 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Sunil K. Sinha, MD, PhD, FRCP, FRCPCH
Department of Paediatrics
South Cleveland Hospital
Middlesbrough TS4 3BW, United Kingdom
Ralph D'Agostino, PhD
Department of Mathematics
Boston University
Boston, MA 02215
on behalf of the SELECT and STAR Investigators
To the Editor.—
We read with interest Dr Kattwinkel's commentary1 on our recently published studies2,3 comparing the new surfactant, lucinactant, to current commercially available surfactants. We write to address specific comments and explicate our purpose in undertaking this research.
As Dr Kattwinkel points out, the simplistic approach of delivering only phospholipids to infants with surfactant deficiency did not prevent or ameliorate respiratory distress syndrome (RDS). It was only with modifications of the formulation, including the addition of alcohols to facilitate the spreading of surfactant proteins (animal-derived in current formulations), that an improvement in the clinical performance of surfactants was noted. As these newer formulations were developed and tested, surfactants became one of the most studied therapies in neonatology. Yet, there are many myths regarding these data. For example, the perception that the data "consistently favored" animal-derived protein-containing "natural" surfactants is only true as they relate to reductions in the incidence of pneumothorax. This advantage has been observed across most of the head-to-head trials, and in a meta-analysis it translated into a number needed to treat of 
20.4 In contrast, a significant reduction in neonatal mortality with animal-derived protein-containing surfactants versus synthetic, non–protein-containing surfactants was observed in only 1 of the 10 comparison trials,5 and in the results of the meta-analysis comparing these 2 classes of surfactants, it was only marginally significant.4 The latter trial, which compared poractant alfa with the non–protein-containing synthetic surfactant Pumactant, was halted prematurely and led to the withdrawal of Pumactant. It is important to highlight that this trial enrolled fewer infants than our trial comparing lucinactant to poractant alfa (STAR). None of the 8 surfactant-comparison trials reporting on bronchopulmonary dysplasia or chronic lung disease showed any significant difference between the 2 classes of surfactants either alone or after pooled analyses.4 An increase in the risk of intraventricular hemorrhage was observed in a trial that compared calfactant with colfosceril and overall with animal-derived surfactants as a class in the systematic review, translating into a number needed to harm of 36.4 Unfortunately, none of these trials reported the competing outcomes of death or intraventricular hemorrhage or the occurrence of periventricular leukomalacia, an important contributor to adverse neurodevelopmental outcome. Moreover, there is no information on outcomes beyond the neonatal period for any of these trials.
The search for an alternative to animal-derived surfactants has been the focus of our group for more than a decade, culminating in the conduct of our 2 global trials of lucinactant, SELECT and STAR. These trials were designed to evaluate how this novel peptide-based surfactant would perform across a variety of regions and neonatal practices in a "real-world" environment compared with other commercially available surfactants. The SELECT and STAR trials were conducted according to current and contemporary clinical-trial standards. The merits of clinical studies should be judged based on how well they are designed and conducted, not on preconceived notions of acceptability depending on their origin. Many well-conducted studies addressing crucial issues in neonatal and perinatal medicine have been conducted primarily or exclusively in countries outside the United States, including the OSIRIS trial of colfosceril and trials of delivery-room management of infants with meconium-stained amniotic fluid.6 These studies have had a major impact on current neonatal practice, not only in the United States but worldwide. Implying that only studies conducted in the United States are worthy of consideration would limit contributions to the scientific literature.
The approaches taken in the conduct of our trials exceeded in detail and quality any of the previously published studies that compared surfactants. Both trials were designed with internationally recognized neonatologists who served as members of the respective study steering committees that provided oversight of the trials. A single data-safety–monitoring board was established to monitor both trials. The principal investigators of both trials were intimately involved throughout the trials and conducted investigator meetings in addition to visiting most individual sites to ensure adherence to the protocol. In contrast to the majority of previous surfactant-comparison trials, which were not blinded, we focused on strict adherence to blinding procedures and used a central adjudication process for the primary outcomes in our pivotal SELECT trial. Given the significant lack of agreement in establishing important neonatal diagnoses described in the literature, we felt very strongly that the use of strict adjudication methods would improve assessment of the primary and key secondary end points. Ventilatory-management guidelines that reflect evidence-based current practice were also developed (see addendum of ref 3 for distinguished committee membership), used, and monitored throughout the trial. The US Food and Drug Administration (FDA) also provided guidance and, on completion of the pivotal SELECT study, formally audited the data and reviewed the trial's execution.
