Pediatric Research and Scholarship: Another Gordian Knot?

Abstract

Appearing in this issue of the journal are four clinical papers on the use of dexmedetomidine in children (1–4). Previous studies of the use of this drug in children have dealt with sedation in the intensive care unit (5), prevention of postoperative agitation associated with sevoflurane anesthesia (6,7), and sedation for noninvasive procedures (8–12). These articles also illustrate a number of subtle concerns that are relevant to drug development in children and the ethical principles that guide research in human subjects as proposed by the National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research, and summarized in the Belmont Report. In a randomized study of 30 pediatric cardiac surgical patients anesthetized with isoflurane and undergoing cardiopulmonary bypass (CPB), Mukhtar et al. (1) evaluated the effects of a continuous dexmedetomidine infusion on patients’ underlying hemodynamics and neuroendocrine system compared with patients receiving a saline control infusion. The authors infused dexmedetomidine after anesthetic induction until the termination of CPB. Dexmedetomidine attenuated the increase in the neuroendocrine markers of stress caused by sternotomy and CPB. In addition, dexmedetomidine provided better intraoperative hemodynamic stability in children with congenital heart disease. The report focused on hemodynamic outcome and did not address how this drug affected tracheal extubation time, recovery, length of hospital stay, or long-term outcome. Thus, the reader is left to ponder whether the reduction in the cortisol and norepinephrine, markers of neuroendocrine response, affect outcome. Moreover, we do not know how to titrate dexmedetomidine or if the optimum dose was administered. Are there dexmedetomidine dose-related adverse effects in this population? Should the dose be altered at the start of CPB? What is the effect of the fluctuation in the patient’s temperature (37° to 25° and return) on plasma levels achieved using a standard infusion rate? Why was dexmedetomidine not continued until the end of surgery? Why did the dexmedetomidine infusion not allow for less anesthetic administration, as had been shown in an adult study (13)? Did dexmedetomidine improve hemodynamic stability after bypass? Although this study measured biochemical endpoints to assess stress response, can we infer that the effect of this drug will be as salutary as the effect of opioids in the mortality of children having surgery for congenital heart disease? These questions were not answered because the study was not funded, limiting the number of subjects and the resources available to the investigators. Had this study been funded by the pharmaceutical manufacturer, the authors could have enrolled more patients, perhaps included other study centers, assayed blood for dexmedetomidine concentrations, developed pharmacokinetic models of dexmedetomidine in this population, and performed a pharmacodynamic concentration-versus-response analysis. In a study of 60 children undergoing magnetic resonance imaging, randomized to receive either dexmedetomidine or propofol for sedation, Koroglu et al. (2) noted that children sedated with a dexmedetomidine infusion had more stable hemodynamics than children sedated with a propofol infusion. Both drugs decreased mean arterial blood pressure, heart rate, and respiratory rate, but these changes were greater with propofol than with dexmedetomidine. However, although statistically significant, these differences may not be clinically important. Considering that 4 of the 30 children receiving propofol had low oxygen saturations, which were not observed in any child receiving dexmedetomidine, perhaps the real benefit of this drug is in the reduced effect on ventilation. On the other hand, children sedated with propofol had faster onset and recovery and discharge times than did those who received dexmedetomidine. How can we assess whether there is a safety benefit to the faster onset, recovery, and discharge with propofol that offsets its more profound depression of ventilation? Even more concerning, there are no prior dose versus response studies of dexmedetomidine in children, so we have no way of knowing whether the empirically selected doses used in this study were optimum and equipotent or represented a relative overdose of either drug. As in the study by Mukhtar et al., we find investigators struggling to provide dexmedetomidine dosing information for children without the support of the drug manufacturer. In a “pilot” study of 62 children undergoing computed tomography imaging, Mason et al. (3) reported that a continuous infusion of dexmedetomidine caused the heart rate and mean arterial blood pressure to decrease by 15%. No patient had a serious “Type A” adverse event (i.e., no patient required resuscitation, or had a cardiovascular complication, decreased oxygen saturation, allergic reaction, an episode of aspiration, or required a reversal drug). However, 10 children (16%) had a “Type B” adverse event (i.e., required greater than 3 h to recover, agitation, or vomiting). This pilot study was initiated to find a better way to sedate children. Dissatisfaction with the standard sedation recipe of