Diuretics for Management of Bronchopulmonary Dysplasia Associated Pulmonary Hypertension in Preterm Infants—A Systematic Review

ABSTRACT

Bronchopulmonary dysplasia associated pulmonary hypertension (BPD‐PH) in preterm infants results from impaired pulmonary vascular development and contributes significantly to morbidity and mortality. Diuretics are commonly used to reduce pulmonary vascular congestion and improve oxygenation, but their efficacy and safety remain unclear. This systematic review aimed to evaluate the efficacy and safety of diuretic use in preterm infants with BPD‐PH. Primary outcomes were mortality and need for home respiratory support; secondary outcomes were echocardiographic and clinical parameters, and adverse effects such as hyponatremia, nephrocalcinosis. We searched MEDLINE, Embase, Cochrane Central, and ClinicalTrials.gov from inception to October 2025. We included randomized controlled trials (RCTs) and non‐randomized studies of interventions (NRSI) evaluating diuretics in preterm infants (< 32 weeks' gestation) with BPD‐PH diagnosed by echocardiography at 36 weeks postmenstrual age. Two reviewers independently performed data extraction and risk‐of‐bias assessment (ROBINS‐I). Outcomes were synthesized narratively, and a meta‐analysis was performed where sufficient data were available. Of 492 studies identified, 2 NRSI (n = 49) met inclusion criteria, both were at serious risk of bias. No RCTs exist on diuretics for BPD‐PH in preterm infants. Mortality, respiratory, discharge outcomes and adverse effects were reported inconsistently. Meta‐analysis of limited data demonstrated small improvements in echocardiographic markers, but the certainty of evidence was very low due to heterogeneity between the two included studies. To conclude, evidence supporting the use of diuretics in BPD‐PH in preterm infants remains limited and of low certainty. High quality studies are needed to better define their safety and efficacy.

1. Description of the Condition

Interstitial pulmonary edema plays a central role in the pathogenesis of bronchopulmonary dysplasia–associated pulmonary hypertension (BPD‐PH) in preterm infants, contributing to increased pulmonary vascular resistance (PVR) and impaired lung function. This review describes the evidence for use of diuretics either loop, thiazide or potassium sparing diuretics in preterm infants with pulmonary arterial hypertension (PAH).

Bronchopulmonary dysplasia (BPD) represents a specific form of chronic lung disease (CLD) characterized by disrupted lung development and/or injury resulting from preterm birth [1, 2]. In extremely preterm infants, the incidence of BPD, defined as the need for respiratory support required at 36 weeks postmenstrual age (PMA), ranges from 23% to 85%, with the highest rates observed in those born at the lowest gestational ages [3, 4, 5, 6, 7]. Reported prevalence of BPD‐PH varies widely, ranging from 3% to 44% among infants with BPD. Mortality in this population is also high, with rates between 20% and 70%, depending on the severity of the pulmonary hypertension (PH) [8, 9, 10, 11]. BPD‐PH is associated with increased morbidity during hospitalization, including extended length of stay, longer duration of ventilation, increased need for home respiratory support, and ongoing respiratory complications. It is also associated with persistent pulmonary vascular disease and poorer neurodevelopmental outcomes compared with preterm infants who do not have PH [12, 13, 14, 15].

BPD‐PH may result from (i) increased vascular resistance in pulmonary arteries [pre‐capillary or PAH—most common phenotype], (ii) excessive pulmonary blood flow for example, from left‐to‐right shunts such as atrial septal defect (ASD), ventricular septal defect (VSD), or patent ductus arteriosus (PDA) [flow‐mediated PH] (iii) left ventricular dysfunction (due to intrinsic impaired left ventricle (LV) relaxation in preterm infants, systemic hypertension) or pulmonary vein stenosis (PVS) [post‐capillary PH] or (iv) a combination of any of these [16, 17, 18, 19, 20, 21, 22, 23, 24, 25]. In this review, we focus on PAH phenotype. Other phenotypes were excluded because they have not been consistently reported in larger cohorts; the literature consists mainly of case reports and small case series, and no standardized diagnostic criteria, or proposed treatment options have been established [16, 21, 26, 27]. Hence, in this review, the term BPD‐PH refers exclusively to the PAH phenotype.

Major factors driving the development of BPD‐PH (increased PVR) include ongoing inflammation, disrupted alveolar growth, immature lymphatic function, and constricted, stiff pulmonary arteries. In the early stages of BPD‐PH, elevated PVR and abnormal vasoreactivity are the primary features. As the disease advances, structural changes emerge, such as intimal proliferation, thickening of the adventitia, and smooth muscle hypertrophy [28]. These alterations ultimately produce fragile, permeable pulmonary capillaries, where elevated intravascular pressures place hydrodynamic stress on the vessel walls. This contributes to interstitial edema and diminished lung compliance, which in turn further elevates extra‐alveolar PVR i.e. resistance to blood flow in large pulmonary vessels (arteries and veins) that are not directly in contact with the alveoli [18, 29, 30, 31]. Above changes result in capillary compression, impaired gas exchange, and worsening hypoxia and vasoconstriction. If not interrupted, this can lead to dilation of right ventricle (RV), right heart failure, and higher mortality in this high‐risk population [32, 33].

