Red Handed: An Independent Review of IRBs

A decade or so ago, the Government Accounting Office published a series of reports faulting human protections at VA hospitals, raising concerns about HHS oversight of human research, and urging continued vigilance in human research. 

After a seeming pendulum swing toward protections bashing (witness the outpouring of condemnation after OHRP sanctioned a team of researchers investigating ways to cut infection rates in intensive care units, or the Tricouncil draft policy's use of the phrase "over-protection"), a recent GAO report may show a revival of concerns about independent review operations.

In it, GAO reports the result of a "sting" operation in which a sham device study was presented to three private IRBs.  The IRBs were selected on the basis of having "less burdensome initial paperwork requirements."  Two of the IRBs rejected the protocol, describing it as "awful" and "the riskiest thing I've ever seen on this board."  But one– identified in a New York Times article as Coast Independent Review Board (whose webpage lists "speed" as the first selling point), unanimously approved the protocol, characterizing it as "probably very safe." 

Congress, GAO, and press reports are framing this as evidence of flaws in private IRB review. I'm actually impressed that the glass is 2/3 full.  And without suggesting endorsement of the private IRB model, I actually wonder how well academic medical center IRBs-- most of whose members are salaried by an institution with a significant financial and professional stake in seeing research protocols approved-- would have performed.  (photo credit: laughing squid, 2008)

Centralized Revue

In the most recent issue of Molecular Therapy, U Penn researcher Hildegrund Ertl provides a strong and eloquent defense of the Recombinant DNA Advisory Committee (RAC).  RAC was initially formed to evaluate the safety of studies involving recombinant DNA. In the last decade, however, its most visible function has been to provide advise to researchers pursuing novel gene transfer protocols in human beings.

Many researchers resent RAC, viewing it as yet another layer of oversight for their clinical studies. Understandably, they question whether it makes sense to have a separate review track for gene transfer. Other scientists and ethicists might question whether gene transfer is so exceptional as to be singled out for separate review, but would nevertheless argue that the RAC model should be extended to other ethically contentious areas of medical research. Nevertheless, the RAC model of centralized review of trial protocols has yet to be extended to comparably novel and contentious human clinical research areas like cell transfer, embryonic stem cell research, or tissue engineering.

Ertl provides a clear and persuasive description of RAC's role in improving gene transfer trial safety, enhancing scientific value of studies, and ensuring appropriate informed consent practices. But the structure of her argument embeds three assumptions that, in my view, need to be questioned.

1- Why demand solid preclinical evidence? Ertl answers "if such data are not available, the risk outweighs the potential benefit for human volunteers– and that is not acceptable." I would argue that preclinical evidence is of greater use in improving the scientific value of clinical studies. 

2- How are risks justified in early phase studies? In the above quote, Ertl seems to suggest the answer is therapeutic benefit for the volunteer. In my view, risks in first-in-human trials are justified by the potential for scientific gain, not direct medical benefit.

3- What is the purview of ethics? Ertl, like many others, partitions "technical" concerns like study validity / value / preclinical evidence from "ethical" concerns like informed consent and conflict of interest. But why is the former any less ethical than the latter, given that technical questions implicate problems of risk-benefit balance and the ultimate ends of research. In my view, there is no clear division between the technical and ethical, and few if any decisions in designing and executing clinical protocols are devoid of ethical content. (photo credit: 416style, 2005)

Departing Milano Stazione? ADA-SCID and Gene Transfer

Greetings after a hiatus for teaching, grants, committees, book deadlines, wiping runny noses, and more. Much has happened since my last posting, and in the next two or three weeks, I hope to catch up.

First item on the agenda is a Jan 29 report in New England Journal of Medicine (NEJM) describing successful reconstitution of immune function in eight of ten children receiving gene transfer for adenosine deaminase severe combined immune deficiency (ADA-SCID). The paper follows on a previous report in Science, 2002, and almost certainly counts as gene transfer's greatest clinical accomplishment to date.

