European Journal of Cardio-Thoracic Surgery Advance Access published October 13, 2014
Erythropoietin administration for prevention of cardiac surgery-associated acute kidney injury: a meta-analysis of randomized controlled trials Hong-Tao Tiea, Ming-Zhu Luob, Dan Linb, Min Zhanga, Jing-Yuan Wanb,* and Qing-Chen Wua,* a b
Department of Cardiothoracic Surgery, The First Affiliated Hospital of Chongqing Medical University, Chongqing, China Chongqing Key Laboratory of Biochemistry and Molecular Pharmacology, Chongqing Medical University, Chongqing, China
* Corresponding authors. Chongqing Key Laboratory of Biochemistry and Molecular Pharmacology, Chongqing Medical University, Chongqing, China. Tel: +86-23-68485038; fax: +86-23-86134172; e-mail: [email protected] ( J.-Y.W.); Department of Cardiothoracic Surgery, The First Affiliated Hospital of Chongqing Medical University, Chongqing, China. Tel: +86-13708352012; fax: +86-23-89011138; e-mail: [email protected] (Q.W.). Received 15 April 2014; received in revised form 7 August 2014; accepted 26 August 2014
Abstract The effect of erythropoietin (EPO) on the prevention of cardiac surgery-associated acute kidney injury (CSA-AKI) is controversial. Therefore, we undertook the meta-analysis of randomized controlled trials (RCTs) to assess the efficacy and safety of EPO on the prevention of CSA-AKI in adult patients and to explore whether risk factors for AKI could explain the inconsistent effects. PubMed and EMbase databases were searched to identify eligible RCTs. The meta-analysis was performed with fixed- or random-effects models according to the heterogeneity, and the subgroup analysis stratified by risk factors for AKI was carried out. Five RCTs involving 423 patients were included. Overall, EPO administration was not associated with a reduced incidence of CSA-AKI [relative risk (RR): 0.64, 95% confidence interval (CI): 0.35–1.16], with a moderate heterogeneity (I2 = 67.4%, heterogeneity P = 0.02). Subgroup analysis showed that, in patients without high risk factors for AKI, EPO administration could significantly reduce the incidence of CSA-AKI (RR: 0.38, 95% CI: 0.24–0.61), intensive care unit length of stay [standardized mean difference (SMD): −0.54, 95% CI: −1.05 to −0.04] and hospital length of stay (SMD: −0.48, 95% CI: −0.94 to −0.02). The test of heterogeneity was not significant in the two subgroups. EPO administration could significantly reduce the incidence of CSA-AKI, but not in patients with high risk factors for AKI. Substantial heterogeneity across trials could be attributed to high risk factors for AKI. However, our findings should be interpreted cautiously because of the limited studies included, and highquality RCTs are warranted. Keywords: Acute kidney injury • Cardiac surgery • Erythropoietin • Meta-analysis
INTRODUCTION Acute kidney injury (AKI), a serious postoperative complication, is especially frequent in patients undergoing cardiac surgery and specifically termed cardiac surgery-associated AKI (CSA-AKI). Owing to a lack of uniform definition of CSA-AKI, the incidence rate of CSA-AKI was reported to vary from 8.9 to 39% [1–3]. Renal replacement therapy (RRT) occurred in approximately 1–5% of patients with CSA-AKI [4]. Additionally, CSA-AKI is associated with increases in intensive care unit (ICU) length of stay [5], hospital length of stay (HLOS) [2, 5, 6], medical resources and hospital mortality [5–8]. Considering the poor prognosis and increased medical cost, the prevention of CSA-AKI is of great importance. Despite extensive research carried out to prevent CSA-AKI [5, 9, 10], the best prophylaxis for CSA-AKI remains to be established [11, 12]. Erythropoietin (EPO) is mainly produced by the renal cortex and could regulate haematopoiesis. In recent years, several experiments demonstrated that EPO had properties of anti-apoptosis, antioxidative stress and anti-inflammation, and consequently could decrease the incidence of CSA-AKI [13–15]. To verify this promising finding, several randomized controlled trials (RCTs) [16–21] have
been well performed in adult patients undergoing cardiac surgery; nevertheless, they present controversial results. Therefore, we performed the current meta-analysis of RCTs to evaluate the efficacy and safety of EPO administration on the prevention of CSA-AKI in patients undergoing cardiac surgery. Patients with high risk factors for AKI often have concurrent diseases and need to take many medications, some of which are toxic to the kidney [22]. Hence, we also tried to explore whether high risk factors for AKI could explain the inconsistent effects of EPO on CSA-AKI prevention.
