Furosemide for patent ductus arteriosus during cyclooxygenase inhibitor therapy: A systematic review

Abstract

Background

Although furosemide is used during cyclooxygenase (COX) inhibitor therapy for patent ductus arteriosus (PDA), there are concerns regarding increased ductal closure failure and acute renal failure (ARF). This systematic review explores the effects of furosemide during COX inhibitor therapy.

Methods

We searched MEDLINE, Embase, Cumulative Index to Nursing and Allied Health Literature, Cochrane Central Register of Controlled Trials, and Igaku Chuo Zasshi databases for randomized clinical trials that assessed furosemide during COX inhibitor therapy for PDA in preterm infants. The primary outcome measure was PDA closure failure. Mortality and other complications were also assessed. The risk of bias was assessed using the Cochrane risk‐of‐bias tool for randomized control trials, and the certainty of evidence was assessed using the Grading of Recommendations Assessment, Development, and Evaluation criteria.

Results

Overall, three trials involving 121 patients were included in the analysis. The overall incidence of PDA closure failure was 28%. Although the result of PDA closure failure, mortality, and ARF were obtained, other outcomes were not described in any of the studies. The risk of bias was high. The risk of PDA closure failure did not increase with furosemide administration. Furosemide was not associated with decreased mortality but was associated with an increased risk of ARF (risk ratio, 4.96 [95% confidence interval: 1.80–13.6]). The certainty of evidence for all outcomes was very low.

Conclusion

Although furosemide is not associated with an increased risk of PDA closure failure or mortality, the risk of ARF increases after furosemide administration during COX inhibitor therapy.

INTRODUCTION

Patent ductus arteriosus (PDA) is a clinical consequence and a cardiovascular comorbidity that causes left‐to‐right shunting in preterm infants. ¹ , ² Increased pulmonary blood flow and decreased systemic flow due to ductal steal by PDA is associated with mortality and severe complications, such as bronchopulmonary dysplasia (BPD), necrotizing enterocolitis (NEC), and pulmonary hemorrhage. ¹ , ² , ³ , ⁴ Approximately 20 to 30% of infants under 30 weeks of gestation receive pharmacotherapy with cyclooxygenase (COX) inhibitors (e.g., ibuprofen and indomethacin) in an attempt to prevent complications of PDA. ² , ⁵ , ⁶ Renal and electrolyte abnormalities are the most common side effects of COX inhibitor therapy as COX inhibitors inhibit prostaglandin synthesis. ⁷ Furosemide increases renal prostaglandin synthesis and can be administered to reduce COX inhibitor‐induced nephrotoxicity; ⁸ , ⁹ however, owing to the increase in renal prostaglandin synthesis, there is concern about the increasing failure of ductal closure in preterm infants treated with furosemide. ¹⁰ , ¹¹

Previous randomized controlled trials (RCTs) have reported inconsistent results regarding the effect of furosemide administration during COX inhibitor therapy in preterm infants. Although the first RCT reported the efficacy of furosemide on the renal side effects of indomethacin therapy, ¹² furosemide failed to reduce the renal side effects of indomethacin in a subsequent RCT. ¹³ Another RCT that assessed the effects of furosemide administration during COX inhibitor therapy revealed an increased incidence of acute renal failure (ARF) with furosemide use during indomethacin therapy. ¹⁴ Although a previous systematic review did not find any association between ductal closure failure and furosemide, furosemide increased creatinine clearance in patients with a low blood urea nitrogen‐to‐creatinine ratio. ¹⁵

In addition to these inconsistent results, the last systematic review was conducted over 20 years ago and has not been updated since. ¹⁵ There is little evidence regarding outcomes relevant to patients (i.e., ARF and neurodevelopmental impairment) and the differences in the effects of different types of COX inhibitors (indomethacin or ibuprofen). We therefore aimed to explore the effects of furosemide during COX inhibitor therapy for PDA on neonatal outcomes, such as failure of PDA closure and renal function.

