.51) 1.41 (0.66-3.02) 2.28 (1.06-4.90) 1.28 (0.60-2.73) 1.89 (0.48-7.44) 6.15 (1.50-25.2) 1.75 (0.73-4.22)

.078 .38 .035 .53 .36 .012 .21

0.99 (0.97-1.00) 0.91 (0.78-1.07) 2.01 (1.29-3.13) 1.00 (1.00-1.00) 0.81 (0.38-1.71) 1.33 (0.55-3.22) 2.32 (1.07-5.02)

.27 .25 .002 .55 .58 .52 .032

Multivariable OR (95% CI)

P value

1.83 (0.75-4.49)

.18

1.42 (0.59-3.43)

.44

2.77 (0.58-13.19)

.20

1.67 (1.05-2.69)

.032

1.90 (0.82-4.39)

.13

OR, Odds ratio; CI, confidence interval; BMI, body mass index; CAD, coronary artery disease; HLD, hyperlipidemia; HTN, hypertension; DM, diabetes mellitus; COPD, chronic obstructive pulmonary disease; CVA, cerebrovascular accident; CKD, chronic kidney disease; MAP, mean arterial pressure; ALT, alanine aminotransferase; IABP, intra-aortic balloon pump; CPR, cardiopulmonary resuscitation; ECMO, extracorporeal membrane oxygenation. Bold values indicate a P value < .05.

864.e4

The Journal of Thoracic and Cardiovascular Surgery c September 2017