International Urology and Nephrology https://doi.org/10.1007/s11255-018-1814-0
NEPHROLOGY - REVIEW
Comparative efficacy of pharmacological interventions for contrast‑induced nephropathy prevention after coronary angiography: a network meta‑analysis from randomized trials Wen‑Qi Ma1 · Yu Zhao2 · Ying Wang1 · Xi‑Qiong Han1 · Yi Zhu1 · Nai‑Feng Liu1 Received: 10 November 2017 / Accepted: 29 January 2018 © Springer Science+Business Media B.V., part of Springer Nature 2018
Abstract Background Contrast-induced nephropathy (CIN) is the major complication related to contrast media administration in patients after coronary angiography (CAG). However, inconsistent results have been published in the literature regarding the effects of pharmacological drugs on CIN prevention. We conducted a network meta-analysis to evaluate the relative efficacy of pharmacological interventions for the prevention of CIN. Methods We searched MEDLINE, EMBASE, Cochrane Central Register of Controlled Trials, and ClinicalTrials.gov from inception to July 2017. We included any randomized controlled trials of eleven pharmacological interventions that reported the prevention of CIN. Results We identified 3850 records through database searches, of which 107 studies comprising 21,450 participants were finally identified. Compared with intravenous saline, intravenous saline plus pharmacological drugs including statin [relative risk (RR) 0.57; 95% credibility interval (CrI) 0.39 to 0.83], N-acetylcysteine (NAC) (RR 0.84; 95% CrI, 0.71 to 0.98), vitamin and its analogues (RR 0.66; 95% CrI 0.45 to 0.97), brain natriuretic peptide (BNP) and its analogues (RR 0.46; 95% CrI 0.30 to 0.70), prostaglandin analogues (RR 0.37; 95% CrI 0.18 to 0.76), NAC plus sodium bicarbonate (SB) (RR 0.60; 95% CrI 0.39 to 0.90), and statin plus NAC (RR 0.39; 95% CrI 0.21 to 0.70), have helped to reduce the incidence of CIN in patients after CAG. The top four ranked treatments were statin plus NAC, BNP and its analogues, statin, and vitamin and its analogues, respectively. NAC plus intravenous saline was associated with lower incidence of short-term all-cause mortality than intravenous saline alone (RR 0.62; 95% CI, 0.40 to 0.96; P = 0.03). However, no evidence indicated that any of the pharmacological drugs were associated with a reduced requirement for dialysis and major adverse cardiac and cerebrovascular events (MACCE). Conclusions Statin plus NAC plus intravenous saline seems to be the most effective treatment for the prevention of CIN in patients after CAG. NAC plus intravenous saline may have a protective role against short-term all-cause mortality. However, none of these drugs has effectively decreased the requirement for dialysis and MACCE. Keywords Contrast-induced nephropathy · Meta-analysis · Pharmacological intervention · Prevention
Introduction Electronic supplementary material The online version of this article (https://doi.org/10.1007/s11255-018-1814-0) contains supplementary material, which is available to authorized users. * Nai‑Feng Liu
[email protected] 1
Department of Cardiology, Zhongda Hospital, School of Medicine, Southeast University, 87 Dingjiaqiao, Nanjing 210009, People’s Republic of China
Department of Nephrology, Zhongda Hospital, School of Medicine, Southeast University, 87 Dingjiaqiao, Nanjing 210009, People’s Republic of China
2
Coronary angiography (CAG) and percutaneous coronary intervention (PCI) have been extensively performed in coronary artery disease (CAD) for coronary revascularization. Despite significantly improving cardiovascular outcomes and reducing cardiovascular mortality, one potential renal complication that associated with the administration of contrast media (CM) is contrast-induced nephropathy (CIN), which has been a vexing reality [1–3]. CIN, manifested as an abrupt deterioration in renal function after the administration of CM, is the third leading cause of
13
Vol.:(0123456789)
all hospital-acquired renal insufficiency and associated with prolonged hospitalization, increase in the morbidity of cardiovascular diseases, and other poor outcomes [2, 4–6]. The reported incidence of CIN varies widely, from 2 to 50%, with low incidence in the general population [3, 7]. However, the incidence rate is obviously increased in patients with preexisting renal impairment or other risk factors such as diabetes mellitus (DM), congestive heart failure, advanced age, and hypertension [8, 9]. Although the precise mechanisms of CIN are not fully understood, hypoxia-induced renal tubular injury, rheological alterations, activation of tubuloglomerular feedback, and cytotoxic effects of CM are frequently involved in the pathogenesis of CIN [10–12]. The potential deleterious effects of CM on renal function have been well recognized in the literature, and various preventative strategies including risk assessment before CM exposure, the withdrawal of the nephrotoxic drugs, volume expansion with sodium chloride or sodium bicarbonate (SB), adjuvant pharmacological therapy, hemofiltration or hemodialysis, and the optimal CM policy have been developed to CIN prevention [13–15]. However, the incidence of CIN remains unpredictable. Currently, adequate hydration prior to and following CM exposure is recommended by guidelines (e.g., RIFLE, AKIN, and KDIGO) as the principal prophylactic intervention [16–19]. Available data from randomized controlled trials (RCTs) have demonstrated the importance and benefits of pharmacological drugs in CIN prevention, such as statin, N-acetylcysteine (NAC), SB, and ascorbic acid [20–23]. However, no consensus exists regarding which is the optimal treatment that is preferred for the prevention of CIN; how best to treat the high-risk population; whether CIN is causally related to the in-hospital complications; and whether preventing CIN by pharmacological drugs can reduce the mortality. Previous meta-analyses are limited to comparing the efficacy of all available interventions and vary in their conclusions [24–27]. Furthermore, no meta-analyses have been systematically conducted to investigate all these interventions together in patients following CAG and/or PCI. To address these knowledge gaps, we performed a comprehensive Bayesian network meta-analysis (NMA) to assess the relative efficacy of pharmacological interventions for CIN prevention in above-mentioned population; to estimate the probability of each treatment being the most effective for reducing the CIN risk; and to further investigate the association between CIN and its clinical consequences.
