AKI indicates acute kidney injury; ICU, intensive care unit; RRT, renal replacement therapy (dialysis).
Scores are reported as mean (SD). Serum creatinine values are similar between treatment groups at each time point, and they significantly differ from baseline with an analysis-of-variance test. To convert creatinine to μmol/L, multiply by 88.4. Shaded area indicates normal range of serum creatinine values, 0.5-1.25 mg/dL.
eFigure 1. Planned and post-hoc subgroup analyses for the primary end point: need for Renal Replacement Therapy
eFigure 2. Need for Renal Replacement Therapy in the 13 centers which randomized at least 10 patients
eFigure 3. Need for Renal Replacement Therapy in centers administering >0.1 µg/kg/min for =24 hours
eTable 1. Standardized criteria for the need for Renal Replacement Therapy
eTable 2. Adjustments for multiple ad interim analyses
eTable 3. Patient characteristics according to randomized group assignment, fenoldopam or placebo. Per Protocol analysis
eTable 4. Outcomes, end-points, and adverse events, according to randomized treatment assignment, fenoldopam or placebo. Per Protocol analysis
eTable 5. Indications and treatment variables for renal replacement therapy, according to randomized treatment assignment, fenoldopam versus placebo. Per Protocol analysis
eTable 6. Predictors of Renal Replacement Therapy - Multivariate analyses
eTable 7. Predictors of the composite endpoint Renal Replacement Therapy and/or Mortality - Multivariate analyses
eTable 8. Estimation of the Number Needed to Harm (NNH)
Tiziana Bove, Alberto Zangrillo, Fabio Guarracino, Gabriele Alvaro, Bruno Persi, Enivarco Maglioni, Nicola Galdieri, Marco Comis, Fabio Caramelli, Daniela C. Pasero, Giovanni Pala, Massimo Renzini, Massimiliano Conte, Gianluca Paternoster, Blanca Martinez, Fulvio Pinelli, Mario Frontini, Maria C. Zucchetti, Federico Pappalardo, Bruno Amantea, Annamaria Camata, Antonio Pisano, Claudio Verdecchia, Erika Dal Checco, Claudia Cariello, Luana Faita, Rubia Baldassarri, Anna M. Scandroglio, Omar Saleh, Rosalba Lembo, Maria G. Calabrò, Rinaldo Bellomo, Giovanni Landoni. Effect of Fenoldopam on Use of Renal Replacement Therapy Among Patients With Acute Kidney Injury After Cardiac SurgeryA Randomized Clinical Trial. JAMA. 2014;312(21):2244–2253. doi:10.1001/jama.2014.13573
No effective pharmaceutical agents have yet been identified to treat acute kidney injury after cardiac surgery.
To determine whether fenoldopam reduces the need for renal replacement therapy in critically ill cardiac surgery patients with acute kidney injury.
Design, Setting, and Participants
Multicenter, randomized, double-blind, placebo-controlled, parallel-group study from March 2008 to April 2013 in 19 cardiovascular intensive care units in Italy. We randomly assigned 667 patients admitted to intensive care units after cardiac surgery with early acute kidney injury (≥50% increase of serum creatinine level from baseline or oliguria for ≥6 hours) to receive fenoldopam (338 patients) or placebo (329 patients). We used a computer-generated permuted block randomization sequence for treatment allocation. All patients completed their follow-up 30 days after surgery, and data were analyzed according to the intention-to-treat principle.
Continuous infusion of fenoldopam or placebo for up to 4 days with a starting dose of 0.1 μg/kg/min (range, 0.025-0.3 µg/kg/min).
Main Outcomes and Measures
The primary end point was the rate of renal replacement therapy. Secondary end points included mortality (intensive care unit and 30-day mortality) and the rate of hypotension during study drug infusion.