1.1. Description of Diagnosis of BPD‐PH

Though cardiac cathetrization remains the gold standard for diagnosis of BPD‐PH, due to its invasive nature, it is usually not preferred in routine clinical practice in preterm infants. Pediatric Pulmonary Hypertension Network (PPHNet) and European Pediatric Pulmonary Vascular Disease Network (EPPVDN) suggests echocardiography at around 36 weeks PMA to screen for BPD‐PH in preterm infants who are dependent on respiratory support [23, 34]. Various echocardiographic measures of elevated RV pressure such as septal flattening, RV dilation or hypertrophy and reduced cardiac function, tricuspid regurgitant (TR) jet velocity are helpful in the diagnosis of PH. It has also been suggested to evaluate left ventricle (LV) function, and pulmonary veins assessment.

1.2. Description of the Intervention

Management of BPD‐PH focuses on symptom relief and the prevention of disease progression. This includes close monitoring of clinical status, optimizing ventilation and oxygenation, and ensuring appropriate fluid intake with high calorie intake to minimize fluid overload and support adequate growth [35]. Diuretics are commonly employed to reduce interstitial lung edema and consequently PVR, while pulmonary vasodilators are used to decrease PVR directly on pulmonary vessels and reduce stress on RV [23, 34, 36]. Loop diuretics, such as furosemide, enhance urinary sodium excretion by inhibiting the sodium‐potassium‐chloride cotransporter in the ascending loop of Henle. Diuretic regimens, particularly the combination of hydrochlorothiazide and spironolactone, are commonly employed over few weeks due to their complementary actions: hydrochlorothiazide inhibits the sodium chloride cotransporter in the distal convoluted tubule of kidney, and promotes sodium and water excretion, while spironolactone mitigates potassium loss and offers additional cardiovascular and vascular benefits through aldosterone antagonism. Observational studies have suggested possible clinical benefits such as symptomatic improvement, stabilization of disease, post‐discharge respiratory and neurodevelopmental outcomes with diuretic therapy in preterm infants with BPD‐PH [18, 37, 38]. The European Pediatric Pulmonary Vascular Disease Network includes diuretics as part of first‐line management for BPD‐PH in preterm infants [34]. In the PH in BPD (PUsH BPD) survey done in Germany, most neonatologist respondents reported that they would use diuretics for infants with BPD‐PH [39].

1.3. How the Intervention Might Work

Diuretics may play a therapeutic role in managing BPD‐PH by reducing interstitial edema which in turn can improve alveolar ventilation and lung compliance. This reduction in pulmonary congestion reduces PVR and alleviates respiratory distress, improves respiratory support needs, slowing the progression of PH and decreases RV afterload, thereby mitigating the stress on RV [18, 40, 41]. The extent of symptomatic improvement varies among patients and that diuretics are adjuncts, not definitive treatments for PH.

1.4. Why It Is Important to Do the Review

The trajectory of BPD‐PH is heterogenous, and interstitial pulmonary edema as a potential key contributor to BPD‐PH may be under‐appreciated [18, 28]. During the critical window of rapid postnatal lung development, the impact of diuretic therapy on PVR, RV function, or lung compliance remains unclear. Diuretics are widely used in neonatal intensive care units (NICUs) to manage fluid overload yet their specific role in reducing interstitial pulmonary edema in infants with BPD‐PH remains poorly defined. There is a lack of systematic evidence evaluating their impact on echocardiographic markers of PH and clinically important outcomes. A rigorous synthesis of the available evidence is needed to determine whether diuretics reduce pulmonary edema, improve cardiopulmonary physiology, and translate into meaningful clinical benefit without substantial adverse effects.

1.5. Objective

In this systematic review, we aimed to assess the efficacy and safety of diuretic therapy in preterm infants diagnosed with BPD‐PH at 36 weeks PMA. The primary objective were mortality or need for respiratory support at discharge. Secondary objectives were to assess the (a) improvement in clinical markers and physiological parameters (type of respiratory support, mean airway pressure (MAP), fractional inspired oxygen concentration (FiO₂), blood pressure (BP) and respiratory rate (RR) (b) improvement in echocardiographic markers of PH, and cardiac function, (c) intermediate‐term outcomes—morbidity and healthcare resource utilization such as duration of respiratory support, duration of oxygen therapy, length of stay, (d) adverse effects of diuretics which includes hyponatremia, poor weight gain, ototoxicity, nephrotoxicity and metabolic bone disease.

2. Methods

This review followed the Preferred Reporting Items for Systematic Reviews and Meta‐analyses (PRISMA) 2020 statement and registered with PROSPERO, an international registry for systematic reviews (CRD42024623128). The completed PRISMA checklists for manuscript and abstract are provided as Supporting Information S1: Table S1 and S2.

2.1. Criteria for Selecting Studies for This Review

2.1.1. Type of Studies

We included randomized controlled trials (RCTs) as well as non‐randomized studies of interventions (NRSI), including observational cohort studies. Studies reporting on the same infant population more than once were excluded, with preference given to the most comprehensive or recent dataset. All studies reporting relevant clinical or physiological outcomes were considered, regardless of whether a control group was included.

2.2. Type of Participants

Participants were required to meet all of the following criteria: (1) preterm infants born before 32 weeks gestational age (2) developed BPD (respiratory support need at 36 weeks PMA) and (3) diagnosed as having BPD‐PH by echocardiography at around 36 weeks PMA. PAH is identified in echocardiography by any one of the following, (a) right to left or bidirectional shunt across PDA or ASD (b) flattening of the interventricular septum (IVS) on end‐systole in short axis view at the level of the papillary muscle as measured by the LV end systole eccentricity index (LVs EI) > 1.3, (c) right ventricular systolic pressure (RVSP) as measured by tricuspid regurgitation (TR) jet > 40 mmHg [23, 42]. Echocardiographic evidence of reduced RV function or RV dilation are not needed for the diagnosis of BPD‐PH but if present indicates the severity of PH. Preterm infants who received diuretics for BPD without a PAH diagnosis were excluded to ensure that the effects of diuretic therapy were evaluated specifically in the context of BPD‐PH, rather than for BPD or fluid management alone. This approach reflects clinical practice, where echocardiography and clinical features guide decisions regarding the initiation, continuation, or discontinuation of therapy, instead of using it prophylactically in all infants with evolving or established BPD.