Methods The protocol of this NMA has been registered on the International Prospective Register of Systematic Reviews (Registration number CRD42017057074; http://www.crd.york.
13
International Urology and Nephrology
ac.uk/PROSPERO). We reported this NMA based on the Preferred Reporting Items for Systematic Reviews and MetaAnalyses (PRISMA) statement for NMA [28].
Data sources and searches The literature search was conducted by two independently reviewers (WQM and XQH), and disagreements were resolved by consensus-based discussion. A comprehensive literature search was performed in major electric databases, including MEDLINE, EMBASE, Cochrane Central Register of Controlled Trials (CENTRAL), and ClinicalTrials.gov. The search was restricted to English language articles. The deadline for publication for inclusion in the meta-analysis was July 2017. Our search strategy was based on keywords in combination with both medical subject headings (MeSH) terms and text words. The following search terms and search strategy were used: (acute renal injury OR acute kidney injuries OR renal insufficiency OR acute kidney failure) AND (contrast-associated OR contrast OR contrast-induced OR radiocontrast) AND (controlled clinical trial OR randomized OR placebo OR randomized controlled trial). In addition, relevant meta-analyses of CIN prevention and reference lists of included studies were manually scanned for additional potential studies. Detailed information on the search strategy is provided in supplement material (Literature search strategy).
Study selection The following inclusion criteria were used for final study selection: (1) full text reports of RCTs; (2) evaluating the efficacy of pharmacological drugs for CIN prevention; (3) all the patients included in the eligible studies following CAG and/or PCI; (4) hydration is the co-intervention in the treatment and control groups; (5) comparing any of the following drugs was considered: statin, NAC, vitamin (e.g., ascorbic acid and vitamin E) and its analogues, SB, brain natriuretic peptide (BNP) and its analogues, prostaglandin analogues, theophylline, angiotensin-converting enzyme (ACE) inhibitor, NAC plus SB, and statin plus NAC; (6) If data were reported over 72 h, we used data obtained within the first 5 days; and (7) reported sufficient data and at least one of the following outcomes were mentioned: the incidence of CIN, serum creatinine (SCr) level after procedure, requirement for dialysis, major adverse cardiac and cerebrovascular events (MACCE), and short-term all-cause mortality. Studies were excluded if (1) animal studies, review articles, short communications, abstracts, letters to editor, or case reports; (2) insufficient data or duplicate publication; (3) without clear definitions of CIN or outcomes of interest; and (4) other angiography including computed tomography angiography, intravenous pyelography, peripheral vascular
International Urology and Nephrology
angiography, cerebral angiography, or renal angiography, were all excluded. The primary outcome measure was the incidence of CIN in patients after CAG and/or PCI, which is defined as an impairment in renal function with an increase in SCr ≥ 0.5 mg/dl (44.2 mmol/l) or a 25% elevation of SCr above baseline within 72 h after CM exposure. The second outcome measures were as follows: SCr level after procedure, the requirement for dialysis, MACCE, and short-term all-cause mortality. MACCE was defined as all-cause death, myocardial infarction (MI), unstable angina pectoralis, heart failure, stroke or cerebrovascular accident, repeat revascularization, and sustained or malignant arrhythmia. The short-term mortality was defined as death from any cause within 30 days or in-hospital mortality. If the SCr value was reported at multiple timepoints, the data were extracted at the following rank: 48 h, 72 h, 24 h, or other times.
Data Extraction and Quality Assessment Two reviewers (WQM and YW) extracted data independently on a purpose-built electronic form. Discrepancies were settled by discussion with a third reviewer (NFL). Extracted data included the characteristics of the studies (publication date, study design, country, sample size, CM type and volume, and inclusion criterion of renal function), baseline characteristics of the patients (age, sex, and underlying diseases), treatment characteristics, and clinical outcomes. The quality of each included study was evaluated by using the risk of bias assessment tool from the Cochrane Handbook for Randomized Controlled Trials [29]. The criteria used for quality assessment were random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and other potential bias. This assessment was made independently by two reviewers (WQM, YZ), and disagreements were resolved by a third reviewer (NFL) through consensus.
Data synthesis and analysis For both pairwise meta-analysis and NMA, the relevant treatment effects of pharmacological interventions were estimated using relative risk (RR) for the dichotomous outcomes (incidence of CIN, need for dialysis, MACCE, and short-term all-cause mortality) and weighted mean difference (WMD) for continuous outcomes (SCr level after procedure), respectively. Data were analyzed based on intention-to-treat analysis, when available. The multi-arm studies included in our meta-analysis were treated as multiple independent two-arm studies to perform a pooled analysis. We excluded trials with 0 or 100% events on all arms from the analysis. For dichotomous outcomes, we added 0.5 to each
intervention group when zero outcome events were reported in one intervention arm. Standard pairwise meta-analysis with a random-effects model was carried out to assess treatment effects through the DerSimonian–Laird method [30]. Statistical analyses were performed using STATA 12.0 (Stata Corporation, College Station, Texas, USA), and the results were reported as RR or WMD with 95% confidence intervals (CIs). For indirect and multiple treatment assessments, we performed a NMA to compare multiple prophylactic interventions in a Bayesian framework, using the R statistical software (version 3.1.1, R Foundation for Statistical Computing) for statistical analysis, and the pooled estimates were obtained using the Markov Chains Monte Carlo method [31]. We used 4 parallel chains and obtained 100,000 samples after a 50,000-sample burnin in each chain. Outcomes were reported as RR or WMD with 95% credible intervals (CrIs). In standard pairwise meta-analysis, heterogeneity of treatment effects observed in individual trials of the same comparison will be assessed using the Cochrane Q-test, and index (I2), with P > 0.10 in Q-test or I2 0.05 indicating no inconsistency [35]. Meta-regression was used to assess heterogeneity on the primary outcome based on baseline age, sex ratio, DM ratio, the dosage of CM, and region. Sensitivity analysis was performed on the primary outcome according to the following prespecified variables: publication year (excluding studies published before 2010), statistical models (using fixedeffects model), risks of bias (excluding studies with more than one item indicating a high risk of bias assessed by the Cochrane risk of bias tool), type of CM (excluding studies
13
with high-osmolar CM), sample size (excluding the studies with the sample size less than 100 patients), and the baseline renal function (including studies with renal dysfunction). Subgroup analysis was performed by several major covariates, including average age, CM type, the percent of diabetic patients, and study location. In addition, we draw a comparison-adjusted funnel plot to examine potential publication bias (presence of small-study effects) [33].