MATERIALS AND METHODS The meta-analysis was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) statement [23].
Literature search and selection criteria Two trained reviewers (Hong-Tao Tie and Ming-Zhu Luo) independently conducted the literature search to identify eligible RCTs
© The Author 2014. Published by Oxford University Press on behalf of the European Association for Cardio-Thoracic Surgery. All rights reserved.
REVIEW
REVIEW
European Journal of Cardio-Thoracic Surgery (2014) 1–8 doi:10.1093/ejcts/ezu378
2
H.-T. Tie et al. / European Journal of Cardio-Thoracic Surgery
comparing the effect of EPO versus placebo on the prevention of CSA-AKI. PubMed and EMbase databases were searched from inception to 20 January 2014, using the following terms: (epoetin OR erythropoietin OR EPO) AND (acute kidney injury OR AKI) AND (cardiac surgery OR heart surgery). No limitation was imposed. Moreover, the reference lists of retrieved studies were also manually scanned to include additional eligible studies. This process was performed iteratively until no further study was identified. All included studies were finally approved by consensus. The inclusion criteria were as follows: (i) study population: adult patients undergoing cardiac surgery; (ii) intervention: EPO; (iii) comparative intervention: placebo; (iv) outcome: the incidence of CSA-AKI and (v) study design: RCT.
by Q-statistic and I2 statistics: a fixed-effect model was adopted if I2 < 50% and heterogeneity P > 0.10; otherwise a random-effects model was used. Subgroup analysis stratified by high risk factors for AKI was performed according to our protocol. Publication bias was not assessed because 18 years), of whom the patients in two trials [16, 17] were at risk for AKI and patients in one trial [21] had preoperative anaemia. The surgery type only involved coronary artery bypass grafting in two studies [19, 20], while it included valvular surgery or complex valvular surgery in the other three [16, 17, 21]. All participants in four studies [16, 17, 20, 21] underwent cardiac surgery with extracorporeal circulation (CPB), while only some patients used CPB in the remaining one [19]. The interventions of EPO and placebo were intravenously administered before the surgery in all trials except one [16]. Four studies [17, 19–21] defined CSA-AKI on the basis of an increase in serum creatinine, while the remaining one [16] was defined according to the Acute Kidney Injury Network classification. Of these five studies, all [16, 17, 19–21] reported the incidence of AKI, ICU length of stay, HLOS and hospital mortality; three [17, 20, 21] reported the incidence of arrhythmia; and three [17, 19, 21] reported perioperative transfusion requirement. All but one [21] took AKI as the primary outcome. Each of the included trials was a well-performed RCT with a Jadad score of ≥4.
The primary outcome: acute kidney injury The pooled estimate of the five RCTs suggested that EPO administration was not associated with a significant reduction in the incidence of AKI, compared with controls (RR: 0.64, 95% CI: 0. 35–1.16; P = 0.14), with a moderate heterogeneity among the studies (I2 = 67%, heterogeneity P = 0.02). Subsequently, we performed subgroup analysis to identify potential sources of heterogeneity. When stratified by risk factors for AKI, the heterogeneity was solved in both
3
REVIEW
H.-T. Tie et al. / European Journal of Cardio-Thoracic Surgery
Figure 1: Flow chart of screening RCTs included in the meta-analysis.
subgroups as shown in (Fig. 2). Compared with controls, EPO administration could significantly reduce the incidence of AKI in patients without high risk factors for AKI (RR: 0.38, 95% CI: 0.24– 0.61; P < 0.0001; I2 = 0%; heterogeneity P = 0.81), but not in patients with high risk factors for AKI (RR: 1.22, 95% CI: 0.77–1.94; P = 0.41; I2 = 0%; heterogeneity P = 0.44).