MATERIALS AND METHODS

Review question

This systematic review was designed to address the research question “In preterm infants with PDA who receive COX inhibitors (Population), does management with furosemide (Intervention), compared with management without furosemide (Comparison), decrease or increase morbidity and mortality (Outcome)?” This systematic review was conducted according to the Cochrane Handbook for Systematic Reviews of Interventions and was reported based on the Preferred Reporting Items for Systematic Reviews and Meta‐Analyses (PRISMA) statement. ¹⁶ The protocol for this systematic review was registered in the International Prospective Register of Systematic Reviews (CRD42023492719). Ethical approval was not required given that this study was based on previously published research.

Data sources

We searched MEDLINE (Ovid interface), Embase, the Cochrane Central Register of Controlled Trials (CENTRAL), the Cumulative Index to Nursing and Allied Health Literature (CINAHL), and Igaku Chuo Zasshi (Ichushi, a database of Japanese medical articles) databases from their inception date until November 16, 2023. Trial registrations on ClinicalTrials.gov, the EU Clinical Trials Register, and the Australian New Zealand Clinical Trials Registry were also searched to identify ongoing trials. The search strategies are listed in Supporting Information, Tables S1–S4. The search strategy for Ichushi was constructed in Japanese.

Selection criteria

We included RCTs that compared the outcomes of furosemide with those of controls during COX inhibitor therapy in preterm infants with PDA. Non‐RCTs, interrupted time‐series analyses, cohort studies, case–control studies, before‐and‐after studies, case series analyses, case reports, studies with proceedings only, and animal studies were excluded. Only studies published in English or Japanese were analyzed.

Data extraction

Two authors (HK and YT) screened the titles and abstracts of the articles independently, and disagreements were resolved through discussion between the two authors. The two authors then screened the full‐text articles, and any discrepancies were again resolved through discussion. If the discussion was unsuccessful, the final decision was made by a third person (SM). The reviewers independently collected data on the study design, population, intervention, and outcome from all eligible articles.

Data analysis

The collected data were analyzed using Review Manager 5.4 software (Cochrane Collaboration, Oxford, UK). The primary outcome measure was PDA closure failure. The secondary outcome measures included mortality, ARF, BPD, hyperkalemia, NEC, neurodevelopmental impairment, periventricular leukomalacia, pulmonary hemorrhage, and severe intracranial hemorrhage. Patent ductus arteriosus closure failure was defined as a PDA that remained open after a course of COX inhibitor treatment or as per the authors' definition in their studies. Acute renal failure was defined by the authors in their studies. Only one RCT was used to assess ARF, and in that RCT ARF was defined as an increase in serum creatinine of >1.6 mg/dL after the first indomethacin course during the first 2 weeks after birth. ¹⁴ The detailed definitions of these outcomes are provided in Supporting Information, Table S5. To investigate the effect of furosemide administration on the renal side effects of COX inhibitors, hyperkalemia was set as an outcome measure, in accordance with a previous systematic review. ¹⁵

We reported the effect estimates for binary outcomes as risk ratios (RRs) with 95% confidence intervals (CIs) and for continuous outcomes as mean differences with 95% CIs. Heterogeneity was assessed by visual inspection of the forest plot using the χ ² test (significance level: p < 0.10) and the I ² statistic (significant heterogeneity: I ² > 60%). Meta‐analyses were performed using the random‐effects model with the Mantel–Haenszel method because we anticipated wide variations between studies. Differences with p < 0.05 were considered statistically significant. Subgroup analyses were performed according to the types of COX inhibitors administered (indomethacin or ibuprofen) and gestational age at birth (<28 weeks or ≥ 28 weeks).

Assessments of risk of bias and evidence certainty

The risk of bias for the RCTs was assessed independently using the Cochrane risk‐of‐bias tool (RoB2). The risk of bias for each domain and the overall risk of bias were determined. The risk of bias for each domain was assessed as low, high, or concerning. Meanwhile, the certainty of evidence for each outcome was evaluated based on the risk of bias, inconsistency, imprecision, indirectness, and publication bias following the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) criteria. ¹⁷