I have previously argued in Lancet and Developing World Bioethics, as well as in my forthcoming book, that this study raised important justice concerns because it recruited volunteers from economically disadvantaged settings without clearly fulfilling the requirement, articulated in the Declaration of Helinki, of responsiveness. The NEJM article does not say where subsequent volunteers were recruited, though the fact that all but one new volunteer received PEG-ADA (a very expensive standard of care available only in high-income countries) suggests that later patients were not economically disadvantaged.

Rather than dwell on justice, I'd like to focus on the significance of this study. As indicated, eight of ten children with a life threatening immune disorder had their immune systems reconstituted. Five of these children had T-cell counts that were "above the lower limits of normal." These children were able to enjoy normal social relations parents and other children.

There do not appear to have been any adverse events relating to the gene transfer vector. A major concern was the possibility that gene transfer might trigger a leukemia-like syndrome observed in two X-SCID studies. Blood tests of children in this ADA-SCID study, however, do not evidence of either the leukemia syndrome or its precursors– at least within the time frame of the study (median follow-up of 4 years; range: 1.8-8 years).

So is ADA-SCID gene transfer ready to leave Milan and conquer ADA-SCID?  For children lacking haplo-identical bone marrow donors, maybe so given the morbidity associated with marrow  transplantation. Still, there are lingering concerns. First, though these results are encouraging, risks of malignancy remain unquantified. Second, this gene transfer regime requires several ancillary treatments- like bone marrow conditioning- that expose patients to risk of infection until the gene transfer intervention kicks in. Several volunteers in this study developed infections and neutropenia, for example. In an accompanying editorial in NEJM, Donald Kohn and Fabio Candotti describe several ways that retroviral gene transfer to blood stem cells might be made safer. Last, it is important to remember that ADA-SCID is a multi-system disorder, with neurological, skeletal, and other effects. Though this approach seems to address what is by far the largest cause of morbidity and mortality in children with ADA-SCID, it does eradicate their condition.

The results of Aiuti et al have been widely celebrated in the gene transfer community.  Kohn and Candotti's editorial, for example, is titled "Gene Therapy Fulfilling its Promise."  More than any single gene transfer study I can think of, this one seems to have earned the vindicating headlines. (photo credit: Paolo Margari, Milano Sazione Centrale Ferrovi, 2008)

Found Figures: Picking up the Pieces after an HIV Vaccine Trial Fails

In the November 29, 2008 issue of Lancet, two reports (plus a commentary) report the famously disappointing outcome of a recent placebo-controlled study testing adenoviral vector-based vaccines against HIV. News reports over a year ago reported that the study was halted after an interim analysis failed to show any prospect of proving effective. More troubling, subgroup analysis suggested that vaccine recipients who had high pre-existing immunity to the adenoviral vectors showed higher rates of sero-conversion compared with placebo. As this vaccine was among the most promising and advanced in terms of development, these results were seen as a major setback.

The recent Lancet reports paint a complicated picture: if I read them correctly, the inference that vector might enhance sero-conversion is muddied by the finding that circumcision status might also have played a role in sero-conversion (men with higher rates of adenoviral immunity were also, coincidentally, less likely to be circumcised).

What is clear, from what I gather, is that this is a good example where rigorous preclinical testing, coupled with rigorous trial design, permits meaningful interpretation of (unfortunately) negative human trial results. As Merlin Robb notes in a commentary accompanying the Lancet reports "the predictive value of the non-human SHIV-challenge model is not supported by this experience. The benchmarks for advancing candidate vaccines to efficacy testing and the priorities for vaccine research have been re-examined."

Well-designed studies, supported by rigorous preclinical testing, should always produce valuable, findings– like the unexpected "found figures" in the bark of a tree  (photo credit: Readwalker, Found figures, 2006)

Prime Time for Embryonic Stem Cells?

According to a recent report in the Washington Post, researchers at Geron have received approval from FDA to initiate the first ever human trial involving stem cells derived from human embryos.  A story in the most recent issue of Nature provides more background.

Briefly, the study will involve transplanting tissues derived from human embryonic stem cells into patients who have recently suffered severe spinal cord injury. The principle behind the study is that the embryo-derived oligodendrocytes might repair myelin and restore the ability for nerves to transmit impulses.  According to the Nature report, Geron, submitted 22,000 pages of material to FDA, including data from 24 studies involving over 2000 animals.

