During the study period, 456 preterm infants below 33 weeks of gestational age were assessed for eligibility. One hundred thirty-one and 201 preterm infants completed the analysis in the intervention and control groups, respectively (Figure 1). There were no differences in the proportions of excluded infants or in specific reasons for exclusion between the study groups (supplementary Table s2).
The clinical characteristics and hemodynamic management of the patients are summarized in Table 1. Twenty-nine (22.1%) preterm infants in the intervention group were diagnosed and treated for LSBF per protocol, whereas 6 (2.9%) received open-label inotrope for LSBF in the control group (p<0.001). The need to treat arterial hypotension was not different between the intervention and control groups [21 (16%) vs 23 (11.4%); p=0.228]. The use of inhaled nitric oxide for pulmonary hypertension was more frequent in the intervention group [11 (8.4%) vs 4 (2%); p=0.007]. Overall, thirty-six (27.5%) preterm infants in the intervention group and 27 (13.4%) preterm infants in the control group were treated with vasoactive medications during the first 3 days of life (p=0.001). Dobutamine was more frequently used in the intervention group (86.1% vs. 25.9%, p<0.001), while dopamine was more frequently used in the control group (88.9% vs. 27.8%, p<0.001). There were no differences in the use of epinephrine or norepinephrine, but the maximal dose of epinephrine was greater in the control group than in the intervention group (0.55 vs. 0.20 mcg/kg/min, p=0.037). Overall, a purely inotropic strategy (63.5%) predominated in the intervention group, and a purely vasopressor strategy (55.6%) predominated in the control group.
All echocardiography measurements had >30% missing values, so the data were not imputed. Complete data on echocardiography measurements of SBF were available for 191 infants from 7 centers (121 in the intervention group and 70 in the control group). Preterm infants in the intervention group had higher values of MPAVpeak and RVO at 6, 12, 24 and 48 hours compared to the control group (p<0.05 in all instances) (supplemental Table s3) The analysis of the trajectories of the echocardiography parameters showed that preterm infants who experienced the main outcome (> grade II IVH or ED) had lower SBF values and remained in a low-flow zone from 6 to 24 hours of life despite belonging to the intervention or control group or receiving inotropic drugs per-protocol or as open-label treatment for LSBF (Figure 2)
Overall, 86 (26%) preterm infants in the cohort experienced IVH of any grade, 45 (13.6%) experienced > grade II IVH, and 27 (8.2%) experienced grade III IVH and/or periventricular infarct. IVH was first detected by ultrasound at a mean (SD) of 35 (22) hours of life, without significant differences between the study groups. No IVH was detected beyond the first 7 days of life.
The IPTW attained an adequate balance of measured prognostic factors of IVH between the intervention and control groups (see Table 2 and supplemental Table s4). The results of the primary analysis based on IPTW are shown in Table 3. Preterm infants in the intervention group had a lower incidence of > grade II IVH or ED than did those in the control group in the original cohort [9.9% vs. 19.4%, odds ratio (OR) 0.458 (95% CI: 0.234-0.895); p=0.022] and in the IPTW cohort [8% vs. 23%, OR 0.285 (0.133-0.611); p=0.001]. The mixed effect model accounting for clustering within neonatal units also confirmed the association between the intervention and reduced > grade II IVH or ED [OR=0.292 (95% CI=0.149-0.572); p<0.001]. The occurrence of any grade IVH or high-grade IVH as well as posthemorrhagic ventricular dilatation was also lower in the intervention group.
Supplementary analysis
Complementary analysis using multivariate regression was concordant with the primary analysis, revealing an independent association between the intervention and > grade II IVH or ED [OR=0.227 (0.09-0.556), p=0.001]. The use of dopamine [OR: 4.8 (95% CI: 2.1-13.3, p=0.001], delivery room resuscitation [OR 3.2 (95% CI: 1.5-.7.4; p=0.003] and CRIB II [OR 1.3 (95% CI: 1.2-1.4); p<0.001] were other factors independently associated with > grade II IVH or ED (supplemental Table s5). Subgroup analysis revealed a significant association between the intervention and > grade II IVH or ED only in preterm infants younger than 29 weeks of gestational age [OR=0.140 (95% CI=0.04-0.451); p=0.001] but not in older infants (supplemental Table s6 and s7)
A new propensity score was estimated by logistic regression after excluding infant who received any hemodynamic intervention (namely, vasoactive medications, hydrocortisone, fluid boluses, inhaled nitric oxide or PDA treatment) during the first 3 days of life. Balance diagnostics were analogous to the main analysis. IPTW achieved a good balance of all prognostic factors used to estimate the PS between the intervention and control groups (standardized differences ranging -4.8% to +10.1%). This sensitivity analysis did not reveal a reduced incidence of > grade II IVH or ED in the intervention group compared to the control group neither in the original cohort [OR 0.402 (95% CI: 0.11-1.4; p=0.165] nor in the IPTW cohort [OR 0.389 (0.17-1.4) p=0.142] (supplemental Table s8)
Multivariate regression also showed a significant association between the intervention and reduced grade III IVH-periventricular venous infarct or ED [OR: 0.359 (95% CI: 0.130-0.991); p=0.048] and reduced grade III IVH-periventricular infarct [odds ratio 0.253 (95% CI: 0.08-0.756); p=0.014] (supplementary Table s9 and s10)
The negative control analysis did not reveal an association between the intervention and a set of selected preterm morbidities (need for mechanical ventilation, PDA treatment, late-onset sepsis, necrotizing enterocolitis bronchopulmonary dysplasia and death) nor in the original cohort neither in the IPTW cohort (supplemental Table s11)