To the Editor: I have read with interest the provocative article by Soorani-Lunsing et al. (1), as well as the accompanying editorial and commentaries by experts in the field (2–5). The authors of this study report that moderate hyperbilirubinemia is associated with an increase in minor neurologic dysfunction (MND) with a “strong” dose-response relationship. In another publication by some of the same investigators, infants with MND were found to be at high risk for subsequent behavioral and learning problems at school age (6). Based on these findings the authors recommend lowering bilirubin levels before they get to the “danger zone” (≥335 μmol/L).

In his editorial, Ohlsson (2) raised several serious methodological concerns with ascertainment of the cohort, small sample size, subjective assessment measures, unblinded assessments, and most surprisingly, lack of bilirubin measurement in the control group. Together these preclude an unbiased answer to the study question. None of these limitations were addressed by the authors in their paper. In recommending a more aggressive approach to phototherapy to lower bilirubin levels, the authors cite a few minor side effects in the neonatal period. However, other investigators have shown that visual and auditory orientation responses in infants treated with phototherapy are compromised when phototherapy is terminated on day 4, and persist for as long as 1 month later (7). Poor ability to follow human face and voice may have the potential to impair early developmental interaction between the infant and the parents. Further, as Ohlsson (2) and Hintz and Stevenson (3) point out, the study period of 12 months is insufficient time to evaluate the ultimate impact of hyperbilirubinemia on later childhood outcomes. Unfortunately, follow-up of this particular cohort to an older age will not necessarily provide more definitive answers due to the biases alluded to earlier.

The three accompanying commentaries provide valuable insights into many important issues on the subject, despite the methodologic shortcomings of the original article. Maisels and Newman (4) rightly caution that before changing current practices, we need better evidence not only that the intervention is effective, but also that the benefits justify the risks and costs. At best, the article by Soorani-Lunsing et al. raises concerns which require further large-scale, well-designed prospective studies to determine whether infants are being harmed by what were previously considered “safe” levels of plasma bilirubin, and whether these effects are indeed long lasting. Given the substantive methodologic concerns, one cannot possibly conclude that a causal relationship exists between moderate hyperbilirubinemia and MND.