Published online June 1, 2006
PEDIATRICS Vol. 117 No. 6 June 2006, pp. 2332-2333 (doi:10.1542/peds.2006-1046)

This Article Extract Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services E-mail this article to a friend Similar articles in this journal Similar articles in Web of Science Similar articles in PubMed Alert me to new issues of the journal Add to My File Cabinet Download to citation manager Request Permissions Citing Articles Citing Articles via CrossRef Citing Articles via Google Scholar Google Scholar Articles by Samuels, M. P. Articles by Southall, D. P. Search for Related Content PubMed PubMed Citation Articles by Samuels, M. P. Articles by Southall, D. P. Related Collections Premature & Newborn Social Bookmarking                         What's this?

Continuous Negative Extrathoracic Pressure (CNEP) in Neonatal Respiratory Failure: Correction

Martin P. Samuels, MD
David P. Southall, MD
University Hospital of North Staffordshire, Stoke-on-Trent,
Staffordshire ST4 6QG, United Kingdom

To the Editor.—

There have been many allegations in the British press in the last 6 years that the conduct of the randomized trial of noninvasive respiratory support for preterm infants with respiratory distress that we published in this journal in December 1996^1,2 was deeply flawed. One independent rebuttal of most of those claims was published in 2000,³ but because the regulatory bodies are still investigating these allegations,⁴ it becomes important to state clearly that the results reported in the original article are correct and require no amendment.

Despite stratified randomization based on gestation and the need for oxygen and other respiratory support at 4 hours, the mean birth weight of infants randomly assigned to receive conventional care turned out to be lower than those offered continuous negative extrathoracic pressure (CNEP) using a Perspex chamber from the neck down. In most trials, random allocation is relied on to provide balanced study groups for comparison, and no retrospective attempt is made to adjust for any residual imbalances uncovered after the trial ends, and that is what was done in this trial.

However, 25 infants weighing <1 kg ended up receiving CNEP, and 40 received standard treatment; 28 of the 122 infants treated with CNEP died, and 22 of the 122 were offered standard care. These differences could have arisen by chance, and we did not think it was appropriate to undertake additional statistical manipulation of the data. Indeed, even if allowance is made for the residual imbalance in the matching variables, gestation, and oxygen need, using conditional logistic regression, the difference in mortality still remains readily explained by the play of chance (odds ratio: 2.56; 95% confidence interval: 0.87–7.51). Adjusting for the imbalance in birth weight also leaves the difference in mortality nonsignificant (odds ratio: 1.99; 95% confidence interval: 0.81–4.89). Adjusting for all the imbalances simultaneously does reveal a treatment difference that just reaches significance at the 5% level, but this inference should be treated with considerable caution because it is based on just 26 of the 122 randomized pairs (those pairs in which 1 infant died and 1 survived). What is more, the decision to perform this analysis was itself suggested by the data, and it is well known than this generally leads to statistical significance being overestimated. Thus, evidence of a difference in mortality is weak, and we now know that the slight imbalance in disability in the 10-year-old survivors is in the reverse direction.⁵

As the initiating researchers for this study, we thank Professor Doug Altman in Oxford for agreeing to conduct this additional analysis jointly with the trial's original statistician. The original (anonymized) data file is now in the public domain,^6,7 which allows others to scrutinize these findings for themselves. Publication of this note allows us to correct the 1 error that did creep into the original article: a statement in the discussion indicated that the unexpected imbalance in birth weight (correctly reported in the results section) would tend to diminish the potential implication of the statistically nonsignificant difference in mortality between the 2 treatment groups, when the converse is obviously the case.

Neonatal care has moved on in the 18 years since this trial was first planned. Even surfactant was not generally available in the United Kingdom when the first patients were recruited. The concept that action to sustain alveolar distension by negative extrathoracic pressure might limit the number of infants needing endotracheal intubation and consequential lung trauma was a novelty when the CNEP trial was first launched. It is now increasingly realized that the concept was right but that constant positive airway pressure through the nose (nasal continuous positive airway pressure) addresses the same need in a simpler way in preterm infants. However, there still remains a real possibility that CNEP has a useful role to play in managing older infants with nasal congestion and a viral respiratory tract infection,⁸ and we remain deeply dismayed that the flawed criticism of our earlier article has put an effective embargo on the additional exploration of this possibility in the United Kingdom for more than 6 years.

REFERENCES

1. Samuels MP, Raine J, Wright T, et al. Continuous negative extrathoracic pressure in neonatal respiratory failure. Pediatrics. 1996;98 :1154 –1160[Abstract/Free Full Text]
2. International Register of Controlled Trials. Continuous negative extrathoracic pressure (CNEP) in neonatal respiratory failure: ISRCTN05982584. Available at: www.controlled-trials.com/isrctn/trial//0/05982584.html. Accessed April 18, 2006
3. Hey E, Chalmers I. Investigating allegations of research misconduct: the vital need for due process. BMJ. 2000;321 :752 –755[Free Full Text]
4. Dyer C. Appeal court rules that GMC must reconsider complaints against Southall. BMJ. 2005;331 :1426 . Available at: http://bmj.bmjjournals.com/cgi/content/full/331/7530/1426-b?maxtoshow=. Accessed April 18, 2006
5. Telford K, Waters L, Vyas H, Manktelow BN, Draper ES, Marlow N. Outcome following neonatal continuous negative pressure ventilation. Lancet. 2006;367 :1080 –1085[CrossRef][Web of Science][Medline]
6. Ho JJ, Subramaniam P, Henderson-Smart DJ, Davis PG. Continuous distending pressure for respiratory distress syndrome in preterm infants. Cochrane Database Syst Rev. 2002;(2) :CD002271
7. Samuels MP, Southall DP. The neonatal Continuous Negative Extrathoracic Pressure (CNEP) trial in 1996: a ten year retrospective. Trials. 2006;7 : In press
8. Al-balkhi A, Klonin H, Marinaki K, et al. Review of treatment of bronchiolitis related apnoea in two centres. Arch Dis Child. 2005;90 :288 –291[Abstract/Free Full Text]

---

PEDIATRICS (ISSN 1098-4275). ©2006 by the American Academy of Pediatrics

CiteULike    Connotea    Del.icio.us    Digg    Facebook    Reddit    Technorati    Twitter    What's this?

This Article Extract Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services E-mail this article to a friend Similar articles in this journal Similar articles in Web of Science Similar articles in PubMed Alert me to new issues of the journal Add to My File Cabinet Download to citation manager Request Permissions Citing Articles Citing Articles via CrossRef Citing Articles via Google Scholar Google Scholar Articles by Samuels, M. P. Articles by Southall, D. P. Search for Related Content PubMed PubMed Citation Articles by Samuels, M. P. Articles by Southall, D. P. Related Collections Premature & Newborn Social Bookmarking                         What's this?