The study was stopped for futility as recommended by the safety committee after a planned interim analysis. Sixty-nine of 338 patients (20%) allocated to the fenoldopam group and 60 of 329 patients (18%) allocated to the placebo group received renal replacement therapy (P = .47). Mortality at 30 days was 78 of 338 (23%) in the fenoldopam group and 74 of 329 (22%) in the placebo group (P = .86). Hypotension occurred in 85 (26%) patients in the fenoldopam group and in 49 (15%) patients in the placebo group (P = .001).
Conclusions and Relevance
Among patients with acute kidney injury after cardiac surgery, fenoldopam infusion, compared with placebo, did not reduce the need for renal replacement therapy or risk of 30-day mortality but was associated with an increased rate of hypotension.
clinicaltrials.gov Identifier: NCT00621790
More than 1 million patients undergo cardiac surgery every year in the United States and Europe alone.1 One of its most common complications is acute kidney injury (AKI), which is associated with morbidity and mortality.2 Moreover, AKI requiring renal replacement therapy (RRT) (dialysis) affects approximately 5% of patients admitted to the intensive care unit (ICU), is associated with a mortality rate of up to 60%,3 and markedly increases the cost of care.
The pathophysiology of AKI after cardiac surgery is complex. However, renal and especially medullary ischemia is a major mechanism of renal injury in this setting. Fenoldopam mesylate is a selective dopamine receptor D1 agonist, which induces vasodilation of the renal, mesenteric, peripheral, and coronary arteries.4 Unlike dopamine, fenoldopam has no significant affinity for D2 receptors; thus, theoretically, it can induce greater vasodilation in the renal medulla than in the cortex.5 Because of these hemodynamic effects, fenoldopam has been widely promoted for the prevention and therapy of AKI in the United States and many other countries with apparent favorable results in cardiac surgery6 and in other settings.7 Meta-analyses of randomized trials reported a reduction in the incidence and progression of AKI8 together with decreased use of RRT and mortality.8 However, the absence of a definitive trial leaves clinicians uncertain as to whether fenoldopam should be prescribed after cardiac surgery to prevent deterioration in renal function. To date, few interventions9 and no pharmacological agents are proven to be efficacious in treating perioperative AKI. Recent literature10,11 on treatment of cardiac surgery–associated AKI suggests that further studies with adequate statistical power are needed for fenoldopam.
Accordingly, we conducted an investigator-initiated, double-blind, randomized, placebo-controlled, multicenter trial to test whether fenoldopam infusion reduces the need for RRT, mortality rates, or both among critically ill cardiac surgery patients with AKI.
We performed a multicenter, prospective, randomized (1:1), double-blind, placebo-controlled, parallel-group study in 19 Italian hospitals in the period from March 2008 to April 2013 after local ethics committee approval. Patients aged 18 years or older signed written consent the day before surgery and were randomized while in the ICU after cardiac surgery if they had AKI, defined by the R (risk) criteria of the RIFLE (Risk, Injury, Failure, Loss, End Stage) classification12 (≥50% postoperative increase in serum creatinine or oliguria, defined as urinary output <0.5 mL/kg/h for ≥6 hours).
Exclusion criteria included previous allergy to fenoldopam, glaucoma, fenoldopam administration within the previous 30 days, use of preoperative RRT (for these patients we did not request preoperative consent), expected ICU stay less than 24 hours after randomization, RRT already started or about to start, do-not-resuscitate orders, and participation in other randomized studies within the previous 30 days (these patients had signed written consent preoperatively but were not randomized even if they developed AKI) (Figure 1).
The study protocol (in Supplement 1)13 and details of planned statistical analyses14 have been published. Our report accords with the CONSORT 2010 statement.
Patients scheduled for cardiac surgery were assessed by a member of the local research team to explain the study protocol and obtain signed informed consent preoperatively. We used a computer-generated permuted block (up to a size of 20 and a 1:1 allocation) randomization sequence. Randomization was stratified by center. Treatment allocation was prepared by an independent operator not otherwise involved in the trial and was concealed by opaque, sealed envelopes that were sequentially numbered. Race/ethnicity was determined by clinicians and collected because there is an association between ethnicity and serum creatinine levels.
