Cpt_0607_cy.indd

PERSPECTIVES
practice presents a threat to public health right balance in oversight of physician opioid discipline of physicians: assessing state and safety—have the ability to issue a prescribing for pain: the role of state medical medical boards through case studies. <http:// boards. J. Law Med. Ethics 31, 21–40 (2003).
aspe.hhs.gov/daltcp/reports/2006/stdiscp.
Hoffmann, D.E. & Tarzian, A.J. Achieving the right balance in oversight of physician opioid 10. Federation of State Medical Boards. Protecting prescribing for pain: a survey of state medical the public: how state medical boards regulate boards. J. Med. Licensing Discipline 89, 159–170
and discipline physicians. <http://www.fsmb.
used. It is more oft en the case that the org/smb_protecting_public.html>. Accessed Bovjberg, R.R., Aliaga, P. & Gittler, J. State to the DHHS study, case closure times “vary considerably according to how far through the disciplinary process a case See ARTICLE page 807
proceeds. Nationally, cases resolved before or during investigation averaged 180 days from intake to closure in 2003, “Personalized Medicine: Elusive
425 days for cases closed aft er investi-gation but before hearing, and 675 days Dream or Imminent Reality?”: A
to reach hearing.”9 Although a signifi -cant majority of cases are closed during Commentary
investigation, for those cases that require a hearing, board efficiency is poor. Moreover, many boards have a backlog of uninvestigated complaints.9 I was invited to comment on the article in this issue by Larry Lesko,
“Personalized Medicine: Elusive Dream or Imminent Reality?”,1
especially with respect to three questions. There are several possible
and Willis, state boards would need an alternative, perhaps a “fast track” review approaches to discussing these questions. One is to simply answer the
question as directly and succinctly as possible; another is to expand
related to the prescribing of opioids for on the questions. For the purpose of this Commentary, I try to do both.
pain. Ideally such a committee would include pain-treatment experts. Argu-ably, boards would need additional In his article, Dr. Lesko lists the following Direct answers
as three of the many unresolved questions “No” would be my simple answer to the fi rst question. I do not believe that RCTs are necessary to establish the usefulness of 1. Are prospective, randomized con- a predictive safety (or a dosing safety) test. trolled trials (RCTs) necessary to qualify RCTs are expensive and used to establish or validate a predictive genetic test (e.g., CONFLICT OF INTEREST
for CYP2C9 activity) to inform dosing in that of a placebo. Safety tests are expected The author declared no conflict of interest.
clinical practice or for regulatory policies to be predictive for an individual. Safety signals are recognized by following treated 2. To what extent can prospective or Reidenberg, M.M. & Willis, O. Prosecution of retrospective observational studies sup- specifi c genetic variants associated with an physicians for prescribing opioids to patients. port genotype-phenotype associations for Clin. Pharmacol. Ther. 81, 903–906 (2007)
adverse event (AE)—including those due Code of Federal Regulations 21, §1306.04(a).
the purpose of qualifying or validating a to dosing related to drug metabolism vari- 3. United States v. Hurwitz, 459 F.3d 463, 468 (4th predictive genetic test to determining opti- ants—may well be the same as the allele 4. Hoover v. Agency for Health Care Administration, frequency in the placebo group. However, 676 So. 2d 1380, 1380–85 (Fla. Dist. Ct. App. 3. What clinical end points are appropri- placebo-treated patients do not receive the ate for qualifying or validating predictive Federation of State Medical Boards. Model genetic tests when given the choice between guidelines for the use of controlled substances of a drug-specifi c AE, even if they carry for the treatment of pain. <http://www.
