Patent ductus arteriosus (PDA), one of the most common cardiac conditions affecting preterm infants, is intimately associated with congestive heart failure, intraventricular hemorrhage (IVH), necrotizing enterocolitis (NEC), bronchopulmonary dysplasia (BPD) and death [1, 2, 3, 4]. However, due to the relatively high incidence of spontaneous closure of PDA in neonates and the controversies on the benefit-to-risk ratio of treatment, there is still no consensus on which PDAs to treat, when to treat, and how to treat [5, 6, 7, 8, 9]. The mainstay of therapy for preterm infants with hemodynamically significant PDA (hsPDA) is thought to be cyclooxygenase inhibitors (COXI), such as ibuprofen, indomethacin and paracetamol. COXI promotes ductal closure by inhibiting cyclooxygenase enzymes, thereby reducing the synthesis of prostaglandins that play a crucial role in maintaining the patency of the ductus. Compared with indomethacin, ibuprofen was reported to have similar efficacy but fewer side effects [5, 10, 11, 12, 13]. Recent meta-analyses found that paracetamol caused fewer adverse effects without significant differences in the PDA closure rates between ibuprofen and indomethacin [14, 15]. Nonetheless, the issue of whether dual-drug combinations may improve treatment results remains unclear, and further study in this area is required [16, 17].

Since Stevenson et al. [18] first reported that excessive fluid administration to preterm infants with respiratory distress syndrome (RDS) may be a predictor associated with the development of PDA complicating RDS, a growing body of research has studied the association between fluid intake and PDA, yielding conflicting results. Subsequent observational studies [19, 20, 21] and Bell’s [22] randomized controlled trial echoed the results from Stevenson’s research, suggesting that premature infants in the high fluid intake group after birth have an elevated risk of developing PDA, while some other randomized controlled trials came to a different conclusion [23, 24]. To address this dilemma, Bell et al. [25] conducted a Cochrane review which concluded that water restriction under the premise of avoiding substantial dehydration reduced the incidence of PDA without significantly raising the risk of adverse events. Based on this evidence, restricted fluid intake has been proposed to be one of the suggested non-pharmacological interventions for the management of PDA [5].

Considering the widely-acknowledged value of fluid restriction in the management of PDA, it is plausible to assume that fluid restriction probably enhances the response rate to COXI. There have been a few studies in an attempt to validate this inference, the majority of which observed no significant difference in fluid intake volume between responders and non-responders to COXI [3, 26, 27, 28, 29]. Only Ahamed et al. [30] provided evidence that fluid intake before indomethacin treatment would affect the incidence of PDA closure. However, there was considerable heterogeneity in these studies in terms of drug selection and the definition of mean fluid intake volume. In addition, these studies were restricted to the association between fluid intake volume and treatment outcome, but the crucial role that fluid balance played was not addressed.

Currently, there are no known studies evaluating whether fluid balance has a link with ibuprofen treatment and outcomes in PDA patients. Therefore this study was undertaken to determine the correlation between fluid balance and outcomes following treatment with ibuprofen for PDA in preterm infants.

In this study, we identified a decreased risk of FIT failure among infants with lower fluid before FIT, especially within the first week after birth. We also observed that gestational diabetes mellitus, higher Apgar scores, and higher neutrophil counts reduced the possibility of successfully responding to FIT. However, the statistically significant association between higher neutrophil counts and non-response to FIT was no longer present after multivariable regression analysis was performed. There was no independently significant correlation between any of the variables and the outcome of SIT and fluid balance after the first week, nor was it significantly related to FIT outcome.

Our study found it was fluid balance, not fluid intake before FIT, which was significantly associated with the FIT outcome. The overall fluid balance, which is simultaneously affected by intake volume and output volume, is a more integrated factor than intake volume and more precisely depicts the overall fluid load. In their randomized controlled trial, Bell et al. [22] proposed that the association between PDA development and fluid overload was probably attributed to water retention and a disturbance in hemodynamic balance when liberal amounts of fluid are given, resulting in larger systemic-to-pulmonary shunting through the ductus arteriosus. However, De Buyst et al. [34] concluded that fluid restriction did not reduce pulmonary circulation overload in their prospective observational study. Additionally, this association is also postulated to be triggered by an increase in prostaglandin E2 levels resulting from excessive fluid intake [35, 36, 37].

In the initial analysis, we noticed that only the fluid balance before FIT had a significant association with FIT outcome, while there was no statistically significant association between treatment outcome and fluid balance before and during SIT, and during the FIT. Potential explanations were the small sample size of infants treated with SIT and the short duration of the treatment period, only three days, which might not be sufficient to have a strong impact on treatment outcomes. In addition, given that the immediate postnatal days are an important period of physiological adaptation for the newborns, with significant hemodynamic changes and a greater susceptibility to hemodynamic instability [38, 39], the inconsistent results could possibly be also interpreted by the hypothesis that only fluid balance within a limited period of time after birth had a significant effect on the outcome of FIT, and the longer the time after birth, the less impact fluid balance had on FIT outcome. Accordingly, we did further investigations on the relationship between fluid balance within the first week of life, fluid balance from day 8 after birth to start of FIT and FIT outcome, respectively. The results of multivariate regressions were compatible with our assumption and showed that only fluid balance within the first week had a significant association with the outcome of FIT (p = 0.027) and was a more robust predictor for the outcome of FIT than the overall fluid balance before FIT (p = 0.042).