2.3. Type of Interventions

The intervention group included preterm infants born before 32 weeks' GA with a BPD‐PH diagnosis at 36 weeks PMA who received either a loop diuretic (furosemide), thiazide diuretics (hydrochlorothiazide plus spironolactone), or a combination of loop and thiazide diuretics, administered orally or intravenously for at least 3 consecutive days. Studies were eligible if they assessed the effects of diuretic therapy, either by comparison with a separate control group or using a pre‐post (self‐controlled) design in infants with BPD‐PH. A planned subgroup analysis included infants receiving additional pulmonary vasodilators (e.g., sildenafil or bosentan) with BPD‐PH.

2.4. Types of Outcome Measures

2.4.1. Primary Outcomes

The primary outcomes are mortality or the need for home respiratory support at discharge. Respiratory support includes low‐flow (LF) oxygen, high‐flow nasal cannula (HFNC), continuous positive airway pressure (CPAP), non‐invasive positive pressure ventilation (NIPPV), or tracheostomy tube ventilation at the time of discharge. These outcomes were chosen for its direct clinical relevance to extremely preterm neonates with BPD‐PH as it reflect both the most serious consequence (death) and ongoing morbidity requiring intervention after discharge. This dual focus provides a comprehensive measure of disease impact and therapeutic effectiveness, aligning with real‐world priorities and established standards in neonatal pulmonary hypertension research [43, 44]. There was no minimum requirement for specific outcome domains ‐ studies reporting physiological measures, acute clinical outcomes, morbidity, mortality, or resource utilization were all eligible.

2.5. Secondary Surrogate Outcomes

Efficacy outcomes include improvement in echocardiographic markers, and clinical markers including ventilatory status, physiological markers such as BP, and RR and outcomes at discharge including resource utilisation. Safety parameters included adverse events due to diuretic use during hospital stay.

2.6. Echocardiographic Markers

PH markers measured by echocardiography include, (a) septal flattening measured by LV sEI, (b) RVSP estimated by TR jet when available for measurement of pressure gradient, (c) Pulmonary artery acceleration time/right ventricular ejection time (PAAT/RVET) ratio ‐ measured in the short‐axis view at the level of pulmonary valve using pulse wave doppler, (d) RV dilatation—measured by RV end diastolic diameter (RVEDD)/LV end diastolic diameter (LVEDD) ratio assessed in the short‐axis view at papillary muscle level, (e) RV systolic function‐ measured by fractional area change (FAC), tricuspid annular systolic plane excursion (TAPSE), and RV output in millilitre per kilogram per minute (mL/kg/min). Echocardiographic measures were included as non‐invasive measures of cardiac function and pulmonary hemodynamics. While the evidence supporting these measures is limited, changes in these measures may be associated with clinical outcomes such as mortality, morbidity, respiratory support [38, 45]. Reporting these markers will be helpful to explore their associations with outcomes and interpret the impact of interventions more comprehensively.

2.7. Clinical Markers

Clinical assessment includes type of respiratory support, MAP and Fio2 concentration, RR, BP, need for addition of pulmonary vasodilators along with diuretics to control PH. These parameters were included as they represent key indicators of cardio‐respiratory status and pulmonary vascular load, enabling assessment of the infant's physiological response to diuretic therapy. Outcomes at discharge includes resolution of PAH pre‐discharge identified either by echocardiography or if the neonate is off respiratory support and off diuretics it is taken as a surrogate marker of PH resolution, need for diuretics or pulmonary vasodilators at discharge, length of stay, ventilation days, and technology dependence at discharge (need for feeding tube, gastrostomy tube, ileostomy, jejunostomy or colostomy, home monitor or oxygen) [46].

2.8. Adverse Effects

Safety outcomes include the proportion of infants who developed hyponatremia ‐ sodium < 135 mEq/L, poor weight gain ‐ defined as < 15 g/day over more than 1 week during the usage of diuretics cumulatively, nephrocalcinosis, metabolic bone disease ‐ identified by serum alkaline phosphatase levels > 500 IU/L in two consecutive weekly samples [47]. These were included as safety outcomes, because they represent well recognized adverse effects of diuretic therapy and has been associated with increased morbidity, prolonged hospitalization, and potential long‐term complications [48, 49, 50].

2.9. Search Methods for Identification of Studies

We performed a comprehensive search from database inception to October 2025 across Ovid Medline (including Epub ahead of print, in‐process, and other non‐indexed citiations, Ovid Medline (R) daily, Ovid Medline (R) (1946–present), Ovid Embase (1974–October 10, 2025), and the Cochrane Central Register of Controlled Trials. Search strategies for each database are provided in Supporting Information S1: Table S3. No date or language restrictions were applied. We also searched trial registries, including Clinicaltrials. gov [51] and the World Health Organization (WHO) International Clinical Trials Registry Platform [52], and screened reference lists of relevant studies, contacting lead authors when necessary. Search terms included, but were not limited to, “diuretics,” “neonates,” “bronchopulmonary dysplasia associated pulmonary hypertension,” “chronic pulmonary hypertension,” and “clinical trials.” The search was initially conducted on October 2024, and all references were imported into EndNote 21 (Clarivate, Philadelphia, PA) and transferred to Covidence (Veritas Health Innovation, Melbourne, Australia) for screening and data extraction. The search was reconducted on October 2025 and no additional studies were identified.