a clinically relevant, causal biomarker, a painpolicy.wisc.edu/domestic/model.htm> surrogate end point used as a substitute for genetic variant is refl ected clinically only a clinical end point, or a clinical outcome? Federation of State Medical Boards. Model policy for the use of controlled substances for the treatment of pain. <http://www.fsmb.org/ 1Genetics Research and Development, GlaxoSmithKline, Research Triangle Park, North Carolina, USA. pdf/2004_grpol_Controlled_Substances.pdf> Correspondence: A Roses (allen.d.roses@gsk.com) Hoffmann, D.E. & Tarzian, A.J. Achieving the CLINICAL PHARMACOLOGY & THERAPEUTICS | VOLUME 81 NUMBER 6 | JUNE
PERSPECTIVES
defi nes the serious need for long-term, could be confi rmed in long-term prospec- tive observation studies rather than RCTs. and should be tested for CYP2C9 variants using a comprehensive screening tool that already been in place for years. An exam- at least covers tag single-nucleotide poly- abacavir as a treatment for HIV, which led to the identifi cation of a genetic marker genetic screening tool.2 When such chips begin with the fi rst patient treated in phase or similar platforms become inexpensive, I and continue through all clinical trials Box 1). It is thus not a totally new concept
their routine application (when mandated to monitor treated patients aft er marketing cussions of and proposals for new schemes has begun, including all patients in a man- will be available for established tests or datory provisional approval group (Figure
that would require fewer patients to defi ne 1). Predictive safety markers ascertained
prescribed “provisional” or “conditional” sake, this will no doubt be part of routine approval process involving continued sur- patients during the provisional approval veillance for safety signals in all patients. period. For a drug with the “approved” records and monitored databases. In fact, No one would argue that 10,000 patients is label, relevant prospective surveillance initial-phase, commercially available SNP platforms are being utilized routinely by cacy for major illnesses to be available to improve safety of coumadin in clinical variants of interest during clinical tri- earlier—under active and regulated risk- practice, based on retrospective studies, als for a drug as well as for combination interactions in adjunctive trials. Th observational in nature, and do not man-date RCTs but do require consistent and Box 1 Abacavir risk-management example.
Abacavir is a drug that is eff ective in treating HIV/AIDs that was developed in the early 1990s, when there was a huge unmet medical need for lifesaving drugs. There were also organized and eff ective patient advocacy groups who engaged the political system. Abacavir had been shown to be eff ective, but a signifi cant proportion of patients also developed a hypersensitivity syndrome (HSS). Abacavir was granted conditional approval in the late 1990s, subject to a comprehensive risk-management program that continues to this day. HSS was defi ned from studies of approximately 200,000 patients receiving the drug between 1996 sive safety databases by surveillance of and 2000 (ref. 8). In 2002, a report published in Lancet from an academic group reported the association with HLA-B5701 in 14 of 18 patients with HSS.9 In the same month the sponsor Large clinical databases, such as those of studies reported in Lancet that 39 of 84 patients were HLA-B57 positive.8 The regulated risk-management program has continued to test the marker in multiple eff ectively for well-informed provisional populations in order to evaluate prospective diagnostic value in diff erent ethnic groups approval studies without reinventing de and to gather information for the regulators. The drug is now widely used for HIV treatment. novo the clinical trial monitoring schemes HSS can be recognized clinically with clinical follow-up appointments, and, when HSS is observed, abacavir can be safely stopped and patients warned never to receive it again.10 stant is the clinical care and monitoring Patient medical records can refl ect this “allergy.” Deaths due to HSS have virtually disappeared environment for prospective surveillance since the mid-1990s. The predictive value of the test is still being evaluated prospectively in that provides both the numerator (safety observational studies of multiple populations receiving the drug. This observational program calls for periodic clinical examinations for HSS. It has also been demonstrated that the test characteristics are diff erent in diff erent ethnic groups and, although highly effi Caucasians, would be dangerous to rely on as a sole predictive test in other ethnic groups cially those that will be mandated by reg- in which HIV is epidemic.11 The cost to GlaxoSmithKline of maintaining this one-off risk- management program is signifi cant, but the drug remains on the market, available to the indications, could also be used to identify “right” patients. In perspective, the cost of the program over 7 years is smaller than the cost new safety data, as well as testing for pro- of a single phase III clinical trial. The cost to patients not receiving the drug would have been spective prediction characteristics.
more than signifi cant; it would have been tragic.