After patients developed AKI as defined, they were randomized to receive placebo (saline) or fenoldopam (Corlopam; Teva Italia) by continuous infusion. Fenoldopam and placebo were identical in shape, color, appearance, and size. The participants and ICU staff were blinded to treatment allocation throughout the entire study.
The study drug was administered as a continuous intravenous infusion for a total of 96 hours or until ICU discharge or death. The starting dose was 0.1 µg/kg/min (range, 0.025-0.3 µg/kg/min). Dose increases to 0.2 μg/kg/min and 0.3 μg/kg/min, reductions (for hypotension) to 0.05 μg/kg/min and 0.025 μg/kg/min, or discontinuations (for hypotension in spite of dose reduction) were allowed, and the dose was recorded hourly. Hypotension was defined by the clinician at the bedside.
Initiation of RRT in the ICU (continuous venovenous hemofiltration or hemodialysis, according to center guidelines and protocols) was at discretion of the attending physicians because of the lack of guidelines or consensus statements to define widely accepted criteria for initiation. However, we collected the number of patients who reached predefined criteria for RRT (eTable 1 in Supplement 2) but did not receive RRT. We also used standardized criteria for ICU discharge to the postoperative cardiac surgery ward (peripheral capillary oxygen saturation ≥94% with a fraction of inspired oxygen ≤0.5 by face mask, adequate hemodynamic stability, absence of clinically significant arrhythmias, chest tube drainage less than 50 mL/h, urine output greater than 0.5 mL/kg/h, no intravenous inotropic or vasopressor therapy, and no seizure activity). From the main ward, patients were then transferred to rehabilitation centers following local standards.
The primary end point was the rate of RRT administration in the ICU. Secondary end points included mortality (ICU mortality and mortality 30 days after surgery), time receiving mechanical ventilation (hours), length of ICU and hospital stay (days), peak serum creatinine level (mg/dL), and the incidence of AKI (according to the RIFLE score I and F definitions).12
Independent monitors verified adherence to required clinical trial procedures and confirmed accurate collection of data according to good clinical practice guidelines following current national and international requirements.15
Following published literature,8,16 we hypothesized a need for RRT of 5% in the treatment group vs 10% in the control group (50% relative risk reduction) and estimated that 870 patients would have to be enrolled (435 patients per group) in the trial to achieve an 80% power at an α of .05. This number was increased by 15% (leading to a total of 1000 patients) to take into account possible loss to follow-up or withdrawal of consent.
The first 2 planned interim analyses were conducted at 250 and 500 patients, and the safety committee recommended continued enrolment for this superiority trial. The third, planned interim analysis was anticipated (from 750 to 667 patients) as requested by the major ethics committee for the trial because of the overall high mortality rate. After such additional interim analysis, the study was stopped for futility as suggested by the independent safety committee. This decision was strengthened by the higher-than-expected rate of RRT (18% instead of the anticipated 10% in the placebo group), which increased the power to detect the hypothesized 50% relative risk reduction. Data analysis followed the method suggested by Reboussin et al17 as applied in the dedicated software PASS (PASS version 08.0.11, NCSS) and based on previous work by Lan and DeMets18 with continuity correction. Adjustments for multiple interim analyses are detailed in eTable 2 in Supplement 2.
Data were stored electronically and analyzed by Stata version 13.0 (StataCorp). All 667 randomized patients completed their follow-up 30 days after surgery for major outcomes (RRT and survival). Missing data for baseline characteristics and secondary outcomes were less than 10% if not otherwise stated in tables. We did not apply any imputation for missing data. All data analysis was carried out according to a preestablished intention-to-treat analysis plan,13,14 including those few patients who were later discovered not to have satisfied the inclusion criteria (the serum creatinine increase from baseline was slightly <50% at randomization in 8 patients and 1 patient was later discovered to have recently received fenoldopam) or those who did not properly receive the study drug (they died or had RRT before receiving the study drug, or they had study drug interruption or the wrong dose per physician decision or mistake). We continued to collect all data about these patients even in the presence of a protocol deviation. We also performed per-protocol analyses as reported in eTables 3, 4, and 5 in Supplement 2.