2.10. Data Collection and Analysis

2.10.1. Selection of Studies

The screening process for study inclusion was conducted independently by two reviewers (UK and AS) in two stages. Stage 1 involves screening titles and abstracts to exclude irrelevant studies, followed by Stage 2, where full‐text articles are reviewed for final eligibility. To minimize bias, reviewers remain blinded to each other's decisions during the initial stages, and any discrepancies were resolved through discussion. Eligibility was determined using predefined PICO (Population, Intervention, Comparison and Outcome) criteria. Population: preterm infants born < 32 weeks’ gestation with BPD‐PH diagnosed by echocardiography at 36 weeks PMA. Intervention: diuretics (furosemide or hydrochlorothiazide and spironolactone) administered for at least three consecutive days. Comparison: placebo or no treatment. Outcomes: primary outcomes included, but were not limited to, mortality and the requirement for home respiratory support. Study types: randomized controlled trials, cohort studies, and case–control studies. Reasons for exclusion were recorded and the entire study selection process regarding inclusion or exclusion decisions was documented using Covidence software (Veritas Health Innovation, Melbourne, Australia).

2.11. Data Extraction and Management

Two reviewers (U.K. and A.S.) independently extracted data using a standardized data extraction template in excel sheet. Extracted data was cross‐checked by the second reviewer to ensure accuracy and consistency. Any discrepancies in data extraction were resolved through discussion; if persistent, they were addressed by a third independent reviewer (C.C.). Study investigators were contacted to request unreported data or additional details where critical data are missing. If missing data cannot be retrieved, these gaps were noted, and sensitivity analyses was planned to assess the potential impact on results. If quantitative analysis cannot be carried out due to missing data, then a narrative review of the results is provided. Conversion of values from median (interquartile range [IQR]) to mean (standard deviation [SD]) was carried out under the assumption of data following a normal distribution, as raw data was unavailable and using a web‐based software (Meta‐converter.com) [53].

2.12. Assessment of Risk of Bias in Included Studies

The Risk of Bias in NRSI (ROBINS‐I) tool was used to assess the risk of bias in nonrandomized studies across seven domains: bias due to confounding, selection of participants, classification of interventions, deviations from intended interventions, missing outcomes, measurement of outcomes, and selection of reported results [54]. For each domain, two reviewers (UK and AS) independently completed the signalling questions to determine domain‐level judgements and an overall risk of bias rating. Assessments were cross verified by other reviewers, and any discrepancy was resolved through discussion and consensus. Overall judgements were categorized as follows: low risk (comparable to a well‐performed randomized trial): low risk across all domains, moderate risk (sound evidence for a nonrandomized study): low or moderate risk in all domains, serious risk (important methodological problems): serious risk in one domain without critical risk in any, critical risk (too flawed to provide reliable evidence due to critical risk in at least one domain), no information (insufficient information to permute judgement in one or more domains). The risk of bias assessment was incorporated into the Grading of Recommendations, Assessment, Development and Evaluation (GRADE) framework to inform overall certainty of evidence [55].

2.13. Measures of Treatment Effect

We planned to do a meta‐analysis if sufficient studies are available to pool results. We reported continuous outcome measures as mean difference (MD) with 95% confidence intervals (CI). The results of individual studies were combined using a random‐effects model, assuming variability in the effects across studies. Treatment effects were measured using the inverse‐variance method for pooling results, which assigns weights to studies based on the precision of their estimates (inverse of the variance). Meta‐analysis was conducted using Data Party software [56]. If enough data were not available, we reported the results narratively.

2.14. Dealing with Missing Data

Study investigators were contacted to address missing data, particularly for key outcomes such as mortality and respiratory support at discharge. Sensitivity analyses excluding studies with imputed values were planned to evaluate the robustness of the pooled estimates.

2.15. Assessment of Heterogeneity

Statistical heterogeneity was assessed using the I ² statistic and χ² test (p less than 0.1 was considered significant). Forest plots were used to visualize individual study results and pooled effect estimates. The I ² statistic was interpreted with thresholds of 25%, 50%, and 75%, interpreted as low, moderate, and high heterogeneity, respectively.

2.16. Assessment of Reporting Biases

Publication bias was planned when ≥ 10 studies were included in a meta‐analysis. Funnel plots were visually inspected for asymmetry. If asymmetry was detected, potential causes were explored and the trim‐and‐fill method was applied where appropriate. Selective outcome reporting was assessed by comparing published outcomes with study protocols or trial registry records when available.

2.17. Subgroup Analysis and Investigation of Heterogeneity

Study‐level subgroups were planned to include RCTs versus NRSI to evaluate how study design influenced outcomes. Participant‐level subgroups were to examine differences based on gestational age (less than 26 weeks vs. 26–32 weeks), severity of PAH with or without RV dysfunction, concurrent use of pulmonary vasodilators (e.g., sildenafil or bosentan alongside diuretics), and the duration of diuretic therapy (short‐term ≤ 2 weeks vs. long‐term > 2 weeks).

