Anip Garg·Naveen Athiraman·Elizabeth Jenkins·David Crossland·Prakash Kannan Loganathan

Abstract Background To compare short and long outcomes between early (≤ 28 days) (EL) and late ligation (LL) groups.To explore factors predicting early extubation (≤ 7 days) after patent ductus arteriosus (PDA) ligation.Methods We conducted a single center,retrospective cohort study of preterm infants <32 weeks who underwent surgical ligation over a 10-year period (2009–2019).Results A total of 133 infants underwent PDA ligation,in the study period.Both groups had similar short-term outcome such as bronchopulmonary dysplasia (BPD) or death (96% vs.98%,P =0.64) and long-term clinical outcomes including Bayley’s assessment at 2 years corrected age.Fewer infants in the EL group developed severe BPD (63% vs.81%,P =0.02).Age at ligation had adjusted odds ratio of 1.04 with over lapping confidence interval (95% CI 1.0–1.1,P =0.02) for severe BPD/death.There was no difference in day of extubation between the EL and LL group (8 days vs.7 days,P =0.85).Left atrium/aortic root ratio of ≥ 1.75 would give sensitivity of 41% and 80% specificity for early extubation (area under the curve of 0.61).There was marginal reduction of hospital stay in the EL group [113 (105–121) days vs.115 (107–123) days;log rank P =0.026].Conclusion EL can be delivered safely with a clinically important lower incidence of severe BPD and shorter duration of hospital stay compared to LL.

Keywords Bronchopulmonary dysplasia·Extubation·Patent ductus arteriosus ligation

Introduction

Persistent patent ductus arteriosus (PDA) has been associated with increased mortality and morbidity.However,the mode and timing of treatment remain a topic of debate[1–3].Medical treatment is often preferred;however,surgical ligation may be the only option in some infants with a haemodynamically significant duct when medical treatment has failed or is contraindicated.Though a newer modality of treatment,transcatheter PDA closure,has emerged as a new alternative to surgical ligation with significantly lower complication rates [4],most centers in the United Kingdom(UK) are still using the traditional surgical approach [5].Previous observational studies have reported both adverse neonatal outcomes [6–8] and favourable outcomes with PDA ligation [9].The number of infants referred for surgical PDA closure varies widely between units and changing over time[10,11].Persistent requirement of significant respiratory support is one of the common reasons for neonatologists to refer for surgical ligation [12].

Our study,therefore,had two objectives:(1) to compare short term hospital and long-term outcomes at 2 years among preterm neonates who underwent early ligation during the neonatal period (≤ 28 days) as compared with late ligation (>28 days);(2) to identify whether early PDA ligation results in early extubation and the factors predicting early extubation (≤ 7 days) post PDA ligation.

Methods

Study participants and setting

In this retrospective cohort study,we included all preterm infants <32 weeks who underwent surgical ligation in a single center surgical level 3 neonatal unit in northeast England over a 10-year period (2009–2019).This cohort also includes out born infants referred for PDA ligation and received most of the intensive care in the surgical center.Infants with major complex congenital heart disease and underwent device closure were excluded from the study.We did not exclude infants with minor congenital abnormalities,such as ventricular septal defect and inter-atrial shunts.All the PDA ligation surgeries were at the regional cardiothoracic center with the surgery lead by a consultant pediatric cardiothoracic surgeon.A decision regarding the need for PDA ligation was taken by the treating neonatal team with confirmatory echocardiography by a cardiologist prior to PDA ligation.

Study exposure and outcome

“Early PDA ligation” was defined as PDA ligation performed on/before ≤ 28 days.A course of antenatal steroids was considered complete when two doses of antenatal steroids were given at least 24 hours prior to delivery.Data collected included patient demographics,day of life at ligation,respiratory support before ligation,day of life at the time of extubation.Collected neonatal morbidities included:necrotising enterocolitis (NEC) requiring surgery (Bell stage >2),retinopathy of prematurity requiring treatment,significant intraventricular haemorrhage (ventricular haemorrhage with dilatation or intraparenchymal haemorrhage),blood culture positive late onset sepsis (>3 days of life),bronchopulmonary dysplasia (BPD) was defined as need for supplemental oxygen or positive pressure respiratory support at 36 weeks (National Heart,Lung and Blood Institutesponsored workshop 2001 definition [13]),requirement of home oxygen therapy,duration of hospital stay,and death before discharge.Severe BPD was defined as need for respiratory support (continuous positive airway pressure/high flow/ventilator) at 36 weeks [13].Though the definition of BPD is undergoing several changes [14],we used the same definition throughout the study period.On most occasions,we used low dose dexamethasone to facilitate extubation in infants who continued to be on mechanical ventilation for more than 2 weeks [15].Bayley scales of infant and toddler development,third edition (BSID III) [16] at corrected age of 2 years was collected.Assessment for BSID III was conducted by a certified neonatologist (unblinded) trained in developmental assessment.

Data collection and analysis

Clinical data were collected from the electronic database and patient file using a standard excel sheet.Pre-operative echocardiographic images of all the infants in the study cohort were independently reviewed by two certified cardiologists (CD and JE).These investigators were blinded to the study exposure and the outcomes.The following echocardiographic parameters were used for assessing PDA with left to right shunt:left ventricular end-diastolic diameter(LVEDD);left ventricular end-systolic diameter (LVESD);ratio of left atrial size to aortic root diameter (LA:AO),and ductal diameter.Details of measurements of echocardiography parameters are presented in Supplementary Table 1.We excluded infants with significant congenital heart defects which may affect the measurements.