Secondary outcomes In patients without high risk factors for AKI, EPO administration correlated with significant decreases in ICU length of stay (SMD: −0.54, 95% CI: −1.05 to −0.04; P = 0.034; I2 = 73.8%; heterogeneity P = 0.022; Fig. 3), HLOS (SMD: −0.48, 95% CI: −0.94 to −0.02; P = 0.04; I2 = 69%; heterogeneity P = 0.04; Fig. 4). While in patients with high risk factors for AKI, EPO administration was not associated with reduced ICU length of stay (SMD: 0.57, 95% CI: −0.26 to 1.41; P = 0.18; I2 = 88%; heterogeneity P = 0.0003; Fig. 3). On the contrary, it was even associated with a significant increment in HLOS (SMD: 0.35, 95% CI: 0.07–0.62; P = 0.01; I2 = 0%; heterogeneity P = 0.79; Fig. 4). EPO administration was not associated with a significant reduction in hospital mortality in patients without high risk factors for AKI (RR: 0.27, 95% CI: 0.05–1.62; P = 0.15; I2 = 0%; heterogeneity P = 0.97; Fig. 5) or in patients with high risk factors for AKI (RR: 1.00, 95% CI: 0.24–4.25; P = 1.00; I2 = 0%; heterogeneity P = 0.41; Fig. 5). Overall, EPO administration had no impact on the incidence of arrhythmia (RR: 0.71, 95% CI: 0.43–1.17; P = 0.18; I2 = 0%; heterogeneity P = 0.90; Supplementary Fig. S1) or perioperative transfusion requirement (SMD: −0.56, 95% CI: −1.25 to
0.13; P = 0.11; I2 = 86%; heterogeneity P = 0.001; Supplementary Fig. S2).
DISCUSSION To our knowledge, this is the first meta-analysis that evaluates the efficacy and safety of EPO on the prevention of CSA-AKI in patients undergoing cardiac surgery. Our findings suggest that, in patients without high risk factors for AKI, EPO administration could effectively decrease the incidence of AKI, ICU length of stay and HLOS. However, EPO administration failed to show efficacy on the prevention of CSA-AKI or shorten the ICU length of stay and HLOS in patients with high risk factors for AKI. Furthermore, EPO administration was not associated with perioperative transfusion requirement, hospital mortality or the incidence of arrhythmia. The pathogenesis of CSA-AKI is multifactorial, including exogenous and endogenous toxins, metabolic factors, ischaemia and reperfusion (IR), neurohormonal activation, inflammation and oxidative stress resulting from hypoxia and ischaemia [22]. Therefore, the mechanisms by which EPO decreases the incidence of CSA-AKI in patients without high risk for AKI could be explained by the following reasons. First, EPO was shown to protect against IR injury and have a reno-protective effect [19]. Both of them could contribute to the preventive effect of EPO on CSA-AKI. Secondly, EPO could inhibit cell apoptosis via down-regulating the ratio of Bax/Bcl-2 and decreasing superoxide production [30, 31]. Thirdly, EPO, as a cytokine for haematopoiesis, could raise the
4
Table 1: Baseline characteristics of the five eligible randomized controlled trials Study
Yoo et al. [21]
De Seigneux et al. [16]
Kim et al. [17]
Tasanarong et al. [20]
Surgery type
EPO
Administration Control time
Study design/Jadad score
CABG/CPB or OP
300 U/kg EPO, IV
NS, IV
Before surgery
R, DB, PC /5
NA/NA; preoperative anaemia At risk for AKIa; NA/NA/NA
Valvular heart surgery/CPB
500 IU/kg rHuEPO+100 ml NS NS, IV with 200 mg iron, IV
Before surgery
R, SB, PC/4
Cardiac surgery/CPB
α-Epoetin: 20 000 ui or α-Epoetin: 40 000 ui, IV
NS, IV
After surgery
R, DB, PC/5
At risk for AKIb; NA/NA
Valvular heart surgery/CPB
300 IU/kg rHuEPO-α, IV
NS, IV
Before surgery
R, DB, PC/5
NA/NA; NA/NA
CABG/CPB
200 U/kg rHuEPO 3 d before operation+100 U/kg rHuEPO at surgery, IV
NS, IV
Perioperative surgery
R, DB, PC/5
Sample; patients age (years); male (%)
BMI (Kg/m2); Hypertension (%);
eGFR(ml/min/1.73 m2); Hb(g/dl) ; SCr (mg/dl)
Risk for AKI; others
71 (36/35); 64.6 ± 10.7/ 68.9 ± 8.4; 83%/ 66% 74 (37/37); 56 ± 12/59 ± 12; 65%/62% 80 (40/20/20); 68.9 ± 12.0/ 66.5 ± 16.5/ 64.7 ± 14.7; 80%/65%/67.5% 98 (49/49); 63 ± 10/62 ± 10; 41%/49% 100 (50/50); 63 ± 16/60 ± 16; 62%/52%
24.7 ± 3.3/ 23.5 ± 2.3; 83%/ 64%
61.60 ± 25.20/59.30 ± 21.70; 13.10 ± 2.30/12.60 ± 1.60; 1.20 ± 0.38/1.08 ± 0.32
NA/NA; NA/NA
23 ± 3/23 ± 4; 24.3%/37.8%
NA/NA; 11.80 ± 0.80/11.60 ± 1.20; NA/NA;
