DISCUSSION
We found that LU scores do not change in the same way in all PTB32W after diuretics.
As we have demonstrated previously 10, BPD patients have higher LU scores than non-BPD patients, and maintain them high until 36 weeks’ PMA, while non-BPD patients experience a decrease in LU scores since the very first week of life, and maintain them close to zero.
In this study, we show that there is a specific group of patients whose LU scores improve with diuretic therapy, while the other group maintain them high. Diuretics were started on day 30 to 35 in both groups, and the number of BPD patients in both groups was similar, so the expected LU score evolution in all of them would have been to remain high until 36 weeks’ PMA, according to our previous study.
In the responders group, the RS was reduced since the very first interval of time studied, when we compared it to the other group. On parallel, LU score also improved in the responders group, although a few days later: probably the small sample size made that not all intervals reached statistical significance when comparing LU score after diuretics.
This can be explained because BPD is a multifactorial disease, not just influenced by an increase in lung extravascular water. LU is a very sensitive method to detect an excess of lung extravascular water15, and that makes LU very useful to study the lung water content in neonates at birth 25, to predict the need of admission and respiratory support 4, and non-invasive ventilation failure 5,6. But the lung with BPD also includes an inflammatory damage, with a decrease of alveolar type cells in the alveoli and of elastin fibres in the extravascular matrix, while there is an increase of collagen fibres that grow under-regulated, and a decrease in pulmonary microvasculature26. Although LU score correlates with the degree of lung inflammation in preterm infants with hyaline membrane disease16, it is very usual to have infants with evolving BPD, or recovering from BPD, whose LU scores remain the same thorough their initial admission in NICU 10. This condition can explain why LU improves later than clinical condition in preterm infants on diuretics.
Our findings should be interpreted with caution: although GA, birth weight, CRIB, CRIB-II and SNAPPE-II indexes were similar in both groups, half of the responders group were extubated at diuretics initiation, while just one patient in the non-responders group. As this is a retrospective analysis, diuretics were prescribed according to the assistant neonatologist’s criteria, thus it could be that the responders group was composed by a subgroup of PTB32W with a milder form of BPD than the non-responders group: in fact, we had less moderate-severe BPD patients in the responders group (25% versus 89% in the non-responders group, p=0.005), and they received shorter courses of diuretics (12 days (IQR 4-14) versus 59 days (IQR 45-65), p=0.035).
As we described earlier 10, LU score decreases later in the most immature patients that don’t develop moderate-severe BPD: patients born before 28 weeks of GA have higher values of LU score physiologically, even though they have healthy lungs, and LU score in them cannot predict BPD as early as the first week of life. It could well be that our responders’ group are those more immature infants without moderate-severe BPD, that would have improved their LU scores anyway, without needing diuretics.
For all these reasons, the main limitation of our study is its retrospective and not-blinded design, because we cannot assure that the changes described in LU score in the responders group are due to diuretic treatment.
Further research on this topic is warranted, as maybe LUS score progression after treatment (diuretics, corticosteroids…) could be a suitable biomarker of DBP development in PTB32W.