2.18. Presentation of the Results and Assessment of Certainty of Evidence

A PRISMA flow diagram outlined the study selection process, including the number of studies identified, screened, excluded (with reasons), and included. Study characteristics, such as design, population, interventions, and outcomes, were summarized in a table. Risk of bias assessments, conducted using ROBINS‐I tool, were presented in tables. Primary outcomes, including mortality and need for home respiratory support, and secondary outcomes, such as physiological parameters ‐ PH markers (RVSP, LVs EI), clinical parameters and adverse effects, were summarized in forest plots and meta‐analysis, where applicable. Recommendation direction and strength were determined using an Evidence‐to‐Decision framework [57]. Additional factors, including values and preferences, resource implications, and feasibility, were incorporated. Recommendations were graded as strong when desirable effects clearly outweighed undesirable effects, and weak/conditional when benefits and harms were closely balanced or evidence certainty was low, in accordance with the GRADE methodology [55]. For outcomes not suitable for meta‐analysis, a narrative synthesis described trends and patterns.

3. Results

3.1. Search Results and Description of Studies

The comprehensive search was conducted from database inception till October 2025 and identified 492 records from four databases: Embase (n = 255), MEDLINE (n = 193), Cochrane Central (n = 39), and ClinicalTrials.gov (n = 5) as described in PRISMA flow diagram in Figure 1. After removing 128 duplicates, 364 records were screened. Of these, 355 were excluded during title and abstract screening. Exclusions were primarily due to wrong population (e.g., pediatric studies outside the target age group, post‐cardiac surgery patients, or maternal hypertension), wrong setting or condition (e.g., studies on diuretics in blood transfusions, BPD without PH, PDA, respiratory distress syndrome, or transient tachypnea of newborn), or inappropriate study design (e.g., pharmacodynamic studies, animal studies). Nine full‐text articles were assessed for eligibility, with seven excluded for reasons such as incorrect study population or irrelevant outcomes as mentioned in Supporting Information S1: Table S4. The final analysis included two retrospective cohort studies, comprising 49 preterm infants with BPD‐PH.

Figure 1.

PRISMA flow diagram.

3.2. Characteristics of Included Studies

The two included studies (Zhu et al. 2023; Sanchez‐Becerra et al. 2023) investigated the efficacy of diuretics in extremely preterm infants with BPD‐PH (Table 1 and Table 2). Both centres had a Targeted Neonatal Echocardiography (TNE) service model in place to support the care of neonates. Zhu et al. reported a total of 27 infants: one (4%) received invasive mechanical ventilation, 20 (75%) received non‐invasive ventilation, 5 (19%) were on LF oxygen and one (4%) infant was in room air at diuretics initiation. Baseline severity of PAH was not reported. Out of 22 screened infants by Sánchez‐ Becerra et al., at diuretic initiation, 3 (14%) were mechanically ventilated, 7 (32%) on CPAP, 2 (9%) on HFNC, and 10 (45%) on LF oxygen. Severity of PAH was not reported at baseline.

Table 1.

Characteristics of included studies.

Study	Zhu 2023 (N = 27)	Sanchez‐Becerra 2023 (N = 22)
Methods	Single‐center retrospective pre–post echocardiographic study conducted at Mount Sinai Hospital (July 2014–July 2018).	Single‐center retrospective study conducted between January 2018 and December 2021 in a tertiary care center in Mexico City.
Population	Infants at < 32 weeks’ gestation with an echocardiographic diagnosis of BPD‐PH who received diuretics and had echocardiograms within 24 h pre‐treatment and within 2 weeks post‐initiation were included. Infants receiving other concurrent PH therapies or diuretics solely for BPD were excluded.	Infants at < 32 weeks’ gestation with BPD‐PH diagnosed by echocardiogram were included.
Intervention	Initial: Furosemide ×3 days (IV 1 mg/kg or enteral 2 mg/kg every 12 hourly) Maintenance: Aldactazide 1 mg/kg/day every 12 hourly until respiratory wean and/or PH resolution Monitoring: Echo every 2–3 weeks Escalation: Cardiac catheterization and/or sildenafil if clinical deterioration	Initial therapy: Furosemide IV 1 mg/kg/dose or PO 2 mg/kg/dose every 12 hourly × 5–7 days Maintenance: Hydrochlorothiazide + spironolactone (if improved) Escalation: Sildenafil 0.5–1.0 mg/kg/dose every 8 hourly (if deterioration or persistent BPD‐PH)
Outcomes	Primary: Change in PAAT Secondary: Pulmonary hemodynamics; ventricular function; cardiorespiratory/ventilatory parameters; adverse effects (e.g., hyponatremia, nephrocalcinosis); clinical outcomes (survival to discharge, home O2, PH resolution, sildenafil use, discharge on diuretics)	Outcomes: Clinical (SBP/DBP, HR, RR, FiO2) and echocardiographic (PAAT, RVET, RV/LV function)
Baseline characteristics	Mean (SD) GA at birth 27.0 (2.8) weeks Mean (SD) birth weight 859 (294) grams	Median (IQR) 29 (27, 31) weeks
Timing of echocardiogram	Pre‐treatment @ mean PMA 36.5 ± 3.4 weeks Interval between pre‐ and post‐treatment 7.8 ± 3.0 days	Post‐treatment scan in 7 days

Table 2.

Results: Definition, indication, missing data, main results and adverse events of included studies.