Results Study characteristics and quality assessment We identified 3850 records from the initial literature search, of which 1745 were removed because of duplicate studies. After reviewing titles and abstracts of 2105 records, 1998
International Urology and Nephrology
articles were excluded, which account for irrelevant content. Finally, 107 RCTs investigating eleven different interventional treatments were included. Figure 1 shows a flow diagram of the literature selection process. Eleven nodes were included in our NMA, and each of the nodes concerned different interventions (Fig. 2). Overall, NAC was the most “common comparators.” The most frequent comparison (n = 39 trials) is between NAC plus intravenous saline and intravenous saline (alone or with placebo). The characteristics of the studies included in the metaanalysis and the citation details are given in the Supplement Table 1 and 2. The publication year ranged from 1999 to 2016. In total, 21,450 participants were included and male participants accounted for 66.5% of the whole population. At baseline, the mean age of participants ranged from 54 to 76 years. The sample sizes of included studies ranged from 40 to 1000 subjects. The regions of the trials where
Fig. 1 Flow diagram of study selection. The flow diagram was depicted following the guideline of Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA)
13
International Urology and Nephrology
Fig. 2 Network of comparisons included in the analyses. Nodes represent the treatments being compared. The size of the node is proportional to the randomly assigned participants and indicates the sample size. Edges represent direct comparisons, and their width corresponds to the number of trials. NAC N-acetylcysteine, SB sodium bicarbonate, ACE angiotensin-converting enzyme, BNP brain natriuretic peptide
participants recruited were as follows: eighteen studies recruited participants from America, thirty-seven studies recruited participants from Europe, fifty studies recruited participants from Asia, and the remaining two studies recruited participants from Africa. Most included studies reported using low-osmolar or iso-osmolar CM, whereas only a minority of studies with high-osmolar CM. Thirtyone studies included the patients with baseline renal insufficiency (defined as SCr levels > 1.5 mg/dl), and eight studies only recruited DM patients. The results of the quality assessment of individual studies are presented in Supplement Fig. 1 and 2. Overall, trials were of generally good quality with low risk of bias. None of the trials was considered to have a high risk of bias for any of the methodological quality items assessed. The risk of bias was low for random sequence generation in 55 trials (50.9%), concealment of treatment allocation in 42 trials (38.9%), masking of participants and personnel in 52 trials (48.1%), blinding of outcome assessment in 60 trials (55.6%), incompleteness of outcome data in 68 trials (63.0%), selective reporting of outcomes in 68 trials (63.0%), and other bias in 19 trials (17.8%).
Treatment outcomes Primary outcome: the incidence of CIN In total, 103 trials involving 21,113 participants evaluated the association of pharmacological interventions with the incidence of CIN. The pooled estimates of the drug treatment estimated with the direct pairwise random-effects meta-analyses are presented in Supplement Table 3.
Compared with intravenous saline (alone or with placebo), intravenous saline plus pharmacological drugs, including statin, NAC, vitamin and its analogues, BNP and its analogues, and prostaglandin analogues, significantly reduced the incidence of CN. The results of NMA that combines direct and indirect evidence are presented in Table 1. The overall results were partly similar to that observed based on direct comparisons, but with greater precision. Statin, NAC, vitamin and its analogues, BNP and its analogues, prostaglandin analogues, NAC plus SB, statin plus NAC, but not SB, theophylline, and ACE inhibitor, are shown to be more effective than intravenous saline in reducing the incidence of CIN [RRs ranging from 0.37 (95% CrI 0.18 to 0.76) for prostaglandin analogues to 0.84 (95% CrI 0. 17 to 0.98) for NAC]. On comparative-effectiveness of various pharmacological drugs, prostaglandin analogues were superior to NAC (RR 0.45; 95% CrI 0.21 to 0.93) and SB (RR 0.44; 95% CrI 0.20 to 0.45), respectively. Statin, vitamin and its analogues, BNP and its analogues, prostaglandin analogues, NAC plus SB, and statin plus NAC were associated with lower CIN risk when compared with ACE inhibitor (0.41 [0.21, 0.81]; 0.47 [0.24, 0.93]; 0.33 [0.17, 0.64]; 0.27 [0.11, 0.66]; 0.43 [0.22, 0.86]; and 0.28 [0.12, 0.63], respectively). Furthermore, we observed that NAC plus statin was significantly more effective than NAC monotherapy (RR 0.46; 95% CrI 0.25 to 0.83) and SB (RR 0.46; 95% CrI 0.24 to 0.88), but not to statin monotherapy (RR 0.68; 95% CrI 0.35 to 1.30). We also found that the combination of NAC and SB was not significantly different from NAC monotherapy (RR 0.72; 95% CrI 0.46 to 1.10) and SB monotherapy (RR 0.71; 95% CrI 0.45 to 1.10). Ranking probabilities for all interventions, which based on surface under the cumulative ranking (SUCRA) curves, are presented in Fig. 3. The SUCRA values provide the hierarchy for eleven interventions that are 11.6, 78.0, 42.2, 62.3, 30.3, 80.5, 54.6, 40.5, 4.0, 53.6, and 92.4% for intravenous saline, statin, NAC, vitamin and its analogues, SB, BNP and its analogues, prostaglandin analogues, theophylline, ACE inhibitor, NAC plus SB, and statin plus NAC, respectively (Supplement Table 4). According to the results of SUCRA, statin plus NAC had the most chance to have a best impact on the incidence of CIN, followed by BNP and its analogues, statin, and vitamin and its analogues, whereas ACE inhibitor ranked worst. We also display the SUCRA values from a network meta-regression in order to avoid small-study effects, and the results indicated the reduction of the relative effectiveness of NAC plus SB in a model adjusted for small-study effects, when compared with the model without adjustment. However, the total ranking of treatments did not alter. In addition, according to multidimensional scaling (MDS), the top four ranked treatments were statin plus NAC, BNP and its analogues, statin, and
13
13
0.68(0.45,1.0) NAC 0.79(0.52,1.2)
