Intravenous sildenafil for treatment of early pulmonary hypertension in preterm infants

Data is lacking on the effect of continuous intravenous sildenafil treatment in preterm infants with early pulmonary hypertension (PH), especially in very low birth weight (VLBW) infants. Preterm infants (< 37 weeks of gestational age) with intravenous sildenafil treatment and diagnosis of PH between 01/12 and 12/21 were retrospectively screened for analysis. The primary clinical endpoint was defined as response to sildenafil according to the improvement of the oxygenation index (OI), the saturation oxygenation pressure index (SOPI) and PaO2/FiO2-ratio. Early-PH was defined as diagnosis < 28 day of life (DOL). 58 infants were finally included, with 47% classified as very low birth weight (VLBW) infants. The primary endpoint was reached in 57%. The likelihood to die during in-hospital treatment was more than three times higher (72 vs 21%, p < 0.001) in infants without response to sildenafil. The echocardiographic severity of PH and right-ventricular dysfunction (RVD) decreased significantly from baseline to 24 h (p = 0.045, and p = 0.008, respectively). Sildenafil treatment leads to significant improvement of the oxygenation impairment in 57% of the preterm infants, with similar response rates in VLBW infants. Intravenous sildenafil treatment is associated with a significant decrease of the PH-severity and RVD.


Material and methods
Study population. Infants treated at the NICU of the University Children´s Hospital of Bonn, Germany, during the study period 01/2012-12/2021 were retrospectively screened for study participation. Inclusion criteria: echocardiographic verification of PH, continuous intravenous sildenafil treatment, < 37 weeks of GA. Exclusion criteria were as follows: ≥ 37 weeks of GA, congenital diaphragmatic hernia (a study analyzing continuous intravenous sildenafil treatment in this subgroup was previously published 19 ), oral sildenafil treatment prior to intravenous treatment or exclusive oral treatment, primary palliative care, congenital heart defect requiring surgical repair.
Ethical approval. The study was approved by the local ethics committee of the Medical Center of the University of Bonn (local running number 138/22). Informed consent was waived by decision of the ethics committee of the Medical Center of the University of Bonn due to the retrospective design of the study. The methods used for the clinical research were performed in accordance with relevant guidelines/regulation and in accordance with the Declaration of Helsinki.
Monitoring data and oxygenation scores. The following hemodynamic parameters were retrospectively recorded from the patient´s chart at baseline, 3,6,9,12,24,48 h, and at day 7 after start of sildenafil treatment: systolic, diastolic, and mean arterial blood pressure, heart rate, pre-and postductal SpO 2 , FiO 2 . Arterial blood gas measurements with pH, paO2, and paCO2 were recorded when available at baseline, 3,6,9,12,24,48 h and day 7. The (a) Oxygenation Index (OI; ) 20 for infants without MV but with continuous positive airway pressure-CPAP or highflow-nasal cannula-HFNC support, and (c) PaO 2 /FiO 2 (P/F)-ratio (for infants with both, MV or non-invasive support) were retrospectively calculated for the timepoints baseline, 12, 24 and 48 h to evaluate oxygenation impairment. The vasoactiveinotropic score (VIS) was calculated for the same timepoints according to the formula described elsewhere for estimation of cardiovascular drug support 21 . Diagnostic and treatment of pulmonary hypertension. PH was defined as evidence of echocardiographic signs of PH and classified as early PH when appearance was in the first 28 DOL. These definition were adopted from the definition of PH in preterm infants described elsewhere 10,11,22 . Echocardiographic PH assessment is described below in detail.
PH drug treatment was conducted by the attending physicians according to in-house standards of clinical practice for vasoactive and PH treatment. Inhaled nitric oxide (iNO) was used as primary drug therapy when PH was treated. Infants were evaluated for sildenafil therapy when echocardiographic assessment demonstrated moderate or severe PH, FiO 2 was > 0.8 despite iNO treatment, and difference in pre-and postductal SpO 2 was > 8% (in case of a patent ductus arteriosus-PDA). Sildenafil was administered routinely as continuous intravenous infusion and a dose regime of 1.6 mg/kg/day, as described previously and according to the in-house clinical standards 19,23,24 . A bolus infusion of 0.4 mg/kg over a period of 3 h was administered followed by a continuous infusion according to the decision of the attending physician. For the period of sildenafil administration all infants were treated simultaneously with iNO.
