Risk/Benefit in Pediatric Phase 1 Cancer Trials: Noble Lie? (part 1)

Photo from art.crazed Elizabeth on Flickr, March 16, 2010.

In years of studying the ethics of early phase trials in patients- for example, cancer phase 1 trials- I’ve become more and more convinced that it is a mistake to think of these trials as having a therapeutic impetus.

To be sure- the issues are complex, many people who share my view do so for the wrong reasons. But in general, it seems to me difficult to reconcile the concept of competent medical care with giving patients a drug that will almost certainly cause major toxicities- and for which there is at best highly fallible animal evidence to support its activity (and at worst- no animal evidence at all).

For this reason, I think groups like ASCO and others – who (in a manner that is self-serving) advocate phase 1 trials as a vehicle for care when patients qualify- do a major disservice to patients and the integrity of medicine.

But surely there are cases where the risks of phase 1 trial enrollment might plausibly be viewed as outweighed by the prospect of direct benefit. As I’ve argued elsewhere, the institution of phase 1 testing is comprised of a heterogeneous set of materials and activities. With a drug, you can specify its composition and dose on a product label, and declare the drug “therapeutic” or “nontherapeutic” for a specific patient population. With phase 1 trials, there is no standard composition or dose- phase 1 trials cannot be put in a bottle and labeled as a homogeneous entity that has or does not have therapeutic value. If this is the case, it seems plausible that there are some phase 1 trials that come closer- and perhaps exceed- the threshold of risk/benefit/uncertainty that establish a therapeutic claim. That is, it seems conceivable that phase 1 trials may be done under conditions, or with sufficiently strong supporting evidence, that one can present them as a therapeutic option for certain patients without lying or betraying the integrity of medicine.

U.S. regulations (and those elsewhere) state that- when exposing children to research risks exceeding “minor increase over minimal,” research risks must be “justified by the anticipated benefit to the subjects…the relation of… anticipated benefit to the risk [must be] at least as favorable to the subjects as that presented by available alternative approaches.” Given that risks of drugs tested in phase 1 cancer trials exceed minor increase over minimal, U.S. regulations require that we view phase 1 trial participation as therapeutic when we enrol children.

Can this regulatory standard be reconciled with my view? I used to think so. Here’s why. Pediatric phase 1 trials are typically pursued only after drugs have been tested in adults. Accordingly, the ‘dogs’ of drug development have been thinned from the pack before testing in children. These trials also test a narrower dose range- and as such, a greater proportion of participants are likely to receive active doses of drug. Finally- rightly or wrongly- the ethos of protection that surrounds pediatric research, plus the stringency of regulations surrounding pediatric testing would- one might think- tend towards demanding higher evidentiary standards for launch of testing.

This week, Marcin Waligora, colleagues, and I published the largest meta-analysis of pediatric phase 1 cancer trials. that fills me with doubt about a therapeutic justification for phase 1 pediatric trials (for news coverage, see here). Before describing our findings, a few notes of caution.

First, our findings need to be interpreted with caution- crappy reporting practices for phase 1 trials make it hard to probe risk and benefit. Also, our analyses used methods and assumptions that are somewhat different than those used in similar meta-analyses of adults. Finally, who am I to impose my own risk/benefit sensibility on guardians (and children) who have reached the end of the line in terms of standard care options?

These provisos aside, our findings suggest that the risk/benefit for pediatric phase 1 cancer trials is not any better than it is for adult trials. Some salient findings:

• on average, every pediatric participant will experience at least one severe or life threatening side effect.
• for monotherapy trials in children with solid tumors (where we can compare our data with previous studies of adults), about 2.5% of children had major tumor shrinkage. This compares with a decade-old estimate of 3.8% in adults. 10.5% for combination therapy vs. 11.7% in adults.
• Contrary to all the latest excitement about new treatment options, our data do not show clear time trends suggesting an improvement of risk/benefit with newer drugs.
• 39% of children in phase 1 studies received less than the recommended dose of the investigational drug.

If- in fact- we reject the view that adult phase 1 studies can generally be viewed as therapeutic, and if, in fact, risk/benefit in pediatric studies has a similar risk/benefit balance despite their building on adult evidence, and if we accept that available care options outside a trial are no better or no worse in terms of their risk/benefit for children and adults- then it follows (more or less- and assuming our meta-analysis presents an accurate view of risk/benefit) that phase 1 trials in children cannot generally be presented as having a therapeutic risk/benefit.

This puts medicine a bind. Phase 1 trials are critical for advancing treatment options for children. But most cannot- in my view- be plausibly reconciled with research regulations. Either a) my above analysis is wrong, b) a lot of substandard phase 1 trials in children are pulling the average estimate of benefit down, making it hard to discern the truly therapeutic ones, c) we must accept- a la Plato- a noble lie and live a fiction that phase 1 studies are therapeutic, d) we must cease phase 1 cancer drug trials in children, or e) regulations are either misguided, or f) phase 1 trials should undergo a specialized review process- so called 407 review.

I would posit (b) and (e) and (f) as the most plausible implications of our meta-analysis.

In my next post, a few reflections. And stay tuned for further empirical and conceptual work on this subject.

2018 February 26 | Leave a comment

@Manual{stream2018-1559,
    title = {Risk/Benefit in Pediatric Phase 1 Cancer Trials: Noble Lie? (part 1)},
    journal = {STREAM research},
    author = {Jonathan Kimmelman},
    address = {Montreal, Canada},
    date = 2018,
    month = feb,
    day = 26,
    url = {http://www.translationalethics.com/2018/02/26/risk-benefit-in-pediatric-phase-1-cancer-trials-noble-lie-part-1/}
}