NA/NA/NA; 85%/60%/65%
NA/ NA/NA; NA/NA/NA; 1.05 ± 0.29/0.98 ± 0.27/ 0.96 ± 0.30
24 ± 3/24 ± 4; 55%/49%
83 ± 28/77 ± 30; NA/NA; 0.92 ± 0.28/0.99 ± 0.27 64 ± 29/67 ± 33; 12.30 ± 1.7/12.20 ± 1.90; 1.05 ± 0.27/1.05 ± 0.45
NA/NA; 79%/63%
Intervention
AKI: acute kidney injury; CABG: coronary artery bypass grafting; CPB: cardiopulmonary bypass; OP: off-pump; IV: intravenous administration; NS: normal saline; eGFR: estimated glomerular filtration rate; Hb: haemoglobin; SCr: serum creatinine; R: randomized; DB: double-blinded; SB: single-blinded ; PC: placebo-controlled; Defined as sepsis, postoperative state, mechanical ventilation, haemodynamic impairment and previous chronic disease. b Defined as involving more than two of the following criteria: female, age >65 years, New York Heart Association (NYHA) functional class IV, left ventricular ejection fraction 1.2 mg/dl or peripheral vascular disease. a
H.-T. Tie et al. / European Journal of Cardio-Thoracic Surgery
Song et al. [19]
Population (EPO/Control)
SCr: serum creatinine; HLOS: hospital length of stay; gr1: group 1, patient receiving α-Epoetin: 20 000 ui; gr2: group 2, patient receiving α-Epoetin: 40 000 ui.
Tasanarong et al. [20]
De Seigneux et al. [16] Kim et al. [17]
Yoo et al. [21]
5
level of erythrocyte anti-oxidative enzymes by increasing the circulating young red blood cells [32]. However, whether the antiinflammation property of EPO plays an important role in this protection mechanism is still controversial [16, 17, 33]. In our subgroup analysis, the pooled results suggested that EPO administration could significantly reduce the incidence of CSA-AKI in patients without high risk factors for AKI [19–21], but not in patients with high risk factors for AKI [16, 17]. Thus, high risk factors seemed to cause the disagreement among current evidence. However, the definition of high-risk factors for AKI was not standardized in the included studies. The study conducted by De Seigneux et al. [16] regarded risk factors for AKI as sepsis, postoperative state, mechanical ventilation, haemodynamic impairment and previous chronic disease, while the other study by Kim et al. [17] defined it as involving more than two of the following criteria: female, age >65 years, New York Heart Association (NYHA) functional class IV, left ventricular ejection fraction < 35%, diabetes mellitus, chronic obstructive pulmonary disease (COPD), preoperative creatinine >1.2 mg/dl or peripheral vascular disease. According to previous published studies, there are several potential explanations for the ineffectiveness of EPO in patients with high risk factors for AKI. First, EPO is only beneficial to prevent AKI before ischaemic injury rather than after such injury [16, 19]. The kidney of patients having high risk factors for AKI might have been injured because renal dysfunction occurs several hours after the kidney injury and current diagnostic methods lack specificity [10]. Moreover, EPO could only protect against injury resulting from ischaemia but not inflammation [16, 17]. Consistently, patients with high risk factors for AKI usually present with some inflammation-associated diseases. Furthermore, many medicines, toxic to kidneys (e.g. beta-lactam, aminoglycoside and amphotericin), are usually required for in these patients [22], and drug-induced renal injury may not be mitigated by EPO. Our meta-analysis additionally revealed that EPO administration could significantly reduce the ICU length of stay and the HLOS in patients without high risk factors for AKI, which could consequently conserve healthcare resources. However, the benefits of EPO disappeared and HLOS even lengthened in patients with high risk factors for AKI. Since high risk factors for AKI (including NYHA IV, LVEF
Erythropoietin administration for prevention of cardiac surgery-associated acute kidney injury: a meta-analysis of randomized controlled trials Hong-Tao Tiea, Ming-Zhu Luob, Dan Linb, Min Zhanga, Jing-Yuan Wanb,* and Qing-Chen Wua,* a b
Department of Cardiothoracic Surgery, The First Affiliated Hospital of Chongqing Medical University, Chongqing, China Chongqing Key Laboratory of Biochemistry and Molecular Pharmacology, Chongqing Medical University, Chongqing, China