In considering the third question, it is Similar conditional approvals for cancer drugs exist. Surely treatments with RCT- critical to understand that identifi cation of cacy for Alzheimer’s disease, depression, chronic obstructive pulmonary a safety marker requires prospective obser- disease, rheumatoid arthritis, diabetes, and other major diseases could similarly benefi t from provisional approvals and closely monitored safety surveillance.
actually quite simple—no matter what may VOLUME 81 NUMBER 6 | JUNE 2007 | www.nature.com/cpt
PERSPECTIVES
defi ned in the drug label for selection of patients to treat. Aft er several decades of involvement with genomic technologies and their application to diagnosis and pharmacogenetics, I have formed views on drug development and surveillance that, in the current electronic world, seem ideas are not original; elements have been discussed in statements by offi Food and Drug Administration (FDA), reports by the Institute of Medicine and tion of various components of the health- Figure 1 A rendition of drug development with provisional or conditional approval. A GAO report says
that a “new expedited process” would “serve as an incentive to increase drug development”.7 AEs, adverse events; Epi, epidemiology; FTIH, first time in humans; PGX, pharmacogenetics.
agencies (in the United States with suf-fi cient government support to regulate and safeguard health care in the twenty- be concluded from retrospective statistical fi rst century). It would involve formal data, predictive accuracy can be evaluated utilization of comprehensive health-care e power of counts is the clinical response to use of systems to provide electronic data in real an epidemiologic conclusion, usually per- the drug. Even if several drug candidates time as physicians are caring for patients formed retrospectively, could be confi rmed for its usefulness in prospective epidemiol- patients, full clinical trials are necessary ogy studies in populations using uniform, the curve in the use of electronic health for safety of a drug for at least the fi rst 200,000 patients need to be observed sys- tematically.5 Serious clinical safety signals, including those due to dosing, need to be defi ned in real time using as few aff ected Figure 1 is a simplistic diagram of
tested prospectively to defi ne patients who predict outcomes can shorten clinical tri- als once they have been established, but those being developed to treat important the time frame for the evaluation of sur- in-a-lifetime, comprehensive genetic vari- diseases or illnesses with major medical ant scans will allow in-time defi nitions of patient profi les that are highly correlated to be eff ective (see Box 1). Th
cholesterol as a surrogate marker, but the with a particular safety signal. Eventually drug discovery and development is still a such data will be routinely used in medi- “fl ow through a pipeline.” However, for rent use were initiated over the past sev- cal care, when the genetic data have been standardized and are available on a “bank sulted with drug-development experts (in of a new drug for Alzheimer’s disease.4 The right safe and effective drug at the
Region-specifi c brain glucose utilization right dose for the right patient at an
as a surrogate marker for the next class of appropriate cost
candidate drugs for Alzheimer’s disease calls for “collaborative efforts among may allow faster screening than the small fi les can also be used to enrich patients CLINICAL PHARMACOLOGY & THERAPEUTICS | VOLUME 81 NUMBER 6 | JUNE
PERSPECTIVES
to . . . identify diseases in great need of approval. . . . To help ensure safety, the structures would be in place to follow a Targets would still be discovered, screen- medicines ascertained on the basis of data ing assays developed, leads processed by medicinal chemists, and candidates tested for preclinical safety and drug metabolism, Conclusion
and then drugs would be tested for the fi rst To return to Dr. Lesko’s questions, large- scale observational studies in a prospec- to work, although certainly the defi nition of animal metabolism could be assisted by ing signals as they occur, and for rapidly performed starting with the FTIH studies. Organ toxicity markers might well be used routinely before FTIH, as is beginning to can provide for real-time individual and tify the adverse clinical signals to associ- ate with a safety test. Patients exhibiting those signals would be tested for genetic panels (including ethnic-specifi c variants) also be available for prospective observa- a new model (Figure 1) will allow
tional monitoring to assess predictability. With all parties in the provision of health Requirements for integrating drug
and eff ectively improve health care with development, safety surveillance,
a cradle-to-grave electronic observational and medical-outcome prospective
epidemiology
determined prospectively over years.
In its simplest form, a prospective moni- CONFLICT OF INTEREST
This article does not necessarily reflect the views Lesko, L.J. Personalized medicine: elusive dream or imminent reality? Clin. Pharmacol.
Ther. 81, 807–816 (2007).
metabolism and transport. Nat. Genet. 37,
84–89 (2005).
Roses, A.D. 2025: the practice of neurology: back from the future. Arch. Neurol. 58, 1766–
Alzheimer’s disease in persons homozygous for the epsilon 4 allele for apolipoprotein E. N. Engl. J. Med. 334, 752–758 (1996).