pentobarbital, midazolam, and fentanyl sedation for children undergoing radiological imaging studies led these authors and other members of the institution’s staff to adopt an institutional change from pentobarbital to dexmedetomidine for sedation of children undergoing radiological imaging procedures. The authors prospectively gathered information for all sedated patients in the institution as a part of their normal practice to support institutional quality assurance. All patients consented to sedation, and the institution granted the investigators permission to access its database and to publish their findings. The “Belmont Report,” published April 1979, specifically addressed “the boundaries between biomedical and behavioral research and the accepted and routine practice of medicine” (http://ohsr.od.nih.gov/guidelines/belmonthtml). The report distinguishes medical practice, defined as “interventions. . .designed solely to enhance the well-being of an individual patient. . .that have a reasonable expectation of success” from research, defined as “an activity designed to test a hypothesis, permit conclusions to be drawn, and thereby to develop or contribute to generalizable knowledge.” The children in this study unquestionably received dexmedetomidine as part of medical practice; the institution felt dexmedetomidine was a better therapeutic choice than the alternative. However, children and parents were not advised that they were receiving a new drug for an off-label indication. Because most pediatric drug usage is off-label anyway, such a disclosure has little meaning. However, the institution used collected data to support their quality assurance efforts. At what point does this collection of data become research? The Belmont Report suggests that it becomes research when the purpose, even if a secondary purpose, is to create generalizable knowledge. If this was envisioned at the outset, then there should have been IRB approval at the outset. Parents should have signed informed consent for participation in the research. However, as the institution had decided that dexmedetomidine was safer than pentobarbital, the alternative choice was not a different therapy, but either refusal to permit use of the data or no sedation (and perhaps no procedure) at all. Many institutions initiate policy changes that alter patient care and then collect data to evaluate that decision. This is, of course, exactly how medicine should be practiced. However, this practice falls between medical practice and research, as one goal of the program is to create knowledge within an institution about the new practice paradigm. Is this generalizable knowledge? Should patients be informed of the change? We do not have concerns about this specific study or the appropriateness of the manner in which IRB approval was sought. Anesthesia & Analgesia will not publish studies that do not meet the highest standards of ethical research. Rather, we use this study to demonstrate the complex relationships among changes in medical practice, prospective gathering of data for continuous quality improvement, and publication of such data to create generalizable knowledge. Mason et al. sought IRB approval to turn data collected by the institution’s quality assurance process into generalizable knowledge. In our view, that was a reasonable point at which to make the transition from medical practice to research. If the manuscript by Mason et al. is challenging in regards to studying off-label use of drugs in children, the case report by Rosen and Daume (4) should heighten our awareness further. This case report exemplifies the difficulty that pediatric practitioners face when the drug label lacks pediatric safety, toxicity, and dosing information. Rosen and Daume describe the anesthetic management of a 3-yr-old, 14-kg, uncooperative child undergoing magnetic resonance angiography 2-1/3 yr after tetralogy of Fallot repair. After unsuccessful sedation with an oral benzodiazepine, the authors administered dexmedetomidine 4 μg/kg intranasally. They calculated this dose from a study of 12 adult volunteers, which reported a dexmedetomidine bioavailability of 16% to 82% after peroral and buccal routes of administration, respectively (14). The authors then administered an IV infusion but at a dose of 1 μg · kg−1 · min−1 over 10 min rather than a bolus of 1 μg/kg over 10 min, resulting in a 10-fold overdose. This resulted in hypertension, as reported by Ebert et al. in adults (15), and markedly prolonged emergence (4 h). However, the child tolerated the overdose well and was discharged home the same day. This makes for an interesting case report: not many anesthetic drugs would be well tolerated if given in a 10-fold overdose. The report provides a valuable clinical observation, and we commend the authors for courageously sharing their difficulties. However, we wonder whether it is appropriate to publish an off-label use of a drug, particularly when it is further compromised by an inadvertent 10-fold overdose. Does publishing such reports reward clinicians for using drugs off-label and making potentially dangerous dosing errors? Is Anesthesia & Analgesia guilty of off-label promotion of dexmedetomidine in suggesting, based on a single case report, that a 10-fold