PEDIATRICS Vol. 114 No. 6 December 2004, pp. 1670-1671 (doi:10.1542/10.1542/peds.2004-2178)
COMMENTARY |
Early Closure of the Watterberg Trial
Emeritus Professor
Henderson Research Centre
Clinical Trials Methodology Group
McMaster University
Hamilton, Ontario, Canada L8V 1C3
The early closure of the Watterberg trial1 caused by excess gastrointestinal (GI) perforations represents the second large National Institute of Child Health and Human Development–funded steroid trial to suffer this fate. Despite substantial literature, the advisability of early steroids to prevent bronchopulmonary dysplasia (BPD) in preterm infants remains in doubt,2 especially with respect to the longer-term outcome. I can’t help but believe that we have missed an opportunity to settle this issue by not allowing these trials to be completed.
A Data Safety Monitoring Board (DSMB) recommendation to stop a trial early for safety reasons ultimately boils down to clinical and ethical judgements. At the heart of the clinical judgement is the trade-off between safety and efficacy: the balance of clinical importance/excess risk on each side of the treatment equation. At the heart of the ethical judgement is the balance between potential risk to trial participants from continuing the study versus potential risk to the population at large from not continuing the study.
Statistical issues are generally more straightforward, although one has not been adequately stressed in my view. Those who craft protocols are careful to avoid multiplicity by specifying a single "primary" efficacy outcome but often mention many potential safety concerns. The resulting increased risk of chance differences requires a more extreme P value to label any particular excess risk as "significant." In addition, adverse effects tend to appear quite early, whereas evidence of efficacy often takes longer to become established.
A DSMB’s main concern is naturally to protect the interests of trial participants. However, when the experimental treatment is generally available, patients being treated outside of the trial also warrant consideration. It may thus sometimes be deemed necessary to allow a trial to go on longer to provide more convincing evidence for the community at large.
The net result is that a DSMB must be prepared to "hang in" when facing an emerging safety issue until (1) they are reasonably sure it is real and (2) they have given efficacy an opportunity to show. At the same time, they should bear in mind external data and the broader view of the population at risk.3
Although the justification to stop the Watterberg trial might seem straightforward (a small observed BPD reduction versus somewhat clearer evidence of excess spontaneous GI perforations), I’m sure it represented an excruciatingly difficult decision for the DSMB. In the hard light of day, the Watterberg data (1.5% reduction in death/BPD, 6% excess risk of any GI perforation) seem quite thin as the basis of stopping the trial, especially when one considers that the new data are statistically consistent with the Cochrane pooled estimates (7% reduction in death/BPD, 3% increase in perforations). Given that the Cochrane data predated the start of the Watterberg trial, the new results should not have been seen as unexpected despite the lower steroid dose.
The Watterberg and Stark et al4 trials had the potential to double the available data on this topic. Although the decisions to foreclose these trials were understandable in some respects, societal good may have been better served by allowing each study to continue to fruition. Arch skeptics might even question the ethics of starting such trials if, as a community, we are not prepared to let them continue to the point of providing a real contribution to scientific knowledge.
 |
FOOTNOTES |
|---|
Accepted Oct 4, 2004.
Address correspondence to Robin S. Roberts, MTech, Henderson Research Centre, Clinical Trials Methodology Group, 711 Concession St, Hamilton, Ontario, Canada L8V 1C3. E-mail: robertsr{at}mcmaster.ca
No conflict of interest declared.
|
REFERENCES |
|---|
 TOP REFERENCES |
|---|
- Watterberg KL, Gerdes JS, Cole CH, et al. Prophylaxis of early adrenal insufficiency to prevent broncopulmonary displasia: a multicenter trial.
Pediatrics. 2004;114
:1649
–1657
[Abstract/Free Full Text] - Halliday HL, Ehrenkranz RA, Doyle LW. Early postnatal (<96 hours) corticosteroids for preventing chronic lung disease in preterm infants. Cochrane Database Syst Rev. 2003;(1) :CD001146
- Slutsky AS, Lavery JV. Data safety and monitoring boards.
N Engl J Med. 2004;350; 1143
–1147
[Free Full Text] - Stark AR, Carlo WA, Tyson JE, et al. Adverse effects of early dexamethasone in extremely-low-birth-weight infants. National Institute of Child Health and Human Development Neonatal Research Network.
N Engl J Med. 2001;344
:95
–101
[Abstract/Free Full Text]
PEDIATRICS (ISSN 1098-4275). ©2004 by the American Academy of Pediatrics
CiteULike 
Connotea 
Del.icio.us 
Digg 
Facebook 
Reddit 
Technorati 
Twitter What's this?
This article has been cited by other articles:
|
|
 |
|
  L. W. Doyle, P. G. Davis, C. J. Morley, A. McPhee, J. B. Carlin, and and the DART Study Investigators Low-Dose Dexamethasone Facilitates Extubation Among Chronically Ventilator-Dependent Infants: A Multicenter, International, Randomized, Controlled Trial Pediatrics, January 1, 2006; 117(1): 75 - 83. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. V. Gordon Weighing Statistical Certainty Against Ethical, Clinical, and Biologic Expediency: The Contributions of the Watterberg Trial Tip the Scales in the Right Direction Pediatrics, May 1, 2005; 115(5): 1446 - 1447. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