Infants with need for inotropic support were treated with dobutamine and with milrinone. Furthermore, levosimendan was administered in infants with cardiac dysfunction despite high-dose inotropic treatment. In cases of oxygenation failure unresponsive to conventional treatment, extracorporeal membrane oxygenation (ECMO) was implemented according to international guidelines and criteria 25,26 . Echocardiographic assessment. For echocardiographic measurements a Philips CX50 Compact Extreme Ultrasound system with a S12-4 sector array transducer (Philips Healthcare, Best, the Netherlands) was used. All infants with diagnosis of early or late PH were evaluated by an echocardiographic assessment daily or (when stabilized) every 48-72 h, according to the in-house standards of clinical practice. All echocardiographic data at baseline (prior to start of sildenafil administration), at 24 h, and 48 h were retrospectively screened for analysis independently by two experienced neonatal echocardiographers. Both echocardiographers have over 10 years' experience in neonatal echocardiography as senior physicians, with profound knowledge of the echocardiographic diagnosis of PH and cardiac dysfunction in preterm and term infants in a tertiary referral medical center. PH was graded as mild, moderate, or severe, using the following echocardiographic parameters: (a) flow pattern of the ductus arteriosus (DA), (b) intraventricular septum position (IVS), and (c) tricuspid valve regurgitation (TVR). Tricuspid valve regurgitation was graded as I°, II° or III°. Mild PH was diagnosed as follows: DA shunt-flow was left-to-right, IVS was flattened, and TVR was I°-II°. Moderate PH was diagnosed as follows: DA shunt-flow was alternating (left-to-right/ right-to-left), IVS was flattened, and TRV was II-III°. Severe PH was diagnosed as follows: DA shunt-flow was right-to-left, IVS was D-shaped (towards left ventricular cavity), and TRV was III°. Additionally, the end-diastolic right-ventricular to left-ventricular (RV/LV) ratio was calculated in a standard four chamber directly distal to the tricuspid and mitral annulus as a horizontal line from endocardium of the RV and LV free wall to endocardium of the interventricular septum. Presence of ventricular dysfunction was defined as (a) right-ventricular dysfunction (RVD) and (b) left-ventricular dysfunction (LVD). For the assessment of ventricular dysfunction a combined approach of quantitative and qualitative measurements was used, based on international guidelines 27 www.nature.com/scientificreports/ Statistical analysis and outcome measures. Infants were retrospectively classified as sildenafil treatment Responder and non-Responder. A response to sildenafil treatment (primary outcome parameter) was defined as: decrease of the OI ≥ 20% (infants with MV) or the SOPI (infants without MV, but with non-invasive support) or an increase of the P/F ratio ≥ 20% (both, infants with and without MV) within the first 12-24 h after sildenafil. The following parameters were defined as secondary outcome measures: decrease of PH severity in the echocardiographic assessment during sildenafil therapy at timepoints 24 and 48 h, duration of MV, days of oxygen supplementation, oxygen supplementation at discharge, diagnosis of bronchopulmonary dysplasia (BPD) at 36 weeks of GA, survival to discharge. For this descriptive analysis, continuous variables were described using median and interquartile range (IQR). Categorical variables were summarized as absolute number (n) and percentage. For comparison of continuous and non-normally distributed variables, a Wilcoxon-test or Mann-Whitney U test was performed to compare continuous variables between timepoints and subgroups (Responder vs non-Responder), as appropriate. For categorical variables the Pearson's Chi 2 test and Fisher's exact test was applied, as appropriate.A Kaplan-Meier plot was used to analyze the survival chance after start of sildenafil treatment. A p-value of < 0.05 was considered significant.