Dichotomous data (including the primary outcome) were compared by 2-tailed χ2 test with the Yates correction or Fisher exact test when appropriate. Continuous measurements were compared using the Mann-Whitney U test. Two-sided significance tests were used throughout. Data are presented as medians (interquartile ranges [IQRs]) or as means (SDs). Means and standard deviations were used when the variables were normally distributed, while medians and IQRs were used with nonnormally distributed variables. Differences between the fenoldopam and placebo groups were assessed using univariate and multivariate regression analysis. For the univariate analysis, the cutoff level for significance was set at P < .05. Risk difference was assessed for categorical variables. Mean or percentile differences were calculated for continuous variables where appropriate. Multivariate regression analyses were performed for RRT and for the composite end point, RRT and/or death. A logistic regression model using stepwise selection was used. The prerandomization clinical data and centers were entered into the model if they had a univariate P value <.10. Treatment group (fenoldopam vs placebo) was forced into the multivariate model. In the multivariate analyses, clinical factors or potential confounding variables were expressed as odds ratio with 95% confidence intervals. In all the subgroup analyses, the heterogeneity was estimated by the χ2 test for heterogeneity and the I2 statistic.
All P values reported are 2-sided. The outcome parameters and the method of statistical analysis, including the subgroup analyses, were defined before unblinding, with the exception of the post hoc subgroup analyses that are described in eFigures 1, 2, and 3 in Supplement 2.
From March 2008 to April 2013, 9235 patients signed the written consent in 19 centers. A total of 667 (7.2%) patients developed early postoperative AKI according to trial criteria and were randomized (338 to the fenoldopam group and 329 to placebo) and analyzed according to the intention-to-treat principle (Figure 1). All patients completed their follow-up 30 days after surgery.
The 2 groups of patients had similar characteristics at baseline (Table 1 and Table 2). Mean age was 70 years, 423 (64%) were men, and 290 (48%) fulfilled the criteria for New York Heart Association (NYHA) class III or IV. Patients were randomized when reaching the AKI criteria at a median of 32 hours (IQR, 26-52 hours) after the beginning of surgery, with serum creatinine increasing from a mean (SD) of 1.1 (0.40) mg/dL to 2.0 (0.69) mg/dL in the fenoldopam group and from a mean (SD) of 1.1 (0.41) mg/dL to 2.0 (0.72) mg/dL in the placebo group.
Study drug was administered at a mean dose of 0.12 µg/kg/min (fenoldopam) and 0.13 µg/kg/min (placebo) for a total of 62 (31.3) hours and 65 (30.3) hours, respectively (Table 3), with 653 of 667 patients effectively receiving it. Six patients in the fenoldopam group and 8 patients in the placebo group died or the decision to start RRT was made after the decision to randomize and before starting the study drug (Figure 1).
Acute kidney injury (Table 3) progressed to treatment with RRT in 69 of 338 patients (20%) in the fenoldopam group and 60 of 329 patients (18%) in the placebo group (P = .47). In addition, predefined criteria for RRT were reached by 25 patients (9 in the fenoldopam group and 16 in the placebo group), with an overall risk to develop severe AKI (Table 3) of 78 of 338 patients (23%) in the fenoldopam group vs 76 of 329 patients (23%) in the placebo group (P = .99). Serum creatinine values in the 2 groups were similar during the study period (Figure 2) and after excluding patients requiring RRT. Indications and treatment variables for RRT are reported in Table 4.
Intensive care unit mortality was 58 of 338 patients (17%) in the fenoldopam group and 58 of 329 patients (18%) in the placebo group (P = .87), while 30-day mortality was 78 of 338 (23%) in the fenoldopam group and 74 of 329 (22%) in the placebo group (P = .86).
