Published online June 2, 2008
PEDIATRICS Vol. 121 No. 6 June 2008, pp. 1287-1288 (doi:10.1542/peds.2007-3450)

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 Citation Map Services E-mail this article to a friend Similar articles in this journal 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 Truog, W. E. Search for Related Content PubMed PubMed Citation Articles by Truog, W. E. Related Collections Premature & Newborn Social Bookmarking                         What's this?

LETTER TO THE EDITOR

Inhaled Nitric Oxide for Preterm Infants

William E. Truog, MD
Department of Pediatrics
University of Missouri
Kansas City, MO 64108

Richard J. Martin, MD
Department of Pediatrics
Case Western Reserve University
Cleveland, OH 44106

Roberta A. Ballard, MD
Department of Pediatrics

Dennis Black, PhD
Department of Biostatistics
University of California
San Francisco, CA 94143

Avital Cnaan, PhD
Department of Biostatistics
University of Pennsylvania
Philadelphia, PA 19104

For the NO CLD Study Group To the Editor.—

In their recent review article in Pediatrics, "Inhaled Nitric Oxide for Preterm Infants: A Systematic Review,"¹ Barrington and Finer report on meta-analyses of nitric-oxide (NO) studies in premature infants. They include our NO CLD study² as 1 of 2 studies testing the effect of inhaled NO (iNO) started beyond the first 3 days of life with the primary outcome of improving survival without bronchopulmonary dysplasia (BPD) measured at 36 weeks' postmenstrual age. In their article, Barrington and Finer stated that "[t]he study by Ballard et al reported a significant benefit of iNO in improving rates of survival without BPD. However, the figures given in the article (165 of 294 vs 182 of 288 patients) were not significant when RevMan software was used to calculate the [relative risk] and its CIs [confidence intervals] ... or when SPSS software was used to perform a simple ² test ... or unadjusted, logistic regression analysis. The reason for this discrepancy is not clear." On the basis of this analysis, they stated in their abstract that "[l]ater use of inhaled nitric oxide to prevent bronchopulmonary dysplasia does not seem to be effective."

We welcome the opportunity to resolve the apparent discrepancy and comment on the statistical analyses performed by Barrington and Finer. As stated in our original article² and a subsequent correction, we identified the special issue raised by second-born and third-born siblings who also qualified to be in our study. These infants were enrolled but not individually randomly assigned; rather, because of anticipated parental preference, other siblings received the same study gas (in a blinded fashion) as the first enrolled infant. We recognized that outcomes for siblings are also correlated because of shared in utero environmental and genetic factors and, therefore, planned a priori to use a statistical analysis that would take into account the clustering effect that occurs by including siblings. This need was explicitly stated in the methods section of our original article.² The dependence of units within a cluster and the need for special statistical techniques to account for this has long been recognized in the statistical literature.^3,4 A number of techniques have been proposed to deal with cluster randomization. The statistical technique that we used originally, multiple outputations,⁵ was suggested by biostatisticians both at the National Heart, Lung, and Blood Institute, the funding source for the study, and at the University of Pennsylvania as being appropriate.

After our data set was moved from Children's Hospital of Philadelphia to the University of California at San Francisco, the data were corrected for incorrect categorization of outcome in 1 infant and for 9 infants who were originally unrecognized as being part of a multiple birth. The data were then reanalyzed by using both the multiple outputations technique and the generalized estimating equation method, the most commonly used technique to account for clustering.³ Both analyses yielded a P value of .03 with a relative risk of 1.26 (95% CI: 1.02–1.55) and an odds ratio of 1.45 (95% CI: 1.03–2.04) for survival without BPD at 36 weeks, favoring NO-treated infants.² The statistical approaches used by Barrington and Finer in their reanalysis of our data do not account for clustering with nonindependence of multiples and, therefore, are inappropriate, explaining the apparent discrepancy noted by the authors.