RESULTS

A total of 118 articles were initially identified, and 27 duplicates were removed (Figure 1). The subsequent title and abstract screening included six articles. No additional eligible articles were identified during a manual search of the retrieved articles. In the full‐text screening, three articles (n = 1, language other than English and Japanese; n = 1, out of scope; n = 1, study protocol) ¹⁸ , ¹⁹ were excluded. Finally, three articles ¹² , ¹³ , ¹⁴ were included in the quantitative synthesis. All trials included in this review compared the outcomes of furosemide versus those of control during indomethacin therapy in preterm infants (Table 1). Although furosemide was administered at 1.0 mg/kg/dose intravenously following each dose of indomethacin in all trials, different doses of indomethacin were administered in each trial. Furosemide was administered simultaneously with indomethacin in all RCTs. The interval between furosemide doses varied according to the timing of indomethacin administration in each RCT (Table 1). Moreover, furosemide was administered in an open‐label fashion in two RCTs, ¹² , ¹³ and none of the patients in the control group received furosemide. In one RCT, ¹⁴ furosemide was administered in a blinded fashion, and five patients in the control group received furosemide against the protocol.

FIGURE 1.

Flow diagram of the search result and study selection.

TABLE 1.

Characteristics of the studies.

Study ID	Country	Study design	Furosemide, n	Control, n	GA and BW	Interventions	Comparisons
Yeh et al. 12	USA	Single‐center RCT	10	9	<37 weeks of GA	Furosemide 1.0 mg/kg i.v. following each dose of indomethacin (0.30 mg/kg/dose)	Indomethacin alone
Romagnoli et al. 13	Italy	Single‐center RCT	16	18	<37 weeks of GA and BW <1750 g	Furosemide 1.0 mg/kg i.v. following each dose of indomethacin (0.20 mg/kg/dose three times every 12 h)	Indomethacin alone
Lee et al. 14	South Korea	Single‐center RCT	35	33	<34 weeks of GA and BW <2000 g	Furosemide 1.0 mg/kg i.v. following each dose of indomethacin (patients aged <48 h received 0.2, 0.1, and 0.1 mg/kg; patients aged >48 h received 0.2, 0.2, and 0.2 mg/kg)	Indomethacin and placebo

Abbreviations: BW, bodyweight; GA, gestational age; i.v., intravenous; RCT, randomized controlled trial.

Data on PDA closure failure were available from all included articles, ¹² , ¹³ , ¹⁴ and mortality and ARF outcomes were described in one article. ¹⁴ We performed a meta‐analysis of these three outcomes, and although one study reported no significant differences between furosemide use during COX inhibitor therapy and BPD or NEC, this was difficult to assess in our review because the data were not described. ¹⁴ All of the included studies evaluated furosemide administration during indomethacin therapy, and none of the studies used ibuprofen; we therefore did not perform a subgroup analysis based on the type of COX inhibitors used. Given that no study conducted a subgroup analysis with a gestational age of <28 weeks and ≥ 28 weeks, we changed the categorization post hoc into <29 weeks and ≥ 29 weeks of gestation, following the categorization in one study. ¹⁴

The risk of bias in the three studies for each outcome is shown in Figure 2. One study ¹² used an envelope for randomization, and another study ¹³ did not describe the detailed randomization process. Regarding missing outcomes, we excluded 16/68 (24%) patients from the study by Lee et al. ¹⁴ and one patient from the study by Yeh et al. ¹² that evaluated PDA closure using cardiac murmur. The registered trial protocols were not shown in the manuscripts for any of the studies. Overall, the outcomes of all studies were considered to have high risk of bias for this systematic review.

FIGURE 2.

Summary of risk of bias. PDA, patent ductus arteriosus.

The results of the meta‐analysis are shown in Figure 3 and summarized in Table 2. In the systematic review, 121 patients from the three RCTs were included. The number of patients for each outcome analysis differed owing to missing outcome data. Regarding PDA closure failure, no significant difference was observed between furosemide and the controls during COX inhibitor therapy (RR: 1.54 [95% CI: 0.80–2.98]; three studies on 116 infants; very low certainty nor statistical heterogeneity [I ²  = 11%]). Given that the outcomes of mortality and ARF were reported in only one RCT, ¹⁴ we only describe the raw result. Furosemide was not associated with decreased mortality (RR: 0.71 [95% CI: 0.17–2.92]; very low certainty), although only one study was included. Meanwhile, furosemide was associated significantly with an increased incidence of ARF (RR: 4.96 [95% CI: 1.80–13.62]; very low certainty). In the subgroup analysis by gestational age, furosemide was associated with an increased incidence of ARF in infants born at <29 weeks of gestation (RR: 4.67 [95% CI: 1.58–13.81]).

