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 Meadow, W. Articles by Wright, L. Search for Related Content PubMed PubMed Citation Articles by Meadow, W. Articles by Wright, L. Social Bookmarking                         What's this?

PEDIATRICS Vol. 113 No. 3 March 2004, pp. 627-628

Importance of Bilirubin Effect Questioned

William Meadow, MD, PhD
Department of Pediatrics
University of Chicago
Chicago, IL 60637

To the Editor.—

I read with interest the manuscript of Oh et al¹ from the National Institute of Child Health and Human Development Neonatal Research Network describing the relationship of peak serum bilirubin (PSB) and outcomes in extremely low birth weight infants. I have 2 residual questions that I’m sure the authors would be able to address: one relatively small and methodological and the other somewhat larger in scope.

To their great credit, the authors point out that the effect sizes are very small, ie, the largest odds ratio (OR) for the effect of PSB on outcomes is only 1.138 (an 14% effect), and the largest OR for an outcome for which the 95% confidence interval does not touch 1.0 is 1.068 (a 7% effect). My first question is how robust would those findings be if obvious outliers were removed. In other words, we all have a sense that infants with birth weight <1000 g whose PSB is, say, 14 mg/dL, are likely to be very ill and at great risk for problems. So my first question is, by removing kids with PSB above various extreme thresholds, how strong would the ORs and regressions be?

The second question has a broader theme. The authors present 6 outcome variables, 1 of which includes both nonsurvivors and survivors (death or neurodevelopmental impairment) and the others of which pertain only to survivors. The only outcome for which the OR doesn’t touch 1.0 is death or neurodevelopmental impairment. All the outcomes that speak only to survivors are marginally significant at best. Moreover, as the authors describe, many of the other potential risk factors are correlated much more powerfully with abnormal outcomes (eg, maternal age, maternal schooling, Medicaid status, as well as more-obvious ones such as intraventricular hemorrhage and periventricular leukomalacia). Yet in both the discussion of the article and the conclusions of the abstract, the authors emphasize the implications of their findings for studying the impact of controlling PSB on potential outcomes for extremely low birth weight survivors. In light of the data presented, I find that emphasis perhaps misplaced.

I would appreciate the authors’ view of this possible interpretation.

REFERENCE

1. Oh W, Tyson JE, Fanaroff AA, et al. Association between peak serum bilirubin and neurodevelopmental outcomes in extremely low birth weight infants. Pediatrics.2003; 112 :773 –779[Abstract/Free Full Text]

---

 
William Oh, MD^a
Betty R. Vohr, MD^a
Jon E. Tyson, MD^b
Avroy A. Fanaroff, MD^c
Barbara J. Stoll, MD^d
Richard A. Ehrenkranz, MD^e
Waldemar Carlo, MD^f
Seetha Shankaran, MD^g
Kenneth Poole, PhD^h
Rebecca Perritt, MS^h
Linda Wright, MD^i
a Brown University
Providence, RI 02912
^b University of Texas
Houston Medical School
Houston, TX 77030
^c Case Western Reserve University
Cleveland, OH 44106
^d Emory University
Atlanta, GA 30322
^e Yale University
New Haven, CT 0652
^f University of Alabama
Tuscaloosa, AL 35487
^g Wayne State University
Detroit, MI 48202
^h Research Triangle Institute
Research Triangle Park, NC 27709
ⁱ National Institutes of Child Health and Development
Rockville, MD 20847

In Reply.—

We appreciate Meadow’s comments on our article. He raised 2 questions: 1) How robust would the findings on the association between peak serum bilirubin (PSB) and neurodevelopmental outcome be if the outliers were excluded from the analysis? 2) Is the emphasis misplaced in regard to the effect of PSB on outcomes versus other major morbidities such as intraventricular hemorrhage and periventricular leukomalacia?

With reference to the first question, we reanalyzed the data by excluding the outliers (infants with PSB >14 mg/dL). The results are shown below.

View this table: [in this window] [in a new window]

 
Thus, excluding the outliers with high PSB did not change the ORs appreciably. There was a small reduction in the OR for death or neurodevelopmental impairment and for Psychomotor Developmental Index <70 and a small increase in the OR value for needing a hearing aid.

In regard to the emphasis of the risk factors affecting neurodevelopmental outcomes, it certainly was not our intention to deemphasize the well-documented effect of such risk factors as severe intraventricular hemorrhage or periventricular leukomalacia on neurodevelopmental outcomes in this population. In fact, we alluded to this association by citing the previously published data from our network.¹ Our discussion was focused on PSB simply because that was the subject of our investigation.

REFERENCE

1. Vohr BR, Wright LL, Dusick AM, et al. Neurodevelopmental and functional outcomes of extremely low birth weight infants in the National Institute of Child Health and Human Development Neonatal Research Network, 1993–1994. Pediatrics.2000; 105 :1216 –1226[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 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 Meadow, W. Articles by Wright, L. Search for Related Content PubMed PubMed Citation Articles by Meadow, W. Articles by Wright, L. Social Bookmarking                         What's this?