1.2(0.69,2.2) 1.8(1.1,2.8) 1.4(0.80,2.5)
1.5(0.68,3.5) 2.2(1.1,4.8) 1.8(0.79,4.0)
2.1(0.96,4.7) 3.6(1.6, 8.2) 1.6(0.75,3.2) Statin + NAC 2.6(1.4,4.8)
0.81(0.48,1.4) 1.40(0.80,2.4) 0.60(0.39,0.90) 0.39(0.21,0.70) IV saline
1.3(0.70,2.6) 2.3(1.2,4.6) NAC + SB 0.64(0.31,1.3) 1.7(1.1,2.6)
Theophylline 1.7(0.81,3.6) 0.74(0.38,1.4) 0.48(0.21,1.0) 1.2(0.74,2.1)
0.58(0.27,1.2) ACE inhibitor 0.43(0.22,0.86) 0.28(0.12,0.63) 0.72(0.41,1.3)
0.84(0.64,1.1) 0.46(0.30,0.70) 0.37(0.18,0.76)
0.57(0.39,0.83) 0.84(0.71,0.98) 0.66(0.45,0.97)
1.0(0.58,1.9) 0.61(0.33,1.1) 1.4(0.89,2.2) 2.2(1.1,4.2) 0.57(0.30,1.1) 0.33(0.17,0.64) 0.77(0.43,1.4) 1.2(0.58,2.5) 0.46(0.19,1.1) 0.27(0.11,0.66) 0.62(0.27,1.4) 0.97(0.38,2.5)
0.70(0.37,1.3) 0.41(0.21,0.81) 0.95(0.54,1.7) 1.5(0.79,2.8) 1.0(0.61,1.8) 0.60(0.34,1.1) 1.4(0.92,2.2) 2.2(1.2,4.0) 0.81(0.43,1.6) 0.47(0.24,0.93) 1.1(0.62,2.0) 1.7(0.84,3.5)
SB sodium bicarbonate, ACE angiotensin-converting enzyme, NAC N-acetylcysteine, BNP brain natriuretic peptide, IV intravenous
Comparisons between treatments should be read from left to right and the estimate is in the cell in common between the column-defining treatment and the row-defining treatment. Values are presented as relative risk (RR) with 95% credibility interval (CrI). Bold values indicate comparisons that are statistically significant. RR 1 favors the treatment in the row
SB 0.55(0.34,0.91) 0.44(0.20,0.95)
0.67(0.43,1.1) 0.99(0.74,1.3) 0.78(0.49,1.3)
1.8(1.1,3.0) 2.2(1.1,4.9) BNP and its analogues 1.2(0.55,2.9) 0.81(0.35,1.8) Prostaglandin analogues 1.4(0.75,2.7) 0.97(0.57,1.6) 1.2(0.64,2.3) 0.96(0.54,1.7) 1.7(0.90,3.4) 2.2(0. 90,5.3) 2.4(1.2,4.8) 1.7(0.93,3.0) 2.1(1.1,4.1) 1.6(0.90,3.1) 3.0(1.5,6.0) 3.7(1.5,9.2) 1.1(0.60,1.8) 0.72(0.46,1.1) 0.91(0.51,1.6) 0.71(0.45,1.1) 1.3(0.71,2.3) 1.6(0.70,3.7) 0.68(0.35,1.3) 0.46(0.25,0.83) 0.59(0.281.2) 0.46 (0.24,0.88) 0.83(0.39,1.7) 1.0(0.40,2.6) 1.8(1.2,2.6) 1.2(1.02,1.4) 1.5(1.1,2.2) 1.2(0.91,1.6) 2.3(1.4,3.3) 2.7(1.3,5.6)
0.87(0.5,1.5) 1.3(0.84,1.9) Vitamin and its analogues 1.5(0.93,2.3) 1.0(0.74.1.4) 1.3(0.80,2.0) 0.82(0.46,1.4) 0.55(0.35,0.87) 0.71(0.40,1.2) 0.66(0.29,1.5) 0.45(0.21,0.93) 0.57(0.25,1.3)
Statin 1.5(0.98,2.2) 1.2(0.67,2.0)
Table 1 Network-estimated relative risk (95% confidence intervals) for the incidence of contrast-induced nephropathy derived from network meta-analysis based on direct and indirect comparisons
International Urology and Nephrology
International Urology and Nephrology Fig. 3 Cumulative ranking probabilities (SUCRA values) for all interventions in the incidence of CIN. Intervention is ranked based on SUCRA. The larger the surface area under the curve and the faster the curve rises, the greater the possibility of being the most efficacious treatment. Blue solid lines represent the unadjusted model, whereas red dashed lines the adjusted for small-study effects model. SUCRAsurface under cumulative ranking curves; NAC N-acetylcysteine, SB sodium bicarbonate, ACE angiotensinconverting enzyme, BNP brain natriuretic peptide
vitamin and its analogues, respectively, whereas intravenous saline ranked worst (Supplement Fig. 3). There are partly differences between SUCRA and MDS rankings accounting for the treatment effects of several drugs are similar. In addition, subgroup analysis was performed to validate the results from overall analysis. When stratifying by average age, CM type, the percent of diabetic patients, and study location, the significant relations of pharmacological drugs on the prevention of CIN were observed in several subgroups (Supplemental Table 5). Secondary outcomes:SCr level after the procedure, dialysis requirement, MACCE, and short‑term all‑cause mortality Data on SCr level after the procedure were available from 77 trials, including 13,737 patients. The results of direct treatment comparisons indicated that NAC (WMD − 0.06; 95% CI − 0.09 to − 0.04), BNP and its analogues (WMD − 0.07; 95% CI − 0.12 to − 0.01), and theophylline (WMD − 0.20; 95% CI − 0.36 to − 0.05) resulted in less increase in SCr level after procedure, when compared with intravenous saline. Compared with vitamin and its analogues, those who received NAC experienced more increase in SCr level after procedure (WMD 0.10; 95% CI 0.001 to 0.20) (Supplement Table 6). The NMA suggested only one comparison reached statistical significance, with statin plus NAC showing less increase in SCr level, when compared with intravenous saline (WMD − 0.24; 95% CrI − 0.38 to − 0.098). We also found that statin plus
NAC was superior to NAC (− 0.19; 95% CrI − 0.33 to − 0.050), vitamin and its analogues (WMD − 0.19; 95% CrI − 0.36 to − 0.025), SB (WMD − 0.22; 95% CrI − 0.38 to − 0.063), ACE inhibitor (WMD − 0.25; 95% CrI − 0.43 to − 0.066), and NAC plus SB (WMD − 0.21; 95% CrI − 0.40 to − 0.018) in this regard, respectively (Supplement Table 7). After exclusion of trials reporting zero events and not reporting the data of dialysis requirement, a total of 22 RCTs were finally in analysis, with 5288 patients and only 50 of which were reported on the requirement for dialysis, indicating low requirement for dialysis rate. Results of direct treatment comparisons indicated that there was no significant difference in dialysis requirement between any of the treatments (Supplement Table 8). Insufficient observations were available to generate evidence networks for dialysis requirement. Data on MACCE and short-term all-cause mortality were available form 33 and 23 RCTs, including 7578 and 5819 patients, respectively. Direct comparison revealed NAC in combination with intravenous saline was more efficient than intravenous saline alone in reducing shortterm all-cause mortality (RR 0.62; 95% CI 0.40 to 0.96; P = 0.03) (Supplement Table 9). However, no significant differences between any of treatments were observed for RRs of MACCE (Supplement Table 10). In NMA, there was no evidence of significantly different associations with MACCE or all-cause mortality between any of the treatments (Supplement Table 11 and 12).
13
International Urology and Nephrology
Network inconsistency Two approaches were developed to deal with inconsistency in the network of interventions. By using loop-specific approach, each ROR with a 95% CI including 1 indicates little evidence of loop inconsistency between direct and indirect evidence. By using node-splitting approach, no significant inconsistency was observed within the networks except 1 inconsistency node (NAC plus SB vs. intravenous saline, P = 0.02715) (Supplement Fig. 4 and 5).