FIGURE 3.

Forest plots of the meta‐analysis. (a) Closure failure of patent ductus arteriosus. (b) Mortality. (c) Acute renal failure. CI, confidence interval; M‐H, Mantel–Haenszel.

TABLE 2.

Summary of study findings: Furosemide versus placebo during cyclooxygenase inhibitor therapy for preterm infants with patent ductus arteriosus.

Outcomes	Anticipated absolute effects (95% CI)	Relative effect (95% CI)	Participants, n (studies)	Certainty of the evidence (GRADE)
Patent ductus arteriosus closure failure	116 per 1000 (−43–424)	RR 1.54 (0.80–2.98)	116 (3 RCTs)	⨁◯◯◯ Very low a , b
Mortality	−35 per 1000 (−101–233)	RR 0.71 (0.17–2.92)	68 (1 RCT)	⨁◯◯◯ Very low a , b
Acute renal failure	410 per 1000 (83–1000)	RR 4.96 (1.80–13.6)	63 (1 RCT)	⨁◯◯◯ Very low a , c

Abbreviations: CI, confidence interval; GRADE, Grading of Recommendations, Assessment, Development, and Evaluation; RCT, randomized controlled trial; RR, risk ratio.

^a High risk of bias in the studies that were included (−1).

^b Wide CI crossing line with no effect (−2).

^c The ratio of the upper to the lower boundary of the CI is more than 3 (−2).

DISCUSSION

This systematic review analyzed three RCTs involving 121 infants and found that furosemide administration during COX inhibitor therapy was not associated with PDA closure failure or mortality. Acute renal failure occurred more frequently in patients who received furosemide than in those who did not. The quality of the evidence, as assessed using the GRADE framework, was low or very low owing to the limited number of studies, small sample size, and high risk of bias. Additionally, we did not consider furosemide administration during ibuprofen therapy because all the included studies only analyzed the effects of furosemide during indomethacin therapy.

The current study shows that furosemide administration during COX inhibitor therapy did not increase the risk for PDA closure failure or mortality, in agreement with the results of a previous systematic review, although our study includes only one new meta‐analysis. ¹⁵ Notably, there are concerns about furosemide weakening the effect of COX inhibitors for ductal closure because furosemide accelerates renal prostaglandin synthesis. ¹⁰ , ¹¹ Although a previous animal study reported that furosemide affected ductus arteriosus dilation, ²⁰ the present study shows that furosemide has minimal effect on PDA closure. A retrospective cohort study involving 43,576 patients also reported that furosemide exposure was not associated with an increased risk of PDA treatment. ²¹ However, analysis of this outcome was limited to 116 patients; prospective studies with large sample sizes are therefore required to confirm the generalizability of these results.

Our study revealed that furosemide use during COX inhibitor therapy was significantly associated with an increased risk of ARF, although the certainty of the evidence was very low. Acute renal failure in the study that was included was defined as increased serum creatinine >1.6 mg/dL after the first indomethacin administration. ¹⁴ A previous systematic review evaluated changes in clinical data, such as urine output, serum creatinine, and blood urea nitrogen. ¹⁵ We selected ARF as one of the study outcomes because it is more relevant than changes in clinical data for patients with PDA. Although furosemide was expected to reduce the renal side effects of COX inhibitors, furosemide administration during COX inhibitor therapy may increase the risk of ARF without any patient benefit. However, the included studies did not set long‐term renal failure as an outcome and these studies reported that renal function improved rapidly. ¹² , ¹⁴ Although no RCT has examined long‐term renal dysfunction with furosemide, renal failure by furosemide administration during COX inhibitor therapy may be transient. From this perspective, furosemide administration during COX inhibitor therapy should not be categorically dismissed. Clinicians should be aware of the risk of ARF associated with furosemide use during COX inhibitor therapy, and larger RCTs are warranted because the trials included in this study had a high risk of bias.