Accounting for clustering, iNO therapy in our trial provided statistically significant benefit for outcome at 36 weeks' postmenstrual age. Favorable outcome was also observed at 40 (P = .008) and 44 (P = .03) weeks' postmenstrual age, and treated infants had significantly fewer hospital and ventilator days during the primary hospitalization (predetermined secondary outcome variables). To date, no adverse effects of iNO therapy in our trial have been identified.^2,6–8 These findings demonstrate that iNO as used in our study safely reduces the incidence and severity of BPD.

We also noted errors in Figs 2 and 3 related to "studies with entry after 3 days based on BPD risk." In Fig 2, labeled "Plot of the effects of INO on the combined outcome of death or BPD," the data presented are for death only. In Fig 3, labeled "Effect of INO on severe intracranial hemorrhage," the outcome of death or BPD for the 2 trials is presented.

We thank Drs Barrington and Finer for providing a comprehensive and transparent review of iNO therapy, not only with our trial but with the multiple other studies that have been published. We fully appreciate the difficulty of interpreting clinical studies with the multiple entry criteria, treatment protocol designs, indications for altering the dose of NO, lack of blinding, and other variables. In addition, we feel strongly that the issue of multiples enrolled in perinatal and neonatal clinical studies needs to be addressed a priori in design and analysis. On the basis of the information provided in this letter, we look forward to a revision of the review article that recognizes the effectiveness of iNO as used in the NO CLD trial to improve survival without BPD in preterm infants.

FOOTNOTES

Statements appearing here are those of the writers and do not represent the official position of the American Academy of Pediatrics or its Committees. Comments on any topic, including the contents of PEDIATRICS, are invited from all members of the profession; those accepted for publication will not be subject to major editorial revision but generally must be no more than 400 words in length. The editors reserve the right to publish replies and may solicit responses from authors and others.

Please see www.pediatrics.org for instructions on submitting letters.

REFERENCES

1. Barrington KJ, Finer NN. Inhaled nitric oxide for preterm infants: a systematic review [published correction appears in Pediatrics. 2008;121(2):451]. Pediatrics. 2007;120 (5):1088 –1099[Abstract/Free Full Text]
2. Ballard RA, Truog WE, Cnaan A, et al. Inhaled nitric oxide in preterm infants undergoing mechanical ventilation [published correction appears in N Engl J Med. 2007;357(14):1444–1445]. N Engl J Med. 2006;355 (4):343 –353[Abstract/Free Full Text]
3. Liang KY, Zeger SL. Longitudinal data analysis using generalized linear models. Biometrika. 1986;73 (1):13 –22[Abstract/Free Full Text]
4. Donner A. Some aspects of the design and analysis of cluster randomization trials. J R Stat Soc Ser C Appl Stat. 1998;1 :95 –113
5. Follmann D, Proschan M, Leifer E. Multiple outputation: inference for complex clustered data by averaging analyses from independent data. Biometrics. 2003;59 (2):420 –429[CrossRef][Web of Science][Medline]
6. Truog WE, Ballard PL, Norberg M, et al. Inflammatory markers and mediators in tracheal fluid of premature infants treated with inhaled nitric oxide. Pediatrics. 2007;119 (4):670 –678[Abstract/Free Full Text]
7. Ballard PL, Merrill JD, Truog WE, et al. Surfactant function and composition in premature infants treated with inhaled nitric oxide. Pediatrics. 2007;120 (2):346 –353[Abstract/Free Full Text]
8. Ballard PL, Truog WE, Merrill JD, et al. Plasma biomarkers of oxidative stress: relationship to lung disease and inhaled nitric oxide therapy in premature infants. Pediatrics. 2008;121 (3):555 –561[Abstract/Free Full Text]

---

PEDIATRICS (ISSN 1098-4275). ©2008 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 Citation Map Services E-mail this article to a friend Similar articles in this journal 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 Truog, W. E. Search for Related Content PubMed PubMed Citation Articles by Truog, W. E. Related Collections Premature & Newborn Social Bookmarking                         What's this?