* Corresponding authors. Chongqing Key Laboratory of Biochemistry and Molecular Pharmacology, Chongqing Medical University, Chongqing, China. Tel: +86-23-68485038; fax: +86-23-86134172; e-mail: [email protected] ( J.-Y.W.); Department of Cardiothoracic Surgery, The First Affiliated Hospital of Chongqing Medical University, Chongqing, China. Tel: +86-13708352012; fax: +86-23-89011138; e-mail: [email protected] (Q.W.). Received 15 April 2014; received in revised form 7 August 2014; accepted 26 August 2014
Abstract The effect of erythropoietin (EPO) on the prevention of cardiac surgery-associated acute kidney injury (CSA-AKI) is controversial. Therefore, we undertook the meta-analysis of randomized controlled trials (RCTs) to assess the efficacy and safety of EPO on the prevention of CSA-AKI in adult patients and to explore whether risk factors for AKI could explain the inconsistent effects. PubMed and EMbase databases were searched to identify eligible RCTs. The meta-analysis was performed with fixed- or random-effects models according to the heterogeneity, and the subgroup analysis stratified by risk factors for AKI was carried out. Five RCTs involving 423 patients were included. Overall, EPO administration was not associated with a reduced incidence of CSA-AKI [relative risk (RR): 0.64, 95% confidence interval (CI): 0.35–1.16], with a moderate heterogeneity (I2 = 67.4%, heterogeneity P = 0.02). Subgroup analysis showed that, in patients without high risk factors for AKI, EPO administration could significantly reduce the incidence of CSA-AKI (RR: 0.38, 95% CI: 0.24–0.61), intensive care unit length of stay [standardized mean difference (SMD): −0.54, 95% CI: −1.05 to −0.04] and hospital length of stay (SMD: −0.48, 95% CI: −0.94 to −0.02). The test of heterogeneity was not significant in the two subgroups. EPO administration could significantly reduce the incidence of CSA-AKI, but not in patients with high risk factors for AKI. Substantial heterogeneity across trials could be attributed to high risk factors for AKI. However, our findings should be interpreted cautiously because of the limited studies included, and highquality RCTs are warranted. Keywords: Acute kidney injury • Cardiac surgery • Erythropoietin • Meta-analysis
INTRODUCTION Acute kidney injury (AKI), a serious postoperative complication, is especially frequent in patients undergoing cardiac surgery and specifically termed cardiac surgery-associated AKI (CSA-AKI). Owing to a lack of uniform definition of CSA-AKI, the incidence rate of CSA-AKI was reported to vary from 8.9 to 39% [1–3]. Renal replacement therapy (RRT) occurred in approximately 1–5% of patients with CSA-AKI [4]. Additionally, CSA-AKI is associated with increases in intensive care unit (ICU) length of stay [5], hospital length of stay (HLOS) [2, 5, 6], medical resources and hospital mortality [5–8]. Considering the poor prognosis and increased medical cost, the prevention of CSA-AKI is of great importance. Despite extensive research carried out to prevent CSA-AKI [5, 9, 10], the best prophylaxis for CSA-AKI remains to be established [11, 12]. Erythropoietin (EPO) is mainly produced by the renal cortex and could regulate haematopoiesis. In recent years, several experiments demonstrated that EPO had properties of anti-apoptosis, antioxidative stress and anti-inflammation, and consequently could decrease the incidence of CSA-AKI [13–15]. To verify this promising finding, several randomized controlled trials (RCTs) [16–21] have
been well performed in adult patients undergoing cardiac surgery; nevertheless, they present controversial results. Therefore, we performed the current meta-analysis of RCTs to evaluate the efficacy and safety of EPO administration on the prevention of CSA-AKI in patients undergoing cardiac surgery. Patients with high risk factors for AKI often have concurrent diseases and need to take many medications, some of which are toxic to the kidney [22]. Hence, we also tried to explore whether high risk factors for AKI could explain the inconsistent effects of EPO on CSA-AKI prevention.