overdose of dexmedetomidine is well tolerated? Or does this case report provide an important data point, one that is otherwise unlikely to ever be repeated in any study? As dexmedetomidine does not have a labeled indication in pediatric patients, we wish to review issues related to this type of research and scholarship in clinical medicine. Nearly 85% of drugs listed in the Physicians’ Desk Reference do not have labeled indications for children. Nonetheless, it is common practice that physicians prescribe pharmaceuticals off-label for children. These off-label prescriptions include dose adjustments (often by weight or body surface area) via the labeled route for adult patients, use for illness or conditions not included in labeling (e.g., clarithromycin for “antiinflammatory” activity instead of antimicrobial activity), and use via routes of administration not included in the product label (e.g., IV hydroxyzine, nasal fentanyl or midazolam, clonidine in regional nerve blockade, or neostigmine in epidural analgesia). Recent studies note that pharmacokinetics and pharmacodynamics in neonates and children may not be well predicted by adult values and that children require significant adjustments in dose or interval for optimal efficacy and to minimize side effects (20). Prior to the Best Pharmaceuticals in Children Act, the United States (US) Congress authorized an extended 6-mo patent protection to sponsors for conducting pediatric studies. However, this was insufficient incentive for many sponsors because of the costs and risks of pediatric trials. Often by the time a pediatric trial has been conceived or performed, the drug is nearing its patent expiration and the trials cannot be completed in time to provide a return on investment for the sponsor. With the common practice of off-label use and the paucity of data, the US Congress passed the Best Pharmaceuticals for Children Act in 2002. This legislation (and its funding) empowered the US Food and Drug Administration (FDA) and the National Institutes of Health to fund studies of generic pharmaceuticals in children in which the sponsors would not support the study because of costs, risks, and lack of economic incentives. These trials have recently begun to evaluate agents in everyday use (e.g., lorazepam, sodium nitroprusside), and others are being planned, including studies to evaluate spironolactone, baclofen, azithromycin, dobutamine, dopamine, furosemide, heparin, rifampin, lithium, and others. There are multiple levels of safeguards to protect children as research subjects. These include the US FDA requirement for a sponsor to present appropriate preclinical data in support of awarding an Investigational New Drug number. Sponsors also have an ethical responsibility to inform participants in human research trials about known risks associated with their drugs. Unfortunately, some side effects are not predictable in children because of unanticipated differences in pharmacokinetics and pharmacodynamics. Therefore, significant experience in adult subjects may poorly predict drug behavior in children, as recently seen with rapacuronium Were these safeguards available to Drs. Mukhtar et al. (1) and Koroglu et al. (2) when they did their clinical trials in children? Almost certainly not. Did Dr. Mason et al. (3) have access to the available preclinical safety data from the manufacturer when making their institutional change to use dexmedetomidine for safety? Probably not. The FDA would consider it grossly irresponsible for a company to conduct these trials without first reviewing the preclinical safety data. However, the authors conducted the trials, and Anesthesia & Analgesia accepted their manuscripts for publication because we are desperate for dosing information to guide the use of drugs in children. Does that justify performing the trial and publishing the results? Alternatively, has the FDA set an unreasonably high bar for pharmaceutical companies to conduct clinical research, one which discourages companies from conducting trials in children? The FDA is responsible for drug evaluation and research in the US. Once data are received in support of a novel drug, the FDA evaluates the adequacy of the data to support labeling: “Labeling shall be informative and accurate without being promotional, false or misleading” (16). Specifically, labeling calls for “adequate and well controlled studies,” which is usually interpreted to mean at least two randomized, double-blind, placebo-controlled trials. Case reports and clinical observations published in the literature are not accepted as the basis for new drug labeling, although they may be used to identify possible risks. Expanding the product label, such as adding an indication for use of the drug in children, requires providing supportive data to the FDA. When a new drug is not expected to be used in children, then the sponsor is not required to perform pediatric studies. Sometimes sponsors can delay studies in children until the drug has been approved (Phase IV). As children are a minority of the population, sponsors often seek to delay studies in children as long as possible. However, FDA guidance to industry states that “In general, drugs should be studied prior to approval in subjects representing a full