So is the decision to initiate studies at this juncture prudent?  That's impossible to know without seeing the supporting data. What I can comment on, however, is the recurrence of a rhetoric that glosses trial initiation– rather than trial outcome– as a medical achievement in itself.  Whereas the former is a regulatory event, the latter is a clinical event. In my book about gene transfer, I argue that this sets up a cycle of expectation that is difficult to sustain given the scientific and clinical uncertainties. We saw this in the early days of gene transfer. Some examples of this "trial initiation"="medical achievement":

• "This... marks the dawn of a new era in medical therapeutics. This approach is one that reaches beyond pills and scalpels to achieve a new level of healing." (Thomas Okarma, Geron chief executive)

• "Today's news... is a milestone in the new era of hope..." (Amy Comstock Rick, Coalition for the Advancement of Medical Research)

• "This is what we've all been waiting for" (Robert Lanza, Advanced Cell Technology)

• "The announcement boosted the price of shares in [Geron]... up 56% from the day before the announcement" (Meridith Wadman, Nature, Jan 27, 2009)

I wish Geron, and the patients enrolled in this study, all the best.  But if embryonic stem cell work is anything like practically every other major medical advancement, be prepared for a very long, tough slog with lots of setbacks.  In one of the stories, Sean Tipton from the Coalition for the Advancement of Medical Research, commented  "This is a trial of one particular application, not a trial of all embryonic stem cells." That sounds just about right. (photo credit: no typographic man, UlamSpiral (negative), 2006)

Age of Risk: Biologicals

Approving new drugs is a risky business. Despite best efforts (and frankly, some less than best efforts), newly approved drugs frequently turn out to have unexpected toxicities. One example is unexpected heart toxicity associated with the use of the common pain-killers like rofecoxib (i.e. Vioxx).  Another is the surprising heart toxicity associated with the wonder drug for AML (a type of leukemia), imatinib mesylate (i.e. Gleevec).

According to a  2002 paper in JAMA, 8% of new drugs approved by FDA receive "black box" labels warning of toxicities that were not originally detected in drug trials. Another 3% are withdrawn from the market because of safety concerns. 

But what about biologics- vaccines, monoclonal antibodies, recombinant protein products, cell derived agents, etc.? There are a number of reasons why one might anticipate even higher rates of "unexpected" toxicities with this class of therapeutics. For one, they frequently cause immune reactions that are exceedingly difficult to anticipate in animal studies. For another, small alterations in production can dramatically change the composition and properties of a biologic product. For still another, biologics often have a very high degree of species specificity, limiting the predictive value of animal studies.

According to a recent report in JAMA led by Thijs Giezen (October 22/29, 2008), 24% of biologics approved for marketing in Europe received "black box" warnings.  For first-in-class agents, five of eight compounds were subject to regulatory action following approval.  A story in the January 2009 issue of Nature Biotechnology (Jim Kling) provides some perspective on these findings: most biologics are used to treat life threatening illnesses, which may make people more susceptible to toxic reactions (on the other hand, toxicity might be difficult to detect amidst the noise of disease course).

Bottom line: as translational researchers pursue biologics, uncertainty will continue to present a major challenges, necessitating new approaches to pharmacovigilence and trial design. (photo credit: teotwawki 2005)

What Will the New Year Bring?

My hope is that, with a new Presidential administration, certain research and ethics policies of the U.S. government might be revisited. One such policy is the FDA's replacement of the Declaration of Helsinki with the International Conference on Harmonization's Guideline for Good Clinical Practice (GCP) for foreign studies.  Recall that, in May, I lamented what at that time was a proposed policy change (it was finalized on October 27, 2008).

In the January 3 issue of Lancet, Charles Weijer, Eric Meslin, and I briefly lay out reasons why FDA's action is regrettable, and how GCP and Helsinki differ.  We urge the incoming administration to review this policy, and for medical societies to take action now and endorse Helsinki.  Our comment is featured, and is thus accessible for free at Lancet. (photo credit: Vilhelm Sjostrom, "Happy New Year 2009 from Helsinki, Finland," last seconds of 2008)