MATERIALS AND METHODS The meta-analysis was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) statement [23].
Literature search and selection criteria Two trained reviewers (Hong-Tao Tie and Ming-Zhu Luo) independently conducted the literature search to identify eligible RCTs
© The Author 2014. Published by Oxford University Press on behalf of the European Association for Cardio-Thoracic Surgery. All rights reserved.
REVIEW
REVIEW
European Journal of Cardio-Thoracic Surgery (2014) 1–8 doi:10.1093/ejcts/ezu378
2
H.-T. Tie et al. / European Journal of Cardio-Thoracic Surgery
comparing the effect of EPO versus placebo on the prevention of CSA-AKI. PubMed and EMbase databases were searched from inception to 20 January 2014, using the following terms: (epoetin OR erythropoietin OR EPO) AND (acute kidney injury OR AKI) AND (cardiac surgery OR heart surgery). No limitation was imposed. Moreover, the reference lists of retrieved studies were also manually scanned to include additional eligible studies. This process was performed iteratively until no further study was identified. All included studies were finally approved by consensus. The inclusion criteria were as follows: (i) study population: adult patients undergoing cardiac surgery; (ii) intervention: EPO; (iii) comparative intervention: placebo; (iv) outcome: the incidence of CSA-AKI and (v) study design: RCT.
by Q-statistic and I2 statistics: a fixed-effect model was adopted if I2 < 50% and heterogeneity P > 0.10; otherwise a random-effects model was used. Subgroup analysis stratified by high risk factors for AKI was performed according to our protocol. Publication bias was not assessed because 18 years), of whom the patients in two trials [16, 17] were at risk for AKI and patients in one trial [21] had preoperative anaemia. The surgery type only involved coronary artery bypass grafting in two studies [19, 20], while it included valvular surgery or complex valvular surgery in the other three [16, 17, 21]. All participants in four studies [16, 17, 20, 21] underwent cardiac surgery with extracorporeal circulation (CPB), while only some patients used CPB in the remaining one [19]. The interventions of EPO and placebo were intravenously administered before the surgery in all trials except one [16]. Four studies [17, 19–21] defined CSA-AKI on the basis of an increase in serum creatinine, while the remaining one [16] was defined according to the Acute Kidney Injury Network classification. Of these five studies, all [16, 17, 19–21] reported the incidence of AKI, ICU length of stay, HLOS and hospital mortality; three [17, 20, 21] reported the incidence of arrhythmia; and three [17, 19, 21] reported perioperative transfusion requirement. All but one [21] took AKI as the primary outcome. Each of the included trials was a well-performed RCT with a Jadad score of ≥4.
The primary outcome: acute kidney injury The pooled estimate of the five RCTs suggested that EPO administration was not associated with a significant reduction in the incidence of AKI, compared with controls (RR: 0.64, 95% CI: 0. 35–1.16; P = 0.14), with a moderate heterogeneity among the studies (I2 = 67%, heterogeneity P = 0.02). Subsequently, we performed subgroup analysis to identify potential sources of heterogeneity. When stratified by risk factors for AKI, the heterogeneity was solved in both
3
REVIEW
H.-T. Tie et al. / European Journal of Cardio-Thoracic Surgery
Figure 1: Flow chart of screening RCTs included in the meta-analysis.
subgroups as shown in (Fig. 2). Compared with controls, EPO administration could significantly reduce the incidence of AKI in patients without high risk factors for AKI (RR: 0.38, 95% CI: 0.24– 0.61; P < 0.0001; I2 = 0%; heterogeneity P = 0.81), but not in patients with high risk factors for AKI (RR: 1.22, 95% CI: 0.77–1.94; P = 0.41; I2 = 0%; heterogeneity P = 0.44).