Descriptive statistics were used for population characteristics.Categorical variables were presented as proportions,while numerical variables were presented as mean with standard deviation (SD) or median with inter-quartile range(IQR) as appropriate.For univariate analysis,we used Chisquare/Fishers exact test for categorical variables and Student’sttest/Mann–WhitneyUtest for continuous variables.An adjusted odds ratio (OR) was calculated using binary logistic regression using the confounding risk factors that were significant (P<0.05) in the univariate analysis as independent factors and “BPD” or “severe BPD” as a dependent factor.For logistic regression,we used the Hosmer–Lemeshow test to understand model fit (P>0.05).Kaplan–Maier analysis was done for duration of hospitalisation and days for extubation post ligation in early versus late ligation groups.APvalue <0.05 was considered statistically significant.All the statistics were performed using IBM SPSS Statistics(version 23;IBM Corp.,Armonk,New York).The project was registered as a service evaluation with local hospital authorities.

Results

A total of 133 infants were ligated over the 10-year study period.Figure 1 provides the study flow diagram for the cohort of infants underwent PDA closure during the study period.Mean gestational age and birth weight (± SD) for infants who underwent ligation was 25 ± 1.6 weeks and 757 ± 211 g,respectively.We had a total of 42 (31.5%)out born preterm infants and these infants underwent PDA ligation at the median (IQR) age of 39 (33–43.5) days.In our cohort 88 (66%) infants received medical treatment for PDA closure and no infants received prophylactic indomethacin.The median (IQR) age for ligation was 32(25–42) days and the median duration of hospital stay was 113.5 (99–136) days.In our study cohort,only two infants before 2 weeks and 14 infants before 3 weeks of age underwent PDA ligation.Eleven (8.2%) infants died before discharge.Five deaths were attributed to NEC,three were attributed to respiratory failure and one was attributed to sepsis.Two infants died after hospital discharge and none of the deaths was attributed to PDA ligation.Echo images were available for 109 infants with a complete data set available for 90 infants.We noticed an increasing trend in the percentage of preterm infants underwent PDA ligation from 2009 to 2013 with highest percentage in 2013(10.4%),following that there was a decreasing trend from 2014 to 2019 (Supplementary Fig.1).Similarly,there was an increasing trend with late PDA ligation from 2009 to 2015,following that there was a decreasing trend.There was no clear pattern with early PDA ligation (Supplementary Fig.2).

Fig.1 Flow diagram for the study.PDA patent ductus arteriosus,GA gestational age,BSID Bayley scales of infant and toddler development,CGA corrected gestational age.a Not eligible for follow-up (did not reach corrected age of 2 years) at the time of writing report

Baseline characteristics of infants with early(≤ 28 days) vs.late (>28 days) ligation

Out of 133 infants,51 (38.3%) had early and 82 (61.7%)had late ligation.Baseline characteristics between the early and late ligation groups are shown in Table 1.Significantly more infants were being ventilated at the time of PDA ligation and had received medical therapy to close the PDA in the early ligation group.Median (IQR) age at the time of ligation in the early ligation group and late ligation group was 24 (20–26.5) days (mean ± SD corrected gestational age:28.4 ± 1.6 weeks) and 38 (33–50) days (mean ± SD corrected gestational age:31 ± 3.2 weeks),respectively.Late ligation group had echocardiographic evidence of volume overloading with statistically higher LVEDD and LVESD.Other baseline characteristics between the groups were comparable.The use of postnatal steroids was comparable between the groups (43% vs.42%,P=0.85).

Outcomes

Significantly more infants developed severe bronchopulmonary dysplasia in the late ligation group compared to the early ligation group [OR=2.4 (95% CI 1.1–5.4),P=0.02](Table 2).There was no significant difference in other neonatal morbidities or mortality between the groups.There was no significant difference in the median age at extubation post-ligation between early and late ligation groups.Regression analysis using “severe BPD/death” as the dependent outcome and medical treatment before ligation (categorical),birth year (continuous) and age at ligation (continuous) as independent variables;none of the factors used in the regression analysis were significantly associated with severe BPD.Age at ligation had an adjusted OR of 1.04 with over lapping confidence interval (95% CI 1.0–1.1,P=0.02) for severe BPD/death (Hosmer and Lemeshow test:0.79).In the posthoc analysis,with severe BPD incidence of 63% &81% in the early (n=51) &late ligation group (n=82),respectively,and with the alpha value of 5%,our study had a power of 63%.BPD and/or death is outcome present in most of the infants in the cohort (97%),hence we performed logistic regression only with “severe BPD” as dependent outcome.Kaplan–Meier analysis showed a marginal reduction of hospital stay in the early ligation group [113 (105–121) days vs.115 (107–123) days;log rankP=0.026] with overlapping confidence interval and with no difference in day of extubation between early vs.late ligation (log rankP=0.78)(Fig.2).Post-operative complications were similar with pneumothorax noted in 3 (6%) infants in the early ligation group compared to 4 (5%) infants in the late ligation group.Vocal cord palsy was noted in 2 (4%) infants in the early ligation group compared to none in the late ligation group.

Fig.2 Kaplan–Meier analysis for duration of hospitalization (a) and day of extubation (b) in early