Results
In total, 75 infants with documented diagnosis of PH and intravenous sildenafil treatment in the electronic patient´s chart system were screened for inclusion and 58 infants were finally enrolled for the analysis. 17 infants were excluded due to the following criteria: in 6 infants the patients' charts, and documentation was incomplete, in 9 infants' sildenafil was administered oral prior to the continuous intravenous infusion, and 2 infants had a diagnosis of late-onset PH (> 28 DOL).
Epidemiological data of the cohort are displayed in Table 1. The primary endpoint (response to sildenafil therapy) was reached in 57% (n = 33), with a response rate of 54% in VLBW infants and 59% in infants with a birth weight > 1500 g. Sildenafil was administered in median at the 2 (1/3) DOL. Almost the half of the cohort (n = 26, 45%) were classified as VLBW infants and 14 (24%) was born with < 28 weeks of GA (extremely low gestational age newborns-ELGAN). www.nature.com/scientificreports/ Significantly more infants in the non-Responder group suffered from a lung-hypoplasia (p = 0.001) and a lower urinary tract obstruction (LUTO) or renal congenital disorder (p = 0.031) at birth. Lung hypoplasia was defined according to the prenatal diagnosis by the obstetricians or clinical/radiological signs after birth (severe respiratory insufficiency with X-ray findings as small lung fields, diaphragmatic domes, and a bell-shaped) 29 . The rates of comorbidities after birth were equally distributed between the subgroups. The mortality rate was 43% in the overall cohort, with a significantly higher mortality rate in the non-Responder group (21 vs 72%, p < 0.001, see Kaplan-Meier-plot in Fig. 1). When looking for the combined endpoint death and/or BPD there between Responder and non-Responder infants there is still a trend to a higher rate of the combined endpoint in non-Responder infants (see Table 1, p = 0.05). Baseline characteristics in the first 24 h (APGAR 5 and 10 min, CRIP score and lowest FiO2) were similar between the subgroups.
Oxygenation scores and treatment data. The course of oxygenation according to the oxygenation scores (OI, SOPI and P/F-ratio) and allocation to the Responder and non-Responder group with related statistical significances is displayed in Fig. 2A The mean VIS scores after sildenafil treatment start are illustrated in Fig. 3A. In the Responder group, the mean VIS score increased significantly from baseline to 24 h (p = 0.004), without significant changes in the non-Responder group. The mean arterial pressure (MAP, see Fig. 3B) tended to increase during the sildenafil treatment. After commencing sildenafil the median arterial lactate levels decreased significantly from baseline (2.95 mmol/l) to timepoint 48 h (2.19 mmol/l, p = 0.037). No difference was found between Responder and non-Responder.
Comparison of outcome and treatment data between subgroups are illustrated in Table 2. No difference was found regarding the use of concomitant drugs for PH (iNO, bosentan or levosimendan) between Responder and non-Responder.
Echocardiographic assessment. In 42 infants (71%) valid echo data were available at baseline and 24 h.
Echo data at all three timepoints (+ 48 h) were available in 35 infants (59%). At baseline, 93% of the infants presented a persistent ductus arteriosus (PDA), with 33% presenting a right-to-left shunt over the PDA. At 24 h 95% of the infants presented a PDA (19% with right-to-left shunt). The course of PH, RVD and end-diastolic RV/LV-ratio is illustrated in Fig. 4A-C. In the overall cohort, the severity of PH and RVD decreased significantly from baseline to 24 h (p = 0.045, and p = 0.008, respectively). The RV/LV-ratio decreased significantly in the Responder group from baseline to 24 h (p = 0.028), without statistically significant difference in the overall cohort. Regarding the presence of LVD at baseline, 24 h and 48 h no differences were found between the Responder and non-Responder group.