Secondary outcomes In patients without high risk factors for AKI, EPO administration correlated with significant decreases in ICU length of stay (SMD: −0.54, 95% CI: −1.05 to −0.04; P = 0.034; I2 = 73.8%; heterogeneity P = 0.022; Fig. 3), HLOS (SMD: −0.48, 95% CI: −0.94 to −0.02; P = 0.04; I2 = 69%; heterogeneity P = 0.04; Fig. 4). While in patients with high risk factors for AKI, EPO administration was not associated with reduced ICU length of stay (SMD: 0.57, 95% CI: −0.26 to 1.41; P = 0.18; I2 = 88%; heterogeneity P = 0.0003; Fig. 3). On the contrary, it was even associated with a significant increment in HLOS (SMD: 0.35, 95% CI: 0.07–0.62; P = 0.01; I2 = 0%; heterogeneity P = 0.79; Fig. 4). EPO administration was not associated with a significant reduction in hospital mortality in patients without high risk factors for AKI (RR: 0.27, 95% CI: 0.05–1.62; P = 0.15; I2 = 0%; heterogeneity P = 0.97; Fig. 5) or in patients with high risk factors for AKI (RR: 1.00, 95% CI: 0.24–4.25; P = 1.00; I2 = 0%; heterogeneity P = 0.41; Fig. 5). Overall, EPO administration had no impact on the incidence of arrhythmia (RR: 0.71, 95% CI: 0.43–1.17; P = 0.18; I2 = 0%; heterogeneity P = 0.90; Supplementary Fig. S1) or perioperative transfusion requirement (SMD: −0.56, 95% CI: −1.25 to
0.13; P = 0.11; I2 = 86%; heterogeneity P = 0.001; Supplementary Fig. S2).
DISCUSSION To our knowledge, this is the first meta-analysis that evaluates the efficacy and safety of EPO on the prevention of CSA-AKI in patients undergoing cardiac surgery. Our findings suggest that, in patients without high risk factors for AKI, EPO administration could effectively decrease the incidence of AKI, ICU length of stay and HLOS. However, EPO administration failed to show efficacy on the prevention of CSA-AKI or shorten the ICU length of stay and HLOS in patients with high risk factors for AKI. Furthermore, EPO administration was not associated with perioperative transfusion requirement, hospital mortality or the incidence of arrhythmia. The pathogenesis of CSA-AKI is multifactorial, including exogenous and endogenous toxins, metabolic factors, ischaemia and reperfusion (IR), neurohormonal activation, inflammation and oxidative stress resulting from hypoxia and ischaemia [22]. Therefore, the mechanisms by which EPO decreases the incidence of CSA-AKI in patients without high risk for AKI could be explained by the following reasons. First, EPO was shown to protect against IR injury and have a reno-protective effect [19]. Both of them could contribute to the preventive effect of EPO on CSA-AKI. Secondly, EPO could inhibit cell apoptosis via down-regulating the ratio of Bax/Bcl-2 and decreasing superoxide production [30, 31]. Thirdly, EPO, as a cytokine for haematopoiesis, could raise the
4
Table 1: Baseline characteristics of the five eligible randomized controlled trials Study
Yoo et al. [21]
De Seigneux et al. [16]
Kim et al. [17]
Tasanarong et al. [20]
Surgery type
EPO
Administration Control time
Study design/Jadad score
CABG/CPB or OP
300 U/kg EPO, IV
NS, IV
Before surgery
R, DB, PC /5
NA/NA; preoperative anaemia At risk for AKIa; NA/NA/NA
Valvular heart surgery/CPB
500 IU/kg rHuEPO+100 ml NS NS, IV with 200 mg iron, IV
Before surgery
R, SB, PC/4
Cardiac surgery/CPB
α-Epoetin: 20 000 ui or α-Epoetin: 40 000 ui, IV
NS, IV
After surgery
R, DB, PC/5
At risk for AKIb; NA/NA
Valvular heart surgery/CPB
300 IU/kg rHuEPO-α, IV
NS, IV
Before surgery
R, DB, PC/5
NA/NA; NA/NA
CABG/CPB
200 U/kg rHuEPO 3 d before operation+100 U/kg rHuEPO at surgery, IV
NS, IV
Perioperative surgery
R, DB, PC/5
Sample; patients age (years); male (%)
BMI (Kg/m2); Hypertension (%);
eGFR(ml/min/1.73 m2); Hb(g/dl) ; SCr (mg/dl)
Risk for AKI; others
71 (36/35); 64.6 ± 10.7/ 68.9 ± 8.4; 83%/ 66% 74 (37/37); 56 ± 12/59 ± 12; 65%/62% 80 (40/20/20); 68.9 ± 12.0/ 66.5 ± 16.5/ 64.7 ± 14.7; 80%/65%/67.5% 98 (49/49); 63 ± 10/62 ± 10; 41%/49% 100 (50/50); 63 ± 16/60 ± 16; 62%/52%