We wish to clarify, as clearly stated in our article, that the SELECT trial was designed as a superiority trial that compared lucinactant to colfosceril, with beractant serving as a reference arm in the study. The hypothesis tested was that sinapultide, the novel synthetic peptide that mimics human surfactant protein B (SP-B), when added to a formulation of phospholipids (the surfactant preparation, lucinactant) would confer superiority over a colfosceril surfactant formulation of phospholipid only, which contains no proteins. We applied uniformly rigorous clinical design, execution, and statistical methodology to all aspects of the study, including developing the primary study hypothesis, obtaining the study data, and testing the statistical comparisons on the primary outcomes measures. The sample size of the SELECT study and the statistical assumptions of a power of 94% to detect differences between these 2 primary comparators are clearly much more robust than those of any previous surfactant trial. The comment that the SELECT study was "designed as [a] ‘noninferiority’" study and was "underpowered" is not accurate. The beractant treatment group was included as a reference arm per request by the FDA so that it could be used for internal validation of the clinical trial. In fact, our data provide the only comparative information in the literature on the prophylactic use of colfosceril and beractant and are consistent with observations from treatment trials of colfosceril and the animal-derived surfactant, beractant, in which no major outcome differences were otherwise seen except for air leak. Moreover, our SELECT study of RDS prophylaxis (N = 1294) and the Vermont Oxford Network rescue study, which used colfosceril and beractant (N = 1296), are the 2 largest surfactant-comparison trials conducted to date.
We also wish to clarify that, as stated clearly in our article, the STAR trial was a study designed to demonstrate noninferiority of lucinactant compared with poractant. Noninferiority of lucinactant was established even with only half of the planned patients. From a statistical point of view, early termination of the study with only half of the planned patients being enrolled could reduce its power. Any statistically insignificant findings resulting from analysis of this reduced sample could be attributed to the smaller sample. However, any statistically significant result would be valid and is even more convincing, because it is based on a smaller sample size. Power of the study is a concept that is more appropriate for study design and is meaningful before final analyses of the data. After a study is completed and final analysis is performed, power is not a factor that has an impact on the interpretation of statistically significant results.
Consistent with other large multicenter international clinical trials, the number of patients per center in our trials varied from one center to another. In addition, although there were many centers from outside the United States, the quality of the clinical trial was not compromised; the study sites passed rigorous inspection by regulatory agencies (the FDA and European Agency for the Evaluation of Medicinal Products). In addition, the statistical analysis, after adjusting for center effect, still demonstrated the superiority of lucinactant compared with colfosceril in the SELECT trial and noninferiority of lucinactant compared with poractant in the STAR trial. The diversity of the centers is not a "fault" in the design of these studies but rather is an indication of the generalizability of the results.
Administration of animal-derived surfactants results in a more rapid decrease of ventilatory requirements and supplemental oxygen than older-generation synthetic surfactants such as colfosceril. This faster response has been attributed primarily to the presence of surfactant proteins, particularly SP-B, although none of the trials that compared animal-derived surfactants to colfosceril used similar doses of phospholipids. In our study, more rapid reductions of oxygen requirements were also observed among infants who received lucinactant than those in the colfosceril group, which is in keeping with in vitro data demonstrating rapid reductions of surface tension by a mixture of phospholipids and the KL4 peptide (sinapultide). The same response was seen for infants who received beractant. In follow-up comments about the need to know more about the fate of the KL-4 peptide, we believe that we addressed this in the article. We referred to recently published data demonstrating that sinapultide is not only taken up slowly by human type II pneumocytes, but, more importantly, it does not interfere with endogenous surfactant protein expression, the formation of lamellar bodies, or other surfactant-related functions of type II cells.7
Data from our trials and supporting in vitro studies clearly demonstrate that the synthetic peptide, sinapultide, mimics the most important function of SP-B. When added to surfactant phospholipids, sinapultide results in a novel surfactant that outperforms non–protein-containing synthetic surfactants and functions at least as well, if not better, than some animal-derived surfactants. Widespread availability of this surfactant will potentially obviate the need for use of animal-derived products and hence reduce any potential risk associated with such products. These considerations do not even address the issues related to the ethical and religious concerns of administering animal-derived therapeutics when a suitable alternative exists. We can only wonder what parental choices would be if appropriate information were provided regarding not only the comparative data from the lucinactant trials but also the origin of the surfactant preparations commonly in use, in the context of availability of this new alternative.8 In our opinion, data from our clinical trials supporting the use of lucinactant as an alternative therapy for preventing and treating RDS represent a major breakthrough.
REFERENCES
Related articles in Pediatrics:
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
||
|
||
Home | My Pediatrics | Journal Information | Current Issue | Past Issues | Subscriptions & Services | Contact Us Click here for faster international access © 2006 American Academy of Pediatrics. All rights reserved. |
||
|
||