range of patients likely to receive drug once it is marketed. . .” (17). Also, more current regulations specifically require labeling in subpopulations, including the pediatric population (18). The FDA may demand information on a new drug application suitable for use in the pediatric population (19). The FDA received accolades when it accelerated the approval of drugs to combat human immunodeficiency virus/acquired immune deficiency syndrome and other conditions. However, the agency has been bruised by vocal opponents in industry who insist that the process is still too cumbersome. But certain drugs that have been approved have since been withdrawn from the market for safety reasons, and public safety advocates are equally outspoken in their criticism that the relationship between the FDA and the pharmaceutical industry is too cozy. No wonder the FDA is quite careful and deliberate as regards pediatric research! IRBs are empowered to evaluate research design, require appropriate patient protections, and stipulate specific language in informed consent and assent documents. Investigators are expected to minimize the risks to research subjects in the design and execution of clinical trials. Parents are responsible for understanding the merits of a trial and whether the risk-benefit profile is sufficiently compelling to allow their child to participate. Parents must be provided factual, fair, and comprehensible information to assess the risks of participation. Additional regulations set high standards for studies that put a child at greater than minimal risk, and such studies that are of no potential direct benefit to the child must be approved by a specially convened federal review panel. Because a child may not be denied “standard treatment” in the conduct of research, this may ethically preclude a placebo arm if it puts a child at increased risk over the standard of care or experimental treatment. Therefore, the “placebo-controlled” trial should not be the only standard by which trial design is contemplated, implemented, or refereed in review for publication. What are the responsibilities of investigators, academic medical centers, journal reviewers, and editors in publishing clinical research and case reports? Apparent or real financial conflicts of interest have been the center of attention for many years. These conflicts have generated certain safeguards in trial design and investigator disclosures. Trial designs should be sufficiently meritorious and of sufficient power to study the effectiveness of a novel drug. Do we have any way of knowing whether the pediatric dexmedetomidine trials reported in this issue are adequately powered to support safety and efficacy? There is also an ethical obligation to publish research studies, regardless of the outcome. At the very least, all trials, whether investigator initiated or pharmaceutical industry-sponsored, should be entered into a public trials registry. There are dilemmas in the responsibility of reviewers and editors to publish clinical research. Anesthesia & Analgesia will not publish any research, defined as a study to develop generalized knowledge, without IRB approval and written informed consent. Or will it? Mason et al. had retrospective IRB approval, including a waiver of written informed consent, to publish data prospectively gathered as part of a quality assurance program. Are these data as reliable as prospectively gathered research data? If the data are as reliable, then perhaps this was research all along, as defined by the Belmont Report. If so, are these patients entitled to less protection? Should our reviewers speculate about the mindset of authors, attempting to assess whether generalizable knowledge was the intent an along? Anesthesia & Analgesia is also publishing a case report by Rosen and Duame that appears to not need any IRB approval or informed consent, as it describes medical practice as defined by the Belmont Report. However, it is entirely anecdotal. Is anecdotal information better than no information at all? If anecdotal data are not published, then clinicians will have only their personal or local practice to guide their use of drugs in children. We believe that poorly serves the needs of our pediatric patients. Finally, we again turn to the Belmont Report for guidance. The report summarizes basic ethical principles “relevant to the ethics of research involving human subjects. These principles are respect for persons, beneficence and justice.” We submit that these principles apply not only to research conducted in adults but research that is not conducted in children. Respect for persons means that individuals are permitted self-determination. Adults are told the risks and benefits of research studies. They are given the opportunity for informed consent. Respect for persons permits adults to make an informed choice. However, children and their parents are not permitted to make an informed choice. The choice has been made for them by pharmaceutical manufacturers when studies are not done in children. Lacking alternatives, there is no choice, no self-determination. Respect for persons is violated. “Beneficence” is defined by the Belmont Report as an obligation on both individual investigators and society at large to secure