Reply to the Letter by Ramakrishnan et al.: “Methodological Considerations in the Comparison of INSURE and LISA in Very Preterm Infants”

Dear Editor,

We have read the letter by Ramakrishnan et al., titled “Methodological Considerations in the Comparison of INSURE and LISA in Very Preterm Infants” [1]. We provide a reply to their comments below.

Regarding the robustness of our findings, we performed model validation analyses in our original article [2]. In these analyses, group variables of interest remained not statistically significant, supporting the stability of the main results. However, our study was based on national registry observational data (n = 1,112) from the Korean Neonatal Network (KNN), and institutional level information on individual neonatal intensive care units (NICUs) was not available. Consequently, it was impossible to fully control of institution-related confounding variables. This was recognized as an inherent limitation of the available data and was addressed in the original article [2].

For the comments on the dichotomization of ventilation duration using receiver operating characteristic (ROC)-derived cut-offs, both the duration of invasive ventilation and total respiratory support were not normally distributed. The nonnormal distribution of these variables could affect the validity of simple statistical methods and was therefore an important consideration in our statistical approach of this study. Therefore, after consulting with a statistics expert, we used ROC curve analysis to determine the appropriate cutoff values. This approach was specifically chosen to account for the nonnormal distribution of data and establish clinically meaningful cutoff values. Based on the ROC curve analysis, the cutoff values were set at 2 days for the duration of invasive ventilation and 7 days for the duration of total respiratory support. This methodology allowed us to handle data distribution characteristics and assess the variables in question. However, methodological limitations still exist and were mentioned in the original article [2].

We agree that center-specific practices may influence outcomes such as patent ductus arteriosus and sepsis. Unfortunately, however, the KNN database does not include data identifying individual NICUs, making it impossible to adjust for NICU-level influences or clustering effects.

The time to first surfactant administration was slightly shorter in the INSURE group than in the LISA group (51 min vs. 60 min, respectively, p < 0.001), but both groups received the surfactant therapy within the generally recommended early time window and the difference is unlikely to be clinically meaningful. With respect to the higher frequency of repeat surfactant dosing in the LISA group, this finding may be attribute to the greater immaturity of infants in this group as well as potential procedural difficulties. The KNN database does not consistently distinguish the method used for repeat dosing (LISA, INSURE, or intratracheal administration), nor does it allow for a standardized definition of LISA “failure” based on subsequent intubation. Therefore, analyses comparing outcomes between LISA success and failure subgroups could not be reliably performed and represent an important area for future research.

Acknowledgments

The authors would like to thank Professor Kyung Sook Yang of the Institute of Statistics, Korea University College of Medicine, for her helpful advice regarding statistical analyses.

Conflict of Interest Statement

The authors have no conflicts of interest to declare.

Funding Sources

This study was not supported by any sponsor or funder.

Author Contributions

H.C.: conceptualization, writing – original draft preparation, review, and editing. J.L.: conceptualization, writing – review and editing, supervision, and final approval of manuscript.

Funding Statement

This study was not supported by any sponsor or funder.