Corrigendum to “Guidelines for parenteral nutrition in preterm infants: The American Society for Parenteral and Enteral Nutrition”

Abstract

This corrigendum concerns the secondary outcomes for population, intervention, comparison, outcome, and time (PICOT) question 2.¹ The need for these corrections stems from the mistaken inclusion of a study that included more than one intervention. The Strommen et al study² had more than one intervention and so was unable to directly inform on the specific effect of higher vs lower amino acid dosing. The following corrections are put forth as a result of removing this study from analysis.

Removing the Strommen et al study meant removal of the study from Figure S4, which examined the relationship between intravenous amino acid dosing and bronchopulmonary dysplasia. The figure has been corrected in the Supporting Information for the article and is shown here as Supplemental Corrected Figure S4. The resulting statistic remains insignificant. The updated risk difference (RD) and 95% CI for the relationship between intravenous amino acid dosing and bronchopulmonary dysplasia is RD = 0.02 (95% CI = −0.09 to 0.14; P = 0.70). Figure S5, examining the impact of amino acid dosing on sepsis, has also been rerun with the study removed. Figure S5 has been corrected in the Supporting Information for the article and is shown here as Supplemental Corrected Figure S5. The result remains statistically insignificant. The updated RD and 95% CI for the relationship between intravenous amino acid dosing and sepsis is RD = 0.01 (95% CI = −0.07 to 0.09; P = 0.84). Given that funnel plots for assessing publication bias are not accurate when <10 studies are included, the funnel plot for Figure S10 is no longer valid. The figure has been removed from the Supporting Information for the article. To address these changes in the published guideline, on page 840 in the second paragraph of the ‘Secondary outcomes’ section for Question 2, the sentence, “However, one of these trials altered both ILE dose and composition and could not be included in the combined analysis” should be corrected to “However, both of these trials contained multiple interventions and could therefore not be included in the combined analysis.”

Finally, in the recommendation itself in Table 1 on page 832 and the text on page 835 for Question 2, we reference the Strommen et al study as part of the reasoning behind our decision. Had there been no evidence of harm in the Strommen et al study, this would have implied safety in initiating amino acid dosing at 3.5 g/kg/day. This was not the case. We acknowledge that we cannot say for sure whether the amino acid component of this intervention was the problem. For now, we believe that this study, along with the logic that follows in the recommendation, is useful for informing our expert opinion. We feel the verbiage should remain as published.

The supporting information file for the article has been replaced with a new one containing corrected figures. The original Figure S4 is shown below.

The corrected Figure S4 is shown below.

The original Figure S5 is shown below.

The corrected Figure S5 is shown below.

The original Figure S10 is shown below. It is no longer valid and has been removed from the Supporting Information for the article.