Roses, A.D. Genome-based pharmacogenetics and the pharmaceutical industry. Nat. Rev. Drug Discov. 1, 541–549 (2002).
VOLUME 81 NUMBER 6 | JUNE 2007 | www.nature.com/cpt
PERSPECTIVES
Kelly, P., Stallard, N., Zhou, Y., Whitehead, abacavir. Lancet 359, 1121–1122 (2002).
J. & Bowman, C. Sequential genome-wide et al. Association between presence association studies for monitoring adverse events in the clinical evaluation of new drugs. hypersensitivity to HIV-1 reverse-transcriptase having adequate education or training in Stat. Med. 25, 3081–3092 (2006).
inhibitor abacavir. Lancet 359, 727–732 (2002).
10. Phillips, E.J. Genetic screening to prevent New Drug Development: Science, Business, abacavir hypersensitivity reaction: are we there Regulatory, and Intellectual Property Issues yet? Clin. Infect. Dis. 43, 103–105 (2006).
a critical question that the authors have Cited as Hampering Drug Development Efforts, 11. Hughes, A.R. et al. CNA30027 Study Team; not addressed as to whether, and to what GAO-07-49, 35–36 (Government Accountability CNA30032 Study Team. Association of genetic extent, the reluctance of some physicians variations in HLA-B region with hypersensitivity to abacavir in some, but not all, populations. to prescribe adequate pain medications is HLA-B region and hypersensitivity reactions to Pharmacogenomics 5, 203–211 (2004).
attributable to the education and training that physicians receive in medical school and postgraduate training as opposed to See ARTICLE page 903
what the authors term fear of government action. As the DEA stated in the policy The DEA’s Balancing Act to
statement, it is not the agency’s role to provide medical training to physicians Ensure Public Health and Safety
on the general practice of medicine or to specify the precise medical circumstances and patient characteristics that warrant JT Rannazzisi1
the use of opioids to treat pain. Rather, that responsibility is carried by the medi-cal community, and how it does so seems In their article in this issue, Reidenberg and Willis assert that there
likely to be the primary factor determina- are multiple barriers to the adequate treatment of pain and that one
of these barriers is fear of government action against a physician
state boards and associations have issued who prescribes opioids for patients in pain.1 At the same time, the
authors state that the risk of a physician’s being punished by either a
ment of pain patients, while taking steps state medical board or the Drug Enforcement Administration (DEA)
for a patient in pain with adequate medical record documentation is
very small.
(CSA), the DEA is obligated to ensure the sional practice should in no way interfere availability of pharmaceutical controlled with the legitimate practice of medicine or cause any physician to be reluctant to pro- extremely small. Indeed, the DEA recently e DEA to prevent these controlled substances, published a policy statement that sought also stated in that document that it wishes to emphasize this very point and alleviate medical professionals that the agency has channels. In particular, the DEA’s role prescribing controlled substances to treat embarked on a campaign to “target” physi- under the CSA is to ensure that controlled cians who prescribe controlled substances substances are prescribed, administered, the Federal Register on September 6, 2006, for the treatment of pain (or that physi- and dispensed only for legitimate medical can be accessed on the Web sites of both cians must curb their legitimate prescrib- purposes by DEA-registered practitioners the Federal Register and the DEA’s Offi ing of pain medications to avoid liability). acting in the usual course of professional It appears that the fi ndings of Reidenberg and Willis are consistent with the DEA’s assertion that there has been no upsurge in To protect public health and safety, the law that physicians may prescribe control- the number of cases initiated by the agency DEA has always had a legal obligation to led substances only for legitimate medical against physicians who prescribe control- investigate the extremely small fraction of physicians who use their DEA regis-tration to commit criminal acts or oth- 1Office of Diversion Control, Drug Enforcement Administration, Alexandria, Virginia, USA. Correspondence: JT Rannazzisi (dea.diversion.policy@usdoj.gov) this obligation seriously, because even a single physician who uses his or her DEA CLINICAL PHARMACOLOGY & THERAPEUTICS | VOLUME 81 NUMBER 6 | JUNE

Source: http://studwww.ugent.be/~smdkeyse/elusive_dream_imminent_reality.pdf