Discussion:
The respiratory course of premature infants in the NICU plays a significant role in determining oral feeding success. Preterm infants <30w GA are at increased risk for pulmonary complications, especially BPD.7,8 BPD can lead to higher respiratory support requirements, delaying the initiation of oral feeding attempts.9 Gianni et al demonstrated that PMA at the time of achieving full oral feeding was significantly earlier as gestational age or birth weight increased. BPD was significantly associated with an older PMA at full oral feeding.8Our data supports this conclusion as well. We showed that BPD is associated with an older PMAff, the most important variable in our original model describing risk for GT placement.4
Premature infants do not have mature suck-swallow-breathe coordination, limiting their ability to orally feed. Preterm infants are prone to initiate swallowing at unsafe phases of the respiratory cycle, increasing the risk of desaturation and aspiration.10The earliest successful synchronization of sucking, swallowing, respiration and esophageal function has been reported to occur in premature infants sometime between 32-34w PMA.11 The development of this coordination is further delayed in premature infants with BPD, who often require higher levels of respiratory support for longer periods of time.12,13 Positive pressure support has been suspected to inhibit closure of the larynx and interfere with the swallowing reflex.2,3 Infants on nasal CPAP have been shown to have increased risk of aspiration during oral feeds, likely due to interfering signals from the positive pressure.11
In our original model development, although every respiratory variable was associated with the GT/no GT outcome in univariate analysis, only exposure to HFV remained in the final model.4 We suspected that respiratory status is highly correlated to the most significant variable in the model, PMAff, and now provide evidence supporting that hypothesis. However, this conclusion may be confounded because there is variability amongst NICUs regarding when infants may or may not receive oral feeding attempts.14 For instance, our NICU policy restricted oral feeding attempts to infants who were on nasal cannula flow < 0.5 LPM. This fact provides an interesting area for further study. Additionally, this suggests we should be more thoughtful about when we introduce or withhold oral feeding for infants requiring high respiratory support during their NICU course. There are likely critical periods of neurodevelopment during which we need to engage neural and muscular pathways for oral feeding; if this window of time is missed, it may become much more difficult for later development of this skill.15
We noted a significant gender difference in infants requiring a GT (63% male) versus infants not requiring a GT (45% male) in the original test cohort. This was significant in the test cohort on univariate analysis but did not remain in the multivariate model. In the separate validation cohort, this was no longer significant (57% vs. 49%). However, when the two cohorts are combined, the p value is again significant: 47/78 (60%) males in the GT infants, vs. 148/313 (47%) in the non-GT cohort, P = 0.040 by chi-square. Two recent studies reviewing GT trends by collecting retrospective data across multiple hospitals reported no gender differences in infants who received GTs (n = 333 and 360). Studies have shown that males are more likely to develop BPD and more likely to have more severe BPD and this may explain the sex difference in GT outcome.16 Since this finding seems quite variable, it should be examined in larger data sets to assess its reproducibility.
In conclusion, our data show how NICU respiratory course affects achieving full oral feeding. The documentation of respiratory progression during the NICU course, showing improvement in the need for respiratory support from day 30 of age to 32w PMA, and finally 36w PMA, may guide us in finding an opportunity to allow more oral stimulation and feeding during this time. One example, currently being employed at UH Rainbow Babies & Children’s Hospital for older NICU infants with BPD who are on HFNC, is decreasing flow to 2 LPM just for the 20-30 minutes of an oral feeding, which has proven successful thus far (personal communication, Dr. Monika Bhola). The unit policy is to allow oral feeding if an infant is on ≤2 LPM. We are hoping to study this practice as part of a quality improvement project. Are the differences among NICU practices something that can teach us best practice? Our data offers the impetus for those who are experts in oral feeding to help us do better.