Study	Zhu 2023 (N = 27)	Sanchez‐Becerra 2023 (N = 22)
Definition of BPD‐PH	Any one of 1) Flat interventricular system (IVS) in end‐systole, 2) Right ventricular systolic pressure (RVSP) ≥ 40 mmHg, 3) bidirectional or right to left through patent ductus arteriosus (PDA)	Septal flattening with RV dilatation (measured qualitatively)
Treatment indication	Screening all infants between 34‐36 weeks post menstrual age (PMA) Diuretics if symptomatic and evidence of right ventricular dilatation on echocardiogram	Median (IQR) screening at 36 (36, 38) weeks
Missing data	Data from nine patients were unavailable due to echocardiography archiving system changes.	Nil
Mortality; Home respiratory support at discharge	One death (4.5%) due to brain injury; 10 infants (37%) required home oxygen therapy	7 out of 72 screened (9.7%) infants died – all cause mortality; 56 infants (78%) of overall cohort required supplemental oxygen at discharge
Main results	Diuretic therapy showed slight improvement in PVR, RV systolic and LV global function	Diuretics treatment is associated with mild improvement in respiratory support, RVO, and markers of PVR.
Adverse events	Hyponatremia was observed in 17 infants (63%), including 15 cases classified as mild and 2 as moderate; all were managed with oral sodium supplementation, one nephrocalcinosis reported (no routine renal ultrasound)	Median serum sodium (mmol/L): 139.65 (post) vs 137.5 (pre); Renal ultrasound done in all infants – no calcium deposits reported

3.3. Risk of Bias in Included Studies

Detailed risk of bias assessments was performed using the ROBINS‐I tool, with results indicating overall serious risk of bias across domains in both studies (Table 3, Supporting Information S1: Table S5).

Table 3.

Risk of bias assessment in included studies.

Risk of bias	Zhu 2023	Sanchez‐Becerra 2023
Confounding	Serious risk	Serious risk
Selection of participants	Low risk	Low risk
Classification of interventions	Low risk	Low risk
Deviations from intended interventions	Low risk	Low risk
Missing outcome data	Moderate risk	Low risk
Measurement of the outcome	Moderate risk	Moderate risk
Selection of the reported result	Moderate risk	Low risk

3.4. Effects of Interventions and Assessment of Evidence

The included studies reported limited data on mortality, respiratory improvement, and the need for home respiratory support at discharge in preterm infants with BPD‐PH. The effects of diuretics on these outcomes and the corresponding certainty of evidence were assessed using GRADE (Table 4). The absence of a distinct comparator group precluded formal comparative analyses, necessitating a narrative synthesis of the available findings. Given that only two studies were included, subgroup and sensitivity analyses to explore heterogeneity or differences across diuretic regimens were not feasible.

Table 4.

Summary of findings: Diuretics compared to no diuretics in preterm infants with BPD‐PH.

Outcomes	Number of participants (Studies)	Certainty of evidence (GRADE)	Narrative results
Mortality at discharge	49 (2 non‐randomized studies)	⨁◯◯◯ Very lowa,b,c	The evidence is very uncertain about the effect of diuretics on mortality. In Zhu, 1 of 22 infants (4.5%) treated with diuretics died, while in Sanchez‐Becerra 7 of 72 infants (9.7%) screened at 36 weeks PMA died, with mortality not reported separately for the diuretic treated subgroup
Home respiratory support at discharge	49 (2 non‐randomized studies)	⨁◯◯◯ Very lowa,c,d	The evidence regarding the effect of diuretics on the need for home respiratory support is very uncertain. In Zhu et al., 10 of 22 infants (37%) treated with diuretics required home respiratory support. In contrast, in Sanchez‐Becerra et al., 56 of 72 infants (78%) assessed at 36 weeks’ PMA required home respiratory support; however, this outcome was not reported separately for the diuretic‐treated subgroup, limiting comparability.
Echocardiographic markers	49 (2 non‐randomized studies)	⨁◯◯◯ Very lowa,e,f	The certainty of evidence of diuretics on echocardiographic markers is very low due to serious concerns regarding risk of bias, and imprecision. Pooled estimate for PAAT showed a minimal increase (MD 5.2 milli second) (95% CI 0.15 to 10.25), small improvement in TAPSE MD 0.91 mm (95% CI 0.24 to 1.57), without any effect on RVET/PAAT, RVEDD, RVO, and RV FAC.
Physiological parameters	49 (2 non‐randomized studies)	⨁◯◯◯ Very lowa,e,f	There is very low certainty evidence of diuretics on physiological parameters. Pooled estimate showed no effect on RR, SBP or DBP.
Clinical markers	49 (2 non‐randomized studies)	⨁◯◯◯ Very lowa,c,e	The effect of diuretics on clinical markers is supported by very low‐certainty evidence. Zhu et al. 15 infants (56%) reported a reduction in respiratory support after one week of diuretic therapy, while Sanchez‐Becerra et al. observed respiratory improvement in 13 infants (59%). However, in the absence of untreated control group, it remains highly uncertain whether these observed improvements are attributable to diuretic therapy or reflect the natural course of disease over time.
Medications at discharge	49 (2 non‐randomized studies)	⨁◯◯◯ Very lowa,c,e	Evidence for diuretics on outcomes at discharge is limited with only data on medications at discharge is available. In Zhu et al, 16 infants (59%) were discharged on diuretics, Sanchez‐Becerra did not report this outcome. 3 infants (11.1%) needed escalation to sildenafil in Zhu et al as compared to 1 infant (4.5%) in Sanchez‐Becerra et al despite diuretic therapy in both studies.
Adverse outcomes	49 (2 non‐randomized studies)	⨁⨁◯◯ Lowa,c	There is very uncertain evidence for effect of diuretics on safety outcomes. Hyponatremia was reported in 17 infants (63%) in Zhu et al while Sanchez‐Becerra did not report adverse event frequency. Other adverse events were inconsistently reported.