Planned subgroup analyses and per-protocol analyses (eFigure 1 and eTables 3, 4, and 5 in Supplement 2) showed no difference in the rate of RRT in the treatment and control groups (all P > .05, test for interaction = .99). The rate of RRT according to center is reported in eFigure 2 (all P > .05, test for interaction = .48). Exploratory analyses including only centers using study drug at high doses did not show any difference (eFigure 3 with all P > .05, test for interaction = .91).
Factors associated with use of RRT are listed in eTable 6 in Supplement 2, and factors associated with the composite end point, RRT and/or death, are in eTable 7 in Supplement 2.
In the fenoldopam group, there was a statistically significant (P = .003) difference in arterial blood pressure compared with the placebo group (mean [SD], 124 [20.4] mm Hg vs 120 [18.5] mm Hg) 1 hour after starting the study drug (Table 3). Moreover, the number of patients experiencing hypotension during study drug infusion was 85 (26%) in the fenoldopam group vs 49 (15%) in the placebo group (P = .001) (Table 3).
In this multicenter, double-blind, randomized clinical trial (RCT) among cardiac surgery patients with early renal dysfunction, we found that fenoldopam was not effective in attenuating the course of AKI from an early stage to initiation of RRT. There were also no significant differences in mortality rates or in any other secondary outcomes. The effect on RRT rate did not differ significantly in predefined subgroups. Fenoldopam induced greater hypotension than placebo. The study was stopped early for futility, as recommended by the study safety committee, based on a planned interim analysis.
Our results are not in agreement with previous small RCTs,23- 27 subgroup analyses of middle-sized RCTs,6,28 and meta-analyses8,16 that suggested that fenoldopam is an effective nephroprotective agent. In fact, fenoldopam was considered to be one of the few drugs with a potential beneficial effect on renal function in critically ill patients with or at risk for AKI.29- 31
Cardiac surgery appeared to be the best setting to test the putative beneficial renal effects of fenoldopam, both because of previously published small RCTs23- 27 and meta-analyses8,16 and because of pathophysiological principles. Although the etiology of AKI in cardiac surgery is multifactorial,19 the reduction in cardiac output that is frequently observed in patients undergoing cardiac surgery may play a major role by decreasing renal perfusion through a reduction in renal blood flow and through the activation of the sympathetic nervous system and the renin-angiotensin system.32 In this setting, a selective renal vasodilator appeared a logical and promising intervention.
Moreover, dopamine receptors are widely expressed in the kidney. D1 receptors are expressed in the medial layer of renal vessels and induce a dose-dependent vasodilation of renal capacitance vessels. Fenoldopam activates adenylate cyclase, causing arterial vasodilatation and natriuresis by inhibiting sodium-potassium ATPase-dependent processes at the proximal convoluted tubule and in the thick part of the ascending loop of Henle.33 Fenoldopam,34 0.1 μg/kg/min, significantly increases renal blood flow in hemodynamically stable patients with preserved renal function undergoing cardiac surgery.
The results of our study differ from those of several previous small RCTs that showed a preserved creatinine clearance in elective coronary revascularization patients,25 a decrease in postoperative serum creatinine level in patients who had a serum creatinine level greater than 1.5 mg/dL before cardiac surgery,26 and a lower incidence of AKI after complex cardiac operations.27 More importantly, our findings also differ from the results of a double-blind, single-center RCT24 with 193 patients and a subgroup analysis of a multicenter RCT,6 respectively, in which fenoldopam had been used in the cardiac surgery setting to prevent AKI24 and its progression.6
Fenoldopam is available in Europe and the United States. It was approved by the US Food and Drug Administration (FDA) in 1997 and indicated for in-hospital, short-term management of severe hypertension. Fenoldopam has not gained FDA approval for renal indications, although it has been widely used off-label in the United States for kidney protection in various settings.8,16 Our trial demonstrates that fenoldopam is not effective for the treatment of AKI in cardiac surgery and, in addition, suggests that it might not be effective for other patients with early AKI. These findings are in keeping with those of treatment with dopamine and suggest that either dopaminergic stimulation is inadequate in this setting or that the mechanism for AKI after cardiac surgery does not involve renal vasoconstriction or both.