In the subgroup analysis by gestational age, premature infants born at <29 weeks of gestation had a higher risk of ARF with furosemide use during COX inhibitor therapy. Extremely preterm infants are at high risk of developing ARF, and preterm infants who experience ARF have a higher risk of developing chronic kidney disease than those who do not. ²² , ²³ The current results suggest that the routine use of furosemide during COX inhibitor therapy is not recommended, especially in extremely preterm infants. However, this study included old studies and patient backgrounds (e.g., gestational age) and thresholds for indomethacin and furosemide administration may have differed from current neonatal intensive care unit practice. Moreover, furosemide was initiated within 24 h of birth in 80% of the patients in the furosemide group because furosemide was concurrently administered with indomethacin. ¹⁴ However, in the clinical setting, furosemide is usually administered when patients experience congestion and heart failure after COX inhibitor administration. This is because congestion and heart failure due to PDA typically develop 24 h after birth in most PDA patients. ⁹ The method of furosemide administration in this study was different from that in real‐world clinical practice, although the study accounts for a large proportion of the current meta‐analysis. ¹⁴ Thus, it remains uncertain whether furosemide increases the risk of ARF in real‐world settings that aim to reduce congestion with appropriate hemodynamic conditions.

Ibuprofen is typically used as an alternative to indomethacin for PDA treatment as it exhibits equal efficacy and few adverse effects. ²⁴ , ²⁵ We investigated the effects of furosemide during ibuprofen therapy because the previous systematic review did not evaluate these effects. ¹⁵ Although we expected to find new clinical trials investigating furosemide during ibuprofen therapy, we definitively confirm that no currently published studies have investigated these effects. Thus, further research on this topic is warranted.

This study had some limitations. First, we included only three studies, and information on mortality and ARF could only be obtained from only one RCT. ¹⁴ Moreover, one RCT accounted for a large proportion of the results of this study because this RCT included a large number of patients. ¹⁴ However, this RCT was a new trial that was not included in a previous systematic review on this topic. ¹⁴ , ¹⁵ Although RCTs with larger sample sizes are needed to confirm our findings, the results of this study are important as they highlight the paucity of existing evidence. Second, we excluded articles written in languages other than English or Japanese. The trials that were excluded due to this restriction had a small sample size and had results that were similar to those of other included studies. ¹⁵ , ¹⁸ We therefore considered that our results were not influenced by the exclusion of these studies. Third, we were unable to analyze data on furosemide use during ibuprofen therapy. However, this systematic review revealed a clear knowledge gap in the use of furosemide. Regardless of these limitations, this study is informative in that it updates knowledge on furosemide use during COX inhibitor therapy through a systematic review.

CONCLUSIONS

Although the outcomes assessed in this study had a very low certainty, the results showed that furosemide was not significantly associated with an increased risk of PDA closure failure or mortality. However, the risk of ARF was increased significantly following furosemide administration during COX inhibitor therapy.

AUTHOR CONTRIBUTIONS

Hiroki Kitaoka, Yusuke Terada, and Kosuke Tanaka conceptualized and designed the study, designed the data collection instruments, collected the data, conducted the initial analyses, and drafted and revised the manuscript. Masatoshi Nozaki and Satoshi Masutani conceptualized and designed the study and critically revised the manuscript. Tetsuya Isayama and Katsuaki Toyoshima conceptualized and designed the study and critically reviewed and revised the manuscript for important intellectual content. All authors have approved the final manuscript as submitted and agreed to be accountable for all aspects of this study.

CONFLICT OF INTEREST STATEMENT

Kosuke Tanaka is on the Committee for Standardization of Medicine of the Japan Society for Neonatal Health and Development, and received payment from Medicus Shuoppan, Publishers Co., Ltd. Satoshi Masutani is a board member of the Japanese Society of Pediatric Cardiology and Cardiac Surgery. Tetsuya Isayama is on the Board of Directors of the Japan Society for Neonatal Health and Development, is Chair of Japan Evidence‐Based Neonatology of the Committee of Medical Standardization of the Japan Society for Neonatal Health and Development, is on the Board of Directors of the Neonatal Research Network of Japan, and is on the Board of Directors of the Japan Society of Perinatal and Neonatal Medicine. Katsuaki Toyoshima is a member of Japan Evidence‐Based Neonatology. There are no other conflicts of interest to disclose.

Supporting information

Data S1.

Kitaoka H, Terada Y, Tanaka K, Nozaki M, Masutani S, Isayama T, et al. Furosemide for patent ductus arteriosus during cyclooxygenase inhibitor therapy: A systematic review. Pediatr Int. 2024;66:e15822. 10.1111/ped.15822