Sensitivity analysis, meta‑regression, and publication bias In sensitivity analysis, treatment estimates in the primary outcome were similar when restricted to studies publishing after 2010, using fixed-effects model, with no more than one high risk of bias assessed by the Cochrane risk of bias tool, not using high-osmolar CM, and with the sample size more than 100 patients (Supplement Table 13). However, if we included studies involving patients with renal dysfunction, most results were inconsistent with the standard analysis. The meta-regression with the incidence of CIN indicated that baseline age, sex ratio, DM ratio, and region were not the source of heterogeneity in analysis, whereas the dosage of CM was proved to be a potential source of heterogeneity (Supplement Table 14). In addition, the comparison-adjusted funnel plot for the primary outcome is symmetric around the zero line, which indicates the absence of small-study effects (Fig. 4).
Discussion CAG and PCI provide a high-risk condition for the development of CIN, especially in patients with preexisting renal dysfunction. Most patients experienced low adverse effects with the administration of CM, whereas those who develop CIN may have markedly worsened prognoses. In-hospital complications, such as cerebrovascular accident, acute pulmonary edema, sepsis, or major adverse coronary events, are frequently reported [2]. Currently, various pharmacological agents in combination with intravenous saline are routinely administered as potential CIN prophylaxis. However, the incidence of CIN is still high, and nearly none of these drugs appears to have a sound evidence base. Combining evidence from 107 RCTs with 21,450 patients, we made several important observations. First, most of pharmacological drugs in our study, such as statin, NAC, vitamin and its analogues, BNP and its analogues, prostaglandin analogues, NAC plus SB, and statin plus NAC, in combination with intravenous saline, have additional benefits to reduce the incidence of CIN than intravenous saline
13
Fig. 4 Comparison-adjusted funnel plot for the assessment of the incidence of contrast-induced nephropathy. The red line represents the null hypothesis that the study-specific effect sizes do not differ from the respective comparison-specific pooled effect estimates. The blue line is the regression line. Different colors correspond to different comparisons. A = statin; B = N-acetylcysteine (NAC); C = vitamin and its analogues; D = sodium bicarbonate (SB); E = brain natriuretic peptide and its analogues; F = prostaglandin analogues; G = theophylline; H = angiotensin-converting enzyme inhibitor; I = NAC plus SB; J = statin plus NAC; K = intravenous saline (alone or with placebo)
alone. Statin plus NAC had the highest probability to be the best treatment in terms of CIN prevention, followed by BNP and its analogues, statin, and vitamin and its analogues. Second, NAC in combination with intravenous saline may play a protective role against short-term all-cause mortality. However, no clear evidence from available RCTs suggests that any of the pharmacological drugs was associated with reduced the requirement of dialysis and MACCE. Preexistent renal impairment is probably the most important pre-procedural risk factor for CIN [3]. In addition, other risk factors including DM, heart failure, advanced age, and hypertension are associated with additive causative effects on the incidence of CIN. In our study, subgroup analysis for studies that adjusted for main potential confounding factors was performed to assess the robustness of the results. We observed that the effects of several drugs were no longer statistically compared to intravenous saline in some subgroup analyses. The difference of the subgroup analysis from the overall results might be due to the smaller number of included studies and the homogeneity among the studies in subgroup analysis. However, compared to intravenous saline alone, statin plus NAC still significantly reduced the risk for CIN in most subgroup analysis. Our results are consistent with findings from several large-scale RCTs, which reported statistically significant reduction of the incidence of CIN with pharmacological treatment [22, 36, 37]. Recently, a comprehensive meta-analysis conducted by Subramaniam and colleagues
International Urology and Nephrology
comparing multiple prophylactic interventions summarized results from 86 RCTs and observed the reduction in CIN risk with NAC plus intravenous saline and with statins plus NAC plus intravenous saline, but only limited to direct comparisons [25]. Compared with these prior reports, there are several strengths to consider in our analysis. First, to our knowledge, our study is the most recent and comprehensive NMA evaluating the relative efficacy of eleven pharmacological interventions for CIN prevention. Second, we further investigated the association between CIN and clinical consequences, which may provide clinicians with a better understanding of CIN prevention. Third, model fit was good; between-study heterogeneity was low in most analyses; and sensitivity analyses were mostly consistent with the primary outcome; all of which increase robustness of our results. Currently, the definition and staging of CIN are mainly based on the clinical recommendations of RIFLE, AKIN, and KDIGO. However, the clinical utility of these criteria is still uncertain. In our study, we observed that the SCr value was reported at different timepoints among eligible studies, which is partly due to the lack of unified diagnostic criteria of CIN. According to the guidelines of CIN, intravenous volume expansion with either saline or SB solutions is recommended in multiple published guidelines on the prevention of CIN [17]. Antioxidant supplementations such as NAC, vitamin C, vitamin E, and alpha-lipoic acid are widely administered for primary prevention of CIN, which are partly based on the assumption that CIN is caused by reactive oxygen species. Our results are consistent with these recommendations. In addition, some of more recent interventions (e.g., prostaglandin analogues and BNP and its analogues) and the treatment of multi-drug combination (e.g., NAC plus SB and statin plus NAC) confirmed