Note: GRADE Working Group grades of evidence. High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect. a. Downgraded one level for serious risk of bias. Evidence derived from non‐randomized, single‐arm or pre–post studies without a concurrent control group; confounding could not be addressed. b. Downgraded one level for serious indirectness. Mortality was reported as all‐cause mortality, causes of death were unspecified or unrelated to BPD‐PH progression, and subgroup‐specific data for infants treated with diuretics were not separately reported. c. Downgraded two levels for very serious imprecision due to small sample sizes, very few events, and inability to generate comparative effect estimates. d. Downgraded one level for serious indirectness. Outcome reflects overall disease severity rather than a direct measure of diuretic effectiveness. e. Downgraded for indirectness where outcomes represent surrogate markers rather than patient‐important outcomes. f. Downgraded two levels for very serious imprecision due to extremely wide confidence intervals and substantial statistical uncertainty.

3.5. Primary Outcomes

Zhu et al. reported one all‐cause death among 22 infants (4.5%) treated with diuretics, attributed to severe chronic brain injury rather than progression of BPD‐PH. 16 of 22 infants (59%) were discharged on diuretics with persistent evidence of pulmonary vascular disease, and 10 infants (37%) required home respiratory support. Sanchez‐Becerra et al. reported an overall mortality rate of 9.7% (7/72 infants screened at 36 weeks’ PMA but causes of death and mortality data specific to the diuretic‐treated subgroup were not provided. 56 (78%) infants in the overall cohort were discharged on supplemental oxygen, although outcomes specific to diuretic exposure were not reported separately. Due to inconsistent reporting and the absence of comparator groups, no conclusions can be drawn regarding the effect of diuretics on mortality or the need for home respiratory support.

3.6. Secondary Outcomes

We did a meta‐analysis on the available data on echocardiographic and physiological markers (Table 4). No analysis could be done on other secondary outcomes such as clinical markers, outcomes at discharge or adverse events and hence results are provided in narration. Diuretic therapy was associated with small improvements in echocardiographic markers of PH. There is a slight improvement in PAAT with a pooled mean difference (MD) of 5.2 ms (95% CI 0.15–10.25), suggesting a slight reduction in pulmonary vascular resistance (Figure 1S). Similarly, TAPSE improved with a pooled MD of 0.91 mm (95% CI 0.24–1.57), consistent with a small beneficial effect (Figure 1S). However, the certainty of evidence for both outcomes was rated as very low, downgraded due to serious concerns regarding risk of bias and imprecision, with no factors warranting an upgrade. The pooled estimate showed no difference in other echocardiographic parameters including RVET/PAAT, RVEDD, RVO, and RV FAC or in physiological markers such as RR, SBP and DBP.

Both studies demonstrated respiratory improvement in more than half of infants after 1 week of diuretic therapy. Zhu et al. reported 15 infants (56%) had a reduction in respiratory support after 1 week of diuretic therapy, with 10 (37%) infants showing improved in respiratory severity score (RSS), 4 infants (14.8%) weaned off respiratory support, and one infant (3.7%) de‐escalated from NIPPV to CPAP, while 5 infants (19%) required escalation and 7 infants (26%) remained unchanged. Similarly, Sanchez‐Becerra et al. reported respiratory improvement in 13 infants (59%), with reductions in mechanical ventilation, CPAP, and HFNC use, although details on escalation were not provided. However, without untreated control groups, it is very uncertain whether these improvements are attributable to diuretics or reflected time‐dependent clinical improvement. Ventilation days, frequency of desaturation, length of stay, oxygen days could not be measured because of the nature of the study design.

In Zhu et al., three infants demonstrated worsening BPD‐PH and required escalation to sildenafil in addition to diuretics; among them, one infant died and the remaining two were discharged on combined sildenafil and diuretic therapy. Similarly, Sanchez‐Becerra et al. reported one infant who required initiation of sildenafil for worsening pulmonary arterial hypertension. Data on technology dependency at discharge were not reported in either study.

With respect to safety, Zhu et al. reported hyponatremia in 17 infants (63%), comprising 15 mild and 2 moderate cases, all managed with oral sodium supplementation; although routine renal ultrasound was not performed, one case of nephrocalcinosis was identified on chart review and attributed to elevated vitamin D levels. Sanchez‐Becerra et al. reported an increase in median serum sodium levels after diuretic therapy (139.65 vs. 137.5 pre‐treatment) but did not specify the number of infants who developed hyponatremia or required supplementation. Renal ultrasound performed in all infants after 1 week showed no evidence of renal parenchymal calcifications.

4. Discussion

4.1. Summary of Main Results and Certainty of Evidence

This systematic review and meta‐analysis assessed the efficacy and safety of diuretics in managing BPD‐PH in extremely preterm infants. We summarize the findings and certainty of evidence in Table 4 which were rated as low to very low.

The use of all‐cause mortality and the lack of subgroup‐specific data limit our ability to interpret outcomes specifically in the context of diuretic therapy for BPD‐PH, making it impossible to draw definitive conclusions about BPD‐PH related mortality. Both studies highlight a significant proportion of infants requiring home respiratory support at discharge, reflecting the persistent burden of BPD‐PH in this population. Similar like mortality, the results could not be compared with controls with pre‐post study design. Hence, no conclusions could be drawn regarding the role of diuretics in home respiratory support needs.