This trial was randomized and double-blind in design with allocation concealment, thus reducing the risk of selection bias. It focused on patient-centered, objectively verifiable, and clinically relevant outcomes, thus reducing ascertainment bias. The intervention had biological plausibility and was supported by a series of single-center studies with promising results and by meta-analyses, thus justifying the initial trial hypothesis. Our results appear likely to carry external validity because patients were recruited in university and nonuniversity hospitals with the use of pragmatic inclusion criteria and few exclusion criteria. The results are also likely to have high reproducibility, as the trial protocol was simple, with routine practice maintained throughout, except for fenoldopam or placebo infusion.
Cardiac surgery–related AKI is characterized by an abrupt deterioration in kidney function after cardiac surgery as evidenced by a reduction in the glomerular filtration rate. Importantly, this deterioration is not always detected in the first 24 to 48 hours using conventional monitoring by serum creatinine levels, especially because of cardiopulmonary bypass dilution effects. Furthermore, several patients have isolated AKI not requiring ICU care. As a consequence, the low (7.2%) incidence of AKI reported in this study population does not correspond to the overall incidence of postoperative AKI after cardiac surgery, and the same selection bias applies to the apparently high incidence (19%) of RRT in our patients with AKI.
Our study has some limitations. The study was interrupted for futility, and fewer patients were randomized than planned. However, this is the largest multicenter RCT of fenoldopam. In addition, the case for the futility of the intervention was clear and subject to the recommendations of the safety committee.
Hypotension was more frequent in the fenoldopam group, suggesting that the drug may have been administered at too high a dose. It is possible that fenoldopam caused harm (eTable 8 in Supplement 2). However, our study had insufficient power to detect such harm. It is also possible that hypotension may have allowed clinicians to guess which treatment patients were allocated to receive. However, hypotension is very common after cardiac surgery, making such post hoc treatment identification unlikely. We did not collect information on the intensity or duration of hypotension. Such hypotension may have attenuated any beneficial effects that the drug may have on renal function.
Our subgroup analyses failed to identify differences. However, our subgroups were small, and a type II error cannot be excluded. The reported incidence of severe AKI requiring RRT was high; however, this was a high-risk population (patients were on average 70 years old and half were NYHA class III or IV) with a complicated postoperative course after cardiac surgery (ongoing AKI in patients not fit for ICU discharge), and the rate of RRT was high in all centers (eFigure 2 in Supplement 2), indicating that illness severity played an important role in this population. The high surgical risk of these patients, the long time on cardiopulmonary bypass and aortic cross-clamping, the need to receive support with catecholamine, and the development of low cardiac output syndrome all likely contributed to the high incidence of RRT and consequent mortality. The initiation of dialysis was left to the judgment of the attending physician because of the lack of widely accepted criteria or guidelines for initiating RRT. However, RRT initiation is unlikely to have been subject to bias because of the double-blind design of the trial. Furthermore, we found no differences between fenoldopam and placebo when using standardized criteria for RRT or the RIFLE criteria for AKI progression, both of which are independent of clinical decisions (Table 2).
Our findings differ from those of several small previous studies. However, the limitations of single-center randomized trials35 and meta-analyses are well known and may account for the difference in outcome between our study and previous trials or meta-analyses and thus explain the discrepancy between the hypothesized effect and actual results of this multicenter RCT.
Among patients with AKI after cardiac surgery, fenoldopam infusion, compared with placebo, did not reduce the need for RRT or risk of 30-day mortality but was associated with an increased rate of hypotension. Given the cost of fenoldopam, the lack of effectiveness, and the increased incidence of hypotension, the use of this agent for renal protection in these patients is not justified.