existing beliefs in the prevention of CIN, although we are not aware of any guidelines that recommend them; one possible explanation for this is the lack of large-scale RCTs that are adequately powered to determine the effectiveness of these drugs for the prevention of CIN. In our study, statin plus NAC is identified to be the most efficacious drug than other drugs in reducing the incidence of CIN, although the withinclass differences were small and may not be clinically meaningful. Compared with monotherapy, whether combined therapy with stain and NAC has an additive effect on CIN prevention is warranted to be identified in future research. Unexpected result was that ACE inhibitor plus intravenous saline is not superior to intravenous saline alone on the CIN incidence, although it was reported to be helpful in renal function and beneficial in MI [38]. Multiple lines of evidence have consistently suggested that patients experiencing CIN are associated with higher risk of in-hospital complications, which indicated a temporal relationship between CIN and in-hospital complications [8, 39–41]. However, these data could not definitively prove a
causal relationship. A critical question is whether reducing the risk of CIN will lower the in-hospital complications or the mortality rate. In our study, we confirmed that NAC plus intravenous saline has positive effects on reducing shortterm all-cause mortality in direct treatment comparisons, but failed to demonstrate significant differences in NMA. This difference may be related to Insufficient data. Furthermore, none of the drugs had a significant effect on reducing the requirement for dialysis and MACCE. On the other hand, effective prevention of CIN may not equal to alleviate the clinical consequences. Of note, this hypothesis is consistent with the results of Xin et al. [42] and Rezaei et al. [22], both showed the beneficial role of pharmacological drugs in lower the incidence of CIN, but not statistically powered to assess differences in in-hospital complications. Possible explanations for this question were that most patients have underlying risk factors (e.g., hypertension, heart failure, and DM) that can directly increase other in-hospital complications besides CIN. Further studies should be undertaken to explore how these risk factors could contribute to inhospital complications after CM administration, and how much mortality could be avoided if we could reduce the incidence of CIN. It is also worthwhile noting that most of the trials included had a short-duration of follow-up, and the longer-term effectiveness or safety of the available drugs was incompletely understood. Based on the above analysis, at least, pharmacological drugs are essential for the highrisk population to reduce the incidence of CIN, and efforts devoted to reducing the incidence of CIN remain a valuable clinical and a research goal. Several limitations should be addressed. First, detailed analyses were restricted by the availability of data in the included trials, particularly on dialysis requirement, MACCE and all-cause mortality, despite their clinical importance. Only a minority of trials reported the data and most had few or zero events; thus, the CrIs are wide and overlap for many treatments, and it is hard to draw a clear conclusion about the superiority of one intervention over another. Second, many trials suffered from insufficient participant numbers which may overestimate treatment effects. Third, diagnostic criteria of CIN varied across studies, which in turn allow to capture a higher number of events and therefore to better fit statistical models than when the CIN-narrow definition is performed. Fourth, our study enrolled patients with CAG and/or PCI only. These patients, who have an increased burden of cardiovascular risk factors before CM exposure, are higher in the incidence of CIN and in-hospital complications. Thus, it could limit the generalizability of our findings. Fifth, different doses of the same drug were ignored in our analysis, and it is possible that the relative drug efficacies may be dose-dependent and higher doses are probably more effective. However, the number of individual trials assessing individual doses was low in several comparisons. Thus, it
13
may not truly reflect the dose-dependent effects. Finally, the validity of a NMA is threatened by inconsistency in evidence within the network, and heterogeneity is a potential problem that may affect the robustness and accuracy of the results. It is very difficult to make all included studies sufficiently homogeneous in clinical and methodological characteristics, such as participants’ conditions, the duration of treatment, definitions and measurements of outcome, co-interventions, and follow-up. Although we did not find evidence of inconsistency in the current analysis, we further performed sensitivity analysis and meta-regression to investigate potential sources of heterogeneity. However, the results of this NMA should be interpreted with caution. In conclusion, the results of our NMA suggest that intravenous saline in combination with pharmacological drugs, including statin, NAC, vitamin and its analogues, BNP and its analogues, prostaglandin analogues, NAC plus SB, and statin plus NAC have benefits for the prevention of CIN in patients following CAG and/or PCI, when compared with intravenous saline alone. Furthermore, combined treatment (e.g., NAC plus SB or statin plus NAC) is superior to monotherapy for the CIN prevention. Clinical benefits of some more recent interventions (e.g., BNP and its analogues and prostaglandin analogues) are identified in our analysis for CIN prevention. However, well-designed large-scale RCTs evaluating clinical outcomes are urgently needed before implementing these drugs in guidelines. Statin plus NAC plus intravenous may play a positive role in reducing the incidence of short-term all-cause mortality. However, no evidence is available that any of the pharmacological drugs in combination with intravenous saline, has significantly lower requirement for dialysis and MACCE. In future, prospective epidemiologic studies conducted between CIN and longterm clinical outcomes, as well as large-scale RCTs with newly interventions are warranted to elucidate these issues. Funding This work was supported by National Nature Science Foundation of China (No. 81770451).