the well-being of both research subjects and patients. The report uses research in children as a specific example of the complexities of beneficence. Beneficence requires that studies be ethically designed, that they provide a potential benefit to the child, and that the risks to the child be minimized. Disclosures must be honest, and investigators must be forthright about potential conflict of interest. However, how do we get comfortable with most early studies in pediatric patients, where the risks to the pediatric subjects are unknown, and the benefits entirely speculative? We did not ask Drs. Mukhtar et al. (1), Koroglu et al. (2), or Mason et al. (3) to assess the beneficence of their studies. Were we remiss? “Justice” is perhaps the most interesting concept when applied to children. The Belmont Report defines justice with the rhetorical question “Who ought to receive the benefits of research and bear its burdens?” It goes on to observe that “An injustice occurs when some benefit to which a person is entitled is denied without good reason.” Children are unambiguously entitled to the same level of health care as adults. The principle of justice is violated when they are not afforded access to effective treatment because the necessary research has not been conducted to support pediatric labeling. Pharmaceutical companies are the beneficiaries of off-label use of drugs in children. It is not unjust to require companies to study children in a timely manner during the drug approval process, given that they stand to reap direct rewards from the use of their drugs in children. What can we learn from the four papers in this issue of the journal? 1) Not enough pediatric dexmedetomidine studies have been done in children. 2) Basic dexmedetomidine dose versus response curves are missing. 3) In the 6 yr since approval, clinical trials involving appropriate doses and appropriate routes of dexmedetomidine administration for children have not been done, despite an estimated 400,000 adult patient days of dexmedetomidine infusions (personal communication, Hospira Worldwide Inc.). 4) Most pediatric studies involve small numbers of patients in investigator-initiated trials that do not meet the FDA’s definition for adequate and well controlled studies (i.e., could not be used to change the drug label). Where have the FDA, the sponsors, and the pediatric advocacy groups, (American Academy of Pediatrics and Society for Pediatric Anesthesia) been? 5) Based on a very modest number of studies, it appears that dexmedetomidine works similarly in children as it does in adults (a not-too-surprising finding). 6) A shameful approach to drug development continues to be the norm in pediatric medicine. The situation with dexmedetomidine dosing guidelines for children is the norm rather than the exception. The issues in this editorial are not intended to be critical of the four pediatric dexmedetomidine articles in this issue, nor are we expressing specific criticism of Hospira, the manufacturer of dexmedetomidine. Rather, we are taking this opportunity to address several fundamental problems that are, at least in part, unique to pediatric clinical research. The Gordian knot (the mythical intricate knot nearly impossible to untie, or a problem most difficult of solution) is not impossible to unravel, but the solution is complex. In the spirit of protecting children, the FDA may require further adult data to document drug safety before approving pediatric studies. Sponsors must commit to pediatric development early and assemble a team of experienced investigators in the design and implementation of trials. Scholarly manuscripts must be written, reviewed, and published. Children must receive a higher priority than indicated by the size of the pediatric market. If this means providing more economic return on investment for the sponsors (with appropriate caveats), then this should receive further discussion. The ethical responsibilities of publishing studies without controls, studies that are underpowered, or studies that report on quality assurance, present dilemmas for journals. It may be time to develop a consensus statement of advocates for children and obtain commitment of the sponsors and more resources at the FDA to improve this enigmatic situation on behalf of children. ACKNOWLEDGMENTS Dr. Tobin has participated in clinical trials sponsored by Alza, Abbott Laboratories, Baxter Healthcare, GlaxoSmithKline, and Organon, Inc. and has received research support from these sponsors for conduct of the trials. He has been a member of the US FDA Anesthesia and Life Support Advisory Committee, and currently is a consultant to the FDA. Dr. Davis has acted as a paid consultant for Abbott Laboratories, Hospira, Baxter Healthcare, GlaxoSmithKline. He has participated in clinical trials sponsored by Abbott Laboratories, Baxter Healthcare, GlaxoSmithKline and Organon, Inc. and he has received research support from these sponsors for conduct of the trials. Dr. Davis is also the Pediatric Section Editor for Anesthesia & Analgesia. Dr. Shafer, Editor-in-Chief of Anesthesia & Analgesia, has performed Phase I clinical trials on dexmedetomidine for Orion-Farmos. He is presently Chair of the FDA Anesthetic and Life Support Drugs Advisory Committee.