Corresponding Author: Giovanni Landoni, MD, IRCCS San Raffaele Scientific Institute, Via Olgettina 60, 20132 Milan, Italy (firstname.lastname@example.org).
Published Online: September 29, 2014. doi:10.1001/jama.2014.13573.
Author Contributions: Dr Landoni had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Bove, Zangrillo, Alvaro, Comis, Pasero, Pala, Conte, Frontini, Pappalardo, Amantea, Landoni.
Acquisition, analysis, or interpretation of data: Bove, Guarracino, Alvaro, Persi, Maglioni, Galdieri, Caramelli, Renzini, Conte, Paternoster, Martinez, Pinelli, Frontini, Zucchetti, Amantea, Camata, Pisano, Verdecchia, Dal Checco, Cariello, Faita, Baldassarri, Scandroglio, Saleh, Lembo, Calabrò, Bellomo, Landoni.
Drafting of the manuscript: Bove, Alvaro, Galdieri, Pala, Frontini, Amantea, Pisano, Verdecchia, Faita, Scandroglio, Saleh, Lembo, Bellomo, Landoni.
Critical revision of the manuscript for important intellectual content: Bove, Zangrillo, Guarracino, Alvaro, Persi, Maglioni, Comis, Caramelli, Pasero, Renzini, Conte, Paternoster, Martinez, Pinelli, Frontini, Zucchetti, Pappalardo, Amantea, Camata, Dal Checco, Cariello, Baldassarri, Calabrò, Bellomo, Landoni.
Statistical analysis: Alvaro, Galdieri, Conte, Frontini, Amantea, Pisano, Faita, Saleh, Lembo, Landoni.
Obtained funding: Bove, Zangrillo, Alvaro, Comis, Caramelli, Frontini, Amantea.
Administrative, technical, or material support: Bove, Guarracino, Alvaro, Persi, Maglioni, Pasero, Renzini, Frontini, Pappalardo, Amantea, Camata, Dal Checco, Cariello, Baldassarri, Scandroglio, Lembo, Calabrò.
Study supervision: Bove, Zangrillo, Alvaro, Pala, Paternoster, Frontini, Zucchetti, Amantea, Bellomo, Landoni.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. The following centers received a grant from the Italian Ministry of Health to conduct the study: IRCCS San Raffaele Scientific Institute, University Hospital of Pisa, Ospedale Civile “Ca’ Foncello” di Treviso, Siena Hospital, S. Orsola-Malpighi University Hospital, Città della Salute e della Scienza Hospital, University of Turin, and A. O. Spedali Civili di Brescia. The University Hospital of Pisa received the study drug from Teva. San Raffaele Scientific Institute received a donation not related to this study from Teva. No other disclosures were reported.
Funding/Support: This work was supported by a grant from the Italian Ministry of Health.
Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Additional Contributions: We gratefully acknowledge the technical contribution of Paola Zuppelli; Elisa Magistrati, MSc; Fabrizio Monaco, MD; Martina Crivellari, MD; Alice Segantin, MD; Elisa Magistrati, MD; Roberto Dossi, MD; and Marta Eugenia Sassone, MD (San Raffaele Scientific Institute); Giuseppina Angotti, MD; Angela Madeo, MD; and Francesca Gencarelli, MD (Mater Domini Hospital); Michele Clemente, MD (S. Orsola-Malpighi University Hospital); Gabriele Giovenale, MD, and Rosetta Lobreglio, MD (Città della Salute e della Scienza Hospital); Cristina Todisco, MD (Perugia City Hospital, Perugia); Ugolino Livi, MD; Filippo Erice, MD; Walter Vessella, MD; Rodolfo Muzzi, MD (Santa Maria della Misericordia Hospital); Aldo Manzato, MD (Spedali Civili di Brescia Hospital); Maria Cristina Conti, MD; Gianluigi Megliola, MD; Marcello Melone, MD (Città di Lecce Hospital–GVM Care and Research, Lecce, Italy). None received compensation for their contributions.