Compliance with ethical standards Conflict of interest None.
References 1. Mehran R (2007) Contrast-induced nephropathy remains a serious complication of PCI. J Interv Cardiol 20(3):236–240 2. McCullough P (2006) Outcomes of contrast-induced nephropathy: experience in patients undergoing cardiovascular intervention. Catheter Cardiovasc Interv 67(3):335–343 3. Rear R, Bell RM, Hausenloy DJ (2016) Contrast-induced nephropathy following angiography and cardiac interventions. Heart 102(8):638–648
13
International Urology and Nephrology 4. Laville M, Juillard L (2010) Contrast-induced acute kidney injury: how should at-risk patients be identified and managed? J Nephrol 23(4):387–398 5. McCullough PA (2008) Contrast-induced acute kidney injury. J Am Coll Cardiol 51(15):1419–1428 6. Weisbord SD, Gallagher M, Kaufman J, Cass A, Parikh CR, Chertow GM, Shunk KA, McCullough PA, Fine MJ, Mor MK, Lew RA, Huang GD, Conner TA, Brophy MT, Lee J, Soliva S, Palevsky PM (2013) Prevention of contrast-induced AKI: a review of published trials and the design of the prevention of serious adverse events following angiography (PRESERVE) trial. Clin J Am Soc Nephrol 8(9):1618–1631 7. Mehran R, Nikolsky E (2006) Contrast-induced nephropathy: definition, epidemiology, and patients at risk. Kidney Int Suppl 100:S11–S15 8. Narula A, Mehran R, Weisz G, Dangas GD, Yu J, Genereux P, Nikolsky E, Brener SJ, Witzenbichler B, Guagliumi G, Clark AE, Fahy M, Xu K, Brodie BR, Stone GW (2014) Contrastinduced acute kidney injury after primary percutaneous coronary intervention: results from the HORIZONS-AMI substudy. Eur Heart J 35(23):1533–1540 9. Ozcan OU, Er HA, Gulec S, Ustun EE, Gerede DM, Goksuluk H, KayaCetinErol CT (2015) Impact of metabolic syndrome on development of contrast-induced nephropathy after elective percutaneous coronary intervention among nondiabetic patients. Clin Cardiol 38(3):150–156 10. Heyman SN, Rosen S, Rosenberger C (2008) Renal parenchymal hypoxia, hypoxia adaptation, and the pathogenesis of radiocontrast nephropathy. Clin J Am Soc Nephrol 3(1):288–296 11. Ortega LM, Harmouch I, Nayer A (2015) Contrast-induced nephropathy: pathogenesis and new therapeutic options for prevention. Am J Ther 22(6):469–476 12. Wong PC, Li Z, Guo J, Zhang A (2012) Pathophysiology of contrast-induced nephropathy. Int J Cardiol 158(2):186–192 13. Mautone A, Brown JR (2010) Contrast-induced nephropathy in patients undergoing elective and urgent procedures. J Interv Cardiol 23(1):78–85 14. Ludwig U, Keller F (2014) Prophylaxis of contrast-induced nephrotoxicity. Biomed Res Int 2014:308316 15. Solomon R, Deray G, Consensus Panel for CIN (2006) How to prevent contrast-induced nephropathy and manage risk patients: practical recommendations. Kidney Int Suppl 100:S51–S53 16. Stacul F, van der Molen AJ, Reimer P, Webb JAW, Thomsen HS, Morcos SK, Almén T, Aspelin P, Bellin M-F, Clement O (2011) Contrast induced nephropathy: updated ESUR contrast media safety committee guidelines. Eur Radiol 21:2527–2541 17. Palevsky PM, Liu KD, Brophy PD, Chawla LS, Parikh CR, Thakar CV, Tolwani AJ, Waikar SS, Weisbord SD (2013) KDOQI US commentary on the 2012 KDIGO clinical practice guideline for acute kidney injury. Am J Kidney Dis 61(5):649–672 18. Bellomo R, Ronco C, Kellum JA, Mehta RL, Palevsky P, Acute Dialysis Quality Initiative Workgroup (2004) Acute renal failure - definition, outcome measures, animal models, fluid therapy and information technology needs: the Second International Consensus Conference of the Acute Dialysis Quality Initiative (ADQI) Group. Crit Care 8(4):R204–R212 19. Mehta RL, Kellum JA, Shah SV, Molitoris BA, Ronco C, Warnock DG, Levin A, Acute Kidney Injury Network (2007) Acute Kidney Injury Network: report of an initiative to improve outcomes in acute kidney injury. Crit Care 11(2):R31 20. Motohiro M, Kamihata H, Tsujimoto S, Seno T, Manabe K, Isono T, Sutani Y, Yuasa F, Iwasaka T (2011) A new protocol using sodium bicarbonate for the prevention of contrast-induced nephropathy in patients undergoing coronary angiography. Am J Cardiol 107(11):1604–1608