24.7 ± 3.3/ 23.5 ± 2.3; 83%/ 64%
61.60 ± 25.20/59.30 ± 21.70; 13.10 ± 2.30/12.60 ± 1.60; 1.20 ± 0.38/1.08 ± 0.32
NA/NA; NA/NA
23 ± 3/23 ± 4; 24.3%/37.8%
NA/NA; 11.80 ± 0.80/11.60 ± 1.20; NA/NA;
NA/NA/NA; 85%/60%/65%
NA/ NA/NA; NA/NA/NA; 1.05 ± 0.29/0.98 ± 0.27/ 0.96 ± 0.30
24 ± 3/24 ± 4; 55%/49%
83 ± 28/77 ± 30; NA/NA; 0.92 ± 0.28/0.99 ± 0.27 64 ± 29/67 ± 33; 12.30 ± 1.7/12.20 ± 1.90; 1.05 ± 0.27/1.05 ± 0.45
NA/NA; 79%/63%
Intervention
AKI: acute kidney injury; CABG: coronary artery bypass grafting; CPB: cardiopulmonary bypass; OP: off-pump; IV: intravenous administration; NS: normal saline; eGFR: estimated glomerular filtration rate; Hb: haemoglobin; SCr: serum creatinine; R: randomized; DB: double-blinded; SB: single-blinded ; PC: placebo-controlled; Defined as sepsis, postoperative state, mechanical ventilation, haemodynamic impairment and previous chronic disease. b Defined as involving more than two of the following criteria: female, age >65 years, New York Heart Association (NYHA) functional class IV, left ventricular ejection fraction 1.2 mg/dl or peripheral vascular disease. a
H.-T. Tie et al. / European Journal of Cardio-Thoracic Surgery
Song et al. [19]
Population (EPO/Control)
SCr: serum creatinine; HLOS: hospital length of stay; gr1: group 1, patient receiving α-Epoetin: 20 000 ui; gr2: group 2, patient receiving α-Epoetin: 40 000 ui.
Tasanarong et al. [20]
De Seigneux et al. [16] Kim et al. [17]
Yoo et al. [21]
5
level of erythrocyte anti-oxidative enzymes by increasing the circulating young red blood cells [32]. However, whether the antiinflammation property of EPO plays an important role in this protection mechanism is still controversial [16, 17, 33]. In our subgroup analysis, the pooled results suggested that EPO administration could significantly reduce the incidence of CSA-AKI in patients without high risk factors for AKI [19–21], but not in patients with high risk factors for AKI [16, 17]. Thus, high risk factors seemed to cause the disagreement among current evidence. However, the definition of high-risk factors for AKI was not standardized in the included studies. The study conducted by De Seigneux et al. [16] regarded risk factors for AKI as sepsis, postoperative state, mechanical ventilation, haemodynamic impairment and previous chronic disease, while the other study by Kim et al. [17] defined it as involving more than two of the following criteria: female, age >65 years, New York Heart Association (NYHA) functional class IV, left ventricular ejection fraction < 35%, diabetes mellitus, chronic obstructive pulmonary disease (COPD), preoperative creatinine >1.2 mg/dl or peripheral vascular disease. According to previous published studies, there are several potential explanations for the ineffectiveness of EPO in patients with high risk factors for AKI. First, EPO is only beneficial to prevent AKI before ischaemic injury rather than after such injury [16, 19]. The kidney of patients having high risk factors for AKI might have been injured because renal dysfunction occurs several hours after the kidney injury and current diagnostic methods lack specificity [10]. Moreover, EPO could only protect against injury resulting from ischaemia but not inflammation [16, 17]. Consistently, patients with high risk factors for AKI usually present with some inflammation-associated diseases. Furthermore, many medicines, toxic to kidneys (e.g. beta-lactam, aminoglycoside and amphotericin), are usually required for in these patients [22], and drug-induced renal injury may not be mitigated by EPO. Our meta-analysis additionally revealed that EPO administration could significantly reduce the ICU length of stay and the HLOS in patients without high risk factors for AKI, which could consequently conserve healthcare resources. However, the benefits of EPO disappeared and HLOS even lengthened in patients with high risk factors for AKI. Since high risk factors for AKI (including NYHA IV, LVEF