The results suggest a mild improvement in PAAT and TAPSE following diuretic therapy, indicating a slight reduction in PVR and mildly improved RV systolic function. This supports the hypothesized role of diuretics in alleviating pulmonary edema and reducing stress on the right ventricle. Although definitive evidence that improvements in RV function translate into better clinical outcomes is lacking, reduced RV function in preterm infants with BPD‐PH has been consistently associated with higher mortality [58, 59, 60]. However, the absence of significant changes in RVEDD, RV FAC, and RV output raises concerns about the broader clinical benefits of diuretics. The certainty of the evidence was downgraded to very low for most outcomes due to serious concerns regarding risk of bias, inconsistency, and imprecision. The small sample sizes and the observational design of the included studies further contributed to the uncertainty of the findings. In terms of respiratory support, 56%–59% of infants in both studies showed improvement after a week of diuretic therapy; however, details on respiratory support escalation were limited, and comparisons with controls were not possible. Although improvement was observed, it remains unclear whether this was attributable to diuretics or simply reflected time‐dependent maturation, with diuretics potentially acting as a bystander. In one study, 59% of infants were discharged on diuretic therapy. In both studies, fewer than 10% of infants required escalation to sildenafil despite ongoing diuretic treatment. However, given the limited and heterogeneous data, the effect of diuretics on the need for additional pulmonary vasodilator therapy remains highly uncertain. Hyponatremia was commonly reported in one study (63%), whereas the second study did not provide detailed reporting of adverse event frequency. Consequently, the safety profile of diuretics in this population remains uncertain. High heterogeneity across outcomes and the absence of long‐term follow‐up data limits the ability to assess the sustained efficacy and safety of diuretics.

5. Agreements and Disagreements With Other Studies or Review

This is the first review that has evaluated the efficacy and safety of diuretics in BPD‐PH in extremely preterm infants. A Cochrane review on diuretics recommended no routine use of loop diuretics in infants with BPD [61]. A recent systematic review reported that the effectiveness of diuretics in BPD is uncertain with limited number of randomised controlled trials [62]. In both the reviews, objective was to evaluate the effectiveness of diuretics in BPD. The included studies were conducted before the start of the century and include a combination of randomized, non‐blinded and observational studies. We evaluated the efficacy and safety of diuretics in infants with echo confirmed BPD‐PH. Similarly, our review did not find any robust studies or RCTs.

6. Potential Biases in the Review Process

The review included a comprehensive synthesis of available evidence on diuretics in a vulnerable population, and the inclusion of echocardiographic parameters to diagnose BPD‐PH. However, it is limited by the small number of studies with limited sample sizes, reducing statistical power. Conducting the planned meta‐analysis based on small sample sizes may have resulted in imprecise and potentially unstable effect estimates, limiting the reliability of the findings. The non‐randomized design of the studies further weakens the findings, as the effect was only measured within the same patient group before and after treatment, lacking a comparison with a control group.

6.1. Implications for Practice

There is no randomized trial data to comment on the effect of diuretics in BPD‐PH in preterm infants. Mortality, respiratory, discharge outcomes and adverse effects were reported inconsistently in the available studies. Meta‐analysis of limited data demonstrated small improvements in echocardiographic markers, but the certainty of evidence was very low due to heterogeneity between the two included studies. Current evidence supporting the use of diuretics in BPD‐PH in preterm infants remains limited and of very low certainty. High quality studies are needed to better define their safety and efficacy.

6.2. Implications for Research

This review underscores significant gaps in knowledge about the use of diuretics to treat BPD‐PH in preterm infants with BPD and the need for well‐designed RCTs to establish robust conclusions. Future studies should prioritize the development of standardized diagnostic criteria for BPD‐PH, clearly define disease severity, defining the phenotype, and comorbidities at enrollment, and specify diuretic dosing regimens and duration of therapy. The studies should focus on long‐term clinical outcomes, such as survival, neurodevelopment, and quality of life, to better understand the lasting effects of diuretics in preterm infants. It is essential to address the potential adverse effects of chronic diuretic use, including nephrocalcinosis, metabolic bone disease, and electrolyte imbalances, to ensure the safety and overall benefit of the intervention. Future studies may also consider including physiological parameters such as oxygen saturation levels, lung compliance measurements using imaging such as x‐ray vs lung ultrasound, use of additional vasodilator use, cost effectiveness of TNE/lung ultrasound screening and monitoring.

Author Contributions

Uthaya Kumaran Kanagaraj and Ayman Saker conceptualized the study. Uthaya Kumaran Kanagaraj, Ayman Saker, and Carlos Cuello‐Garcia developed the methodology and validated the data. Uthaya Kumaran Kanagaraj and Ayman Saker conducted the investigation and curated the data. Formal analysis was performed by Uthaya Kumaran Kanagaraj, Ayman Saker, and Carlos Cuello‐Garcia. The original draft was prepared by Uthaya Kumaran Kanagaraj and Ayman Saker. Reviewing and editing of the manuscript were performed by all authors (Uthaya Kumaran Kanagaraj, Ayman Saker, Carlos Cuello‐Garcia, Michael Castaldo, and Sandesh Shivananda). Visualization and project administration were carried out by Uthaya Kumaran Kanagaraj and Ayman Saker. All authors have read and approved the final version of the manuscript.

Funding

The authors have nothing to report.

Ethics Statement

Ethical review and approval were waived because this study involved only the secondary analysis of previously published, aggregated, and non‐identifiable data. No primary data collection, patient contact, or access to identifiable health records occurred. Under our institutional policy and standard research ethics guidance, systematic reviews and meta‐analyses that use publicly available, de‐identified sources do not require Research Ethics Board review or participant consent.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.