International Urology and Nephrology 21. Patti G, Ricottini E, Nusca A, Colonna G, Pasceri V, D’Ambrosio A, Montinaro A, Di Sciascio G (2011) Short-term, high-dose Atorvastatin pretreatment to prevent contrast-induced nephropathy in patients with acute coronary syndromes undergoing percutaneous coronary intervention (from the ARMYDA-CIN [atorvastatin for reduction of myocardial damage during angioplasty—contrastinduced nephropathy] trial. Am J Cardiol 108(1):1–7 22. Rezaei Y, Khademvatani K, Rahimi B, Khoshfetrat M, Arjmand N, Seyyed-Mohammadzad MH (2016) Short-term high-dose vitamin E to prevent contrast medium-induced acute kidney injury in patients with chronic kidney disease undergoing elective coronary angiography: a randomized placebo-controlled trial. J Am Heart Assoc 5(3):e002919 23. Jo SH, Koo BK, Park JS, Kang HJ, Kim YJ, Kim HL, Chae IH, Choi DJ, Sohn DW, Oh BH, Park YB, Choi YS, Kim HS (2009) N-acetylcysteine versus AScorbic acid for preventing contrastInduced nephropathy in patients with renal insufficiency undergoing coronary angiography NASPI study-a prospective randomized controlled trial. Am Heart J 157(3):576–583 24. Xu R, Tao A, Bai Y, Deng Y, Chen G (2016) Effectiveness of N-acetylcysteine for the prevention of contrast-induced nephropathy: a systematic review and meta-analysis of randomized controlled trials. J Am Heart Assoc 5(9):e003968 25. Subramaniam RM, Suarez-Cuervo C, Wilson RF, Turban S, Zhang A, Sherrod C, Aboagye J, Eng J, Choi MJ, Hutfless S, Bass EB (2016) Effectiveness of prevention strategies for contrastinduced nephropathy: a systematic review and meta-analysis. Ann Intern Med 164(6):406–416 26. Loomba RSSP, Aggarwal S, Arora RR (2016) Role of N-acetylcysteine to prevent contrast-Induced nephropathy: a meta-analysis. Am J Ther 23(1):e172–e183 27. Brown JR, Block CA, Malenka DJ, O’Connor GT, Schoolwerth AC, Thompson CA (2009) Sodium bicarbonate plus N-acetylcysteine prophylaxis: a meta-analysis. JACC Cardiovasc Interv 2(11):1116–1124 28. Hutton B, Salanti G, Caldwell DM, Chaimani A, Schmid CH, Cameron C, Ioannidis JP, Straus S, Thorlund K, Jansen JP, Mulrow C, Catala-Lopez F, Gotzsche PC, Dickersin K, Boutron I, Altman DG, Moher D (2015) The PRISMA extension statement for reporting of systematic reviews incorporating network metaanalyses of health care interventions: checklist and explanations. Ann Intern Med 162(11):777–784 29. Corbett MS, Higgins JP, Woolacott NF (2014) Assessing baseline imbalance in randomised trials: implications for the Cochrane risk of bias tool. Res Synth Methods 5(1):79–85 30. Salanti G (2012) Indirect and mixed-treatment comparison, network, or multiple-treatments meta-analysis: many names, many benefits, many concerns for the next generation evidence synthesis tool. Res Synth Methods 3(2):80–97
31. Smith TC, Spiegelhalter DJ, Thomas A (1995) Bayesian approaches to random-effects meta-analysis: a comparative study. Stat Med 14(24):2685–2699 32. Salanti G, Ades AE, Ioannidis JP (2011) Graphical methods and numerical summaries for presenting results from multiple-treatment meta-analysis: an overview and tutorial. J Clin Epidemiol 64(2):163–171 33. Chaimani A, Higgins JP, Mavridis D, Spyridonos P, Salanti G (2013) Graphical tools for network meta-analysis in STATA. PLoS ONE 8(10):e76654 34. Song F, Clark A, Bachmann MO, Maas J (2012) Simulation evaluation of statistical properties of methods for indirect and mixed treatment comparisons. BMC Med Res Methodol 12:138 35. Dias S, Welton NJ, Caldwellb DM, Ades AE (2010) Checking consistency in mixed treatment comparison meta-analysis. Stat Med 29(7-8):932–944 36. Morikawa S, Sone T, Tsuboi H, Mukawa H, Morishima I, Uesugi M, Morita Y, Numaguchi Y, Okumura K, Murohara T (2009) Renal protective effects and the prevention of contrast-induced nephropathy by atrial natriuretic peptide. J Am Coll Cardiol 53(12):1040–1046 37. Spargias K, Adreanides E, Demerouti E, Gkouziouta A, Manginas A, Pavlides G, Voudris V, Cokkinos DV (2009) Iloprost prevents contrast-induced nephropathy in patients with renal dysfunction undergoing coronary angiography or intervention. Circulation 120(18):1793–1799 38. Amann U, Kirchberger I, Heier M, Zirngibl A, von Scheidt W, Kuch B, Peters A, Meisinger C (2014) Effect of renin-angiotensin system inhibitors on long-term survival in patients treated with beta blockers and antiplatelet agents after acute myocardial infarction (from the MONICA/KORA Myocardial Infarction Registry). Am J Cardiol 114(3):329–335 39. Solomon RJ, Mehran R, Natarajan MK, Doucet S, Katholi RE, Staniloae CS, Sharma SK, Labinaz M, Gelormini JL, Barrett BJ (2009) Contrast-induced nephropathy and long-term adverse events: cause and effect? Clin J Am Soc Nephrol 4(7):1162–1169 40. Solomon R, Gordon P, Manoukian SV, Abbott JD, Kereiakes DJ, Jeremias A, Kim M, Dauerman HL, Investigators BT (2015) Randomized trial of bicarbonate or saline study for the prevention of contrast-induced nephropathy in patients with CKD. Clin J Am Soc Nephrol 10(9):1519–1524 41. Maioli M, Toso A, Leoncini M, Micheletti C, Bellandi F (2011) Effects of hydration in contrast-induced acute kidney injury after primary angioplasty: a randomized, controlled trial. Circ Cardiovasc Interv 4(5):456–462 42. Xing K, Fu X, Wang Y, Li W, Gu X, Hao G, Miao Q, Li S, Jiang Y, Fan W, Geng W (2016) Effect of rhBNP on renal function in STEMI-HF patients with mild renal insufficiency undergoing primary PCI. Heart Vessels 31(4):490–498
13