From the results of our study and the conclusions of the previous literature [18, 19, 20, 21, 22, 25, 40, 41], fluid restriction is extremely important for infants after birth. However, sometimes fluid restriction has negative repercussions such as nutritional deficiency, dehydration, hypotension, and impaired endorgan perfusion [42, 43]. According to the Canadian Pediatric Society, “aggressive” fluid restriction (120 mL/kg/day) may be potentially hazardous in preterm newborns with hsPDA [44]. Diuretic medications are another option to restrict fluid balance. But a neonatal rodent experimental model demonstrated that furosemide might delay ductal closure and might also lead to dilation of the constricting PDA. Furosemide could promote renal production of prostaglandin E2, a potent ductus arteriosus dilator. Premature infants may be at an increased risk for suffering from electrolyte disturbances, thrombocytopenia, acute kidney injury, nephrocalcinosis, ototoxic sensorineural hearing loss, and direct hyperbilirubinemia due to the use of furosemide [45, 46]. Because of these serious side effects, furosemide is not generally accepted as a standard part of conservative management of the hsPDA [47, 48]. Some pediatric cardiology and neonatology groups have begun recommending chlorothiazide as a first-line for diuretic in hsPDA [49]. Since restricting fluid intake and the use of diuretics pose a higher risk of complications for infants while limiting fluid balance, it is recommended to adopt a judicious and personalized method of fluid restriction while meeting the physiological needs of premature infants [43]. Optimal fluid therapy should be titrated to meet normal physiologic and caloric requirements and regulated to achieve a balance between the pulmonary and systemic blood flow [46].

Our study demonstrated that gestational diabetes mellitus and higher Apgar scores were independent risk factors preventing FIT from closing hsPDA. However, we did not find any study showing that gestational diabetes mellitus has an influence on the closure of hsPDA treated with COXI. A recent international cohort study including 78,126 infants also indicated that there was no significant association between gestational diabetes mellitus and hsPDA after adjusting for confounders [50]. Moreover, the independently significant association between higher Apgar scores and a lower rate of successful FIT was contrary to previous studies showing that newborn infants with lower Apgar scores had a higher risk of adverse outcomes [51, 52]. This may be explained by the small sample size and confounding factors not clearly identified in our study.

Through a linear analysis between weight changes and fluid balance, we found that the accuracy of the measured fluid balance in our study was not highly satisfactory. One potential reason may be that we only calculated fluid balance after admission to the NICU but did not take into account fluid balance prior to NICU admission. We also did not analyze insensible water loss and stool when calculating the fluid output volume. Furthermore, examining the accuracy of fluid balance measurements by testing the correlation between weight changes and fluid balance may not be accurate, as changes in infant weight are also affected by other factors, such as protein and calorie intake [53] and related complications, especially severe pulmonary disease [54].

Aside from the accuracy of the measured fluid balance, our study also has other limitations. First, this was a retrospective study and thus was subject to confounding bias with conclusions implying association and not causation. Second, our study had a relatively small sample size of 110 and 27 infants in FIT and SIT, respectively, which limited the power of analyses. Third, this was a single-center study and the generalizability of our findings may not pertain to other centers. Fourth, our unexplained results may be due to the existence of undetected confounders.

Our data demonstrated that a lower fluid balance volume before FIT, especially within the first week after birth had a significant association with favorable FIT outcome, which might potentially be a therapeutic target to minimize the FIT failure rate. Our study also established the basis for conducting a large randomized controlled clinical trial to examine the effects of restricted fluid administration strategies on the outcome of ibuprofen for hsPDA in preterm infants.

PDA, patent ductus arteriosus; hsPDA, hemodynamically significant patent ductus arteriosus; FIT, the first course of ibuprofen treatment; SIT, the second course of ibuprofen treatment; IVH, intraventricular hemorrhage; NEC, necrotizing enterocolitis; BPD, bronchopulmonary dysplasia; COXI, cyclooxygenase inhibitors; RDS, respiratory distress syndrome; NICU, neonatal intensive care unit; GA, gestational age; BW, birth weight; SGA, small for gestational age; PROM, premature rupture of membranes; ROC, receiver-operating characteristic; AUC, area under the curve; CI, confidence interval; OR, odds ratio.

The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.

CL and YS designed the research study. CL performed the research, analyzed the data and wrote the manuscript. Both authors contributed to editorial changes in the manuscript. Both authors read and approved the final manuscript.

The project was approved by Ethics committee of Children’s Hospital of Chongqing Medical University (2022-182). The need for informed consents was waived due to the retrospective nature of the study.

We thank Yuni Zhang and Dinggang Li for submission of ethical review materials. We thank Long Chen, Kaizhen Liu, and Yin Liu for their help in design of the research and reply to reviewers’ comments.

This work was supported by National Key Research and Development Program of China (2022YFC2704803).

The authors declare no conflict of interest.