Discussion
The present study summarizes a retrospective data set of 58 preterm infants with continuous intravenous sildenafil treatment and represents the biggest cohort to date of preterm infants with intravenous sildenafil treatment of early-PH. According to our results, we could demonstrate that PH severity, the incidence of RVD and the end-diastolic RV/LV ratio decreased significantly during the first 24 h during sildenafil treatment. Sildenafil www.nature.com/scientificreports/ treatment is associated with a rapid improvement in oxygenation impairment in the mayor part of the infants, illustrated and evaluated with established oxygenation indices (OI, SOPI, and P/F ratio), both shown in VLBW and non-VLBW preterm infants. The mortality rate of infants responding to sildenafil was significantly lower compared to non-Responder. Nevertheless, about 40% of the preterm infants did not demonstrate a response to sildenafil treatment and were classified as non-Responder. In recent prospective trials intravenous sildenafil treatment was evaluated in neonates with a gestational age > 34 weeks, but data on intravenous sildenafil in ELGAN and VLBW preterm infants < 28 GA are still lacking 13,14 . Several studies analyzed oral sildenafil treatment and only two recent studies were identified analyzing intravenous sildenafil treatment including VLBW infants with PH 15,17,30,31 . In most of these infants PH was diagnosed secondary to the timepoint of BPD diagnosis (36 weeks of GA). These studies overall report beneficial effects of sildenafil on echocardiographic grading of BPD-PH and oxygenation indices. A prospective multicenter, randomized, placebo-controlled trial on preterm infants < 29 weeks of GA at risk for PH is ongoing (SILDI-SAFE trial) 32 . Our study adds new insights on preterm and VLBW-infants with sildenafil treatment for primary early-PH in the first days of life, and with almost a third of the population classified as ELGANs and nearly half of the cohort classified as VLBW infants.
The response to sildenafil according to the improvement in oxygenation indices in our cohort was documented in 57% of the infants, without influence of GA or birth weight on the response to sildenafil. The updated   13 . All infants were on iNO treatment, which were the same conditions as in our study population. The authors concluded that intravenous sildenafil was not superior to placebo regarding the endpoints: treatment failure rate (need for additional drug for PPHN therapy, need for ECMO, or death prior to discharge, and time on iNO after starting sildenafil in infants without treatment failure). This phase III study raises the question, whether continuous intravenous sildenafil is effective to treat PPHN in preterm and term infants. However, the findings of the study are contrarious to other studies revealing a significant effect of intravenous sildenafil in this population 14,23 . The dose regime of sildenafil used in the phase III trial was half of the dose which was described by Steinhorn et al. 23 , using a bolus infusion of 0.4 mg/kg over 3 h, followed by a maintenance dose of 1.6 mg/kg/day. The missing effect between placebo and sildenafil could potentially be biased by lower plasmatic sildenafil levels due to lower concentrations of the bolus and continuous infusion, and in total cumulative half of the dose used in the open-label, dose-escalation trial by Steinhorn et al. In our study, infants were treated with the dose regime provided by Steinhorn et al., but a bolus infusion was only administered in 50% of the infants, due to the decision of the attending physician. The severity of oxygenation impairment and PPHN in our cohort is comparable with the before mentioned studies, as the median OI at baseline prior to sildenafil start in our cohort was 26.2 (12.0/47.3) and both studies predominantly included infants with an OI > 15 and < 60. In 4 infants of our cohort the continuous sildenafil dose was doubled to 3.2 mg/ kg/day after cessation of the bolus infusion as rescue procedure in case of severe PPHN or oxygenation failure, without difference in response to sildenafil between subgroups (see Table 2).
The preterm infants screened for this retrospective analysis were predominantly those with a delayed cardiopulmonary adaptation and impaired transition of the fetal circulation. As reported recently, preterm infants with early-PH and delayed cardiopulmonary adaptation (PH diagnosis > 48-72 h) are at risk for adverse outcome and BPD, as highlighted recently 33 . This is in line with our findings, with an overall rate of BPD of 26% in the cohort and a rate of 36% of those responding to sildenafil and surviving to discharge.  www.nature.com/scientificreports/ Non-Responder in our cohort suffered significantly more often from lung-hypoplasia, and LUTO or congenital-renal-disorder. Infants with LUTO or congenital-renal-disorder mainly suffer from lung-hypoplasia due to oligo-or anhydramnion. Preterm infants suffering from lung-hypoplasia seem to be prone to respond less to sildenafil, which potentially can be explained by a higher degree of structural lung abnormality (vasculopathy: excessive muscularization, intima thickening, rarefaction of pulmonary arteries and vessels) and higher grade of oxygen impairment at baseline. As sildenafil is a second-line PH-therapy, the use of sildenafil in these infants needs to be evaluated carefully. The term "lung hypoplasia" is not standardized and clear definitions are missing in the current era. Therefore, this diagnosis is to a certain degree subjective.
Our results illustrate, that intravenous sildenafil treatment is associated with a significant and rapid reduction of the PH severity. This is in line with previous studies on infants with PPHN and neonates suffering from PH due to congenital lung disease or BPD-PH 15,17,19,30,31 . According to our results, the RVD improves significantly during the continuous sildenafil infusion. Several studies analyzed the relationship between sildenafil treatment and right ventricular diastolic and systolic function, as shown by magnet resonance tomography scans, in right heart catheterization and in animal models for right ventricular dysfunction [34][35][36][37][38] . In summary, sildenafil has afterload-reducing effects with improvement of RV unloading, leads to an improvement of diastolic RVD, improves RV myocardial remodeling due to upregulation of gene-markers for hypertrophy and inflammation, and there are conflicting results regarding the improvement of the RV systolic function. During the sildenafil treatment the RV/LV ratio decreases significantly, potentially illustrating the effect of RV unloading and higher preload of the left-ventricle in preterm infants. The incidence of LVD initially decreased from baseline to 24 h in non-Responder, followed by a significant increase from 24 to 48 h after start of sildenafil treatment in non-Responder. Sildenafil possibly can lead to a deterioration of a preexisting LV dysfunction, due to a aggravation of the LV dysfunction caused by a higher LV preload and filling pressures, which is in line with previous findings of sildenafil treatment in infants with a congenital diaphragmatic hernia 19 . On the other hand, results from adult RCT trials revealed beneficial effects of sildenafil on LV function 39 . Limitations. Our data provide important information in these patient population and we hypothesize that continuous sildenafil treatment in preterm infants with early-PH seems to be well tolerated. However, this need to be interpreted carefully, as the frequent use of concomitant vasoactive treatment could have mitigated an arterial hypotension as drug related adverse event during the sildenafil treatment. During the treatment period in 9 infants a severe arterial hypotension (5 mmHg < GA for > 60 min) was identified in the documentation system, but these episodes were not related to the sildenafil treatment, as the episodes occurred unrelated to the start of the bolus or continuous sildenafil infusion. The present study was conducted retrospectively and not designed to evaluate the safety profile of sildenafil. This needs to be considered when planning future prospective randomized-controlled trials.
A comparator cohort is missing and as mentioned before, prospective randomized trials are warranted. Therefore, it is important to get knowledge from retrospective studies as provided by our results. Furthermore, the interpretation of the echocardiographic data is at risk for bias because echocardiographic assessment is based to some extent on a qualitative grading. Finally, sildenafil treatment was started as second-line therapy in critically ill infants with a high-degree of preexisting comorbidities. This can have biased the incidence of the response rate of infants to sildenafil treatment.

Conclusion
Continuous intravenous sildenafil treatment is associated with a significant improvement of the oxygenation impairment in preterm infants, including ELGAN and VLBW infants. Additionally, continuous intravenous sildenafil treatment is associated with an improvement of PH, RVD and RV/LV ratio in the echocardiographic assessment in this population. Nevertheless, a substantial proportion (43%) of preterm infants do not respond to sildenafil, especially infants with prenatal diagnosed lung-hypoplasia. These infants are at high risk for inhospital mortality.

Data availability
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.

Funding
Open Access funding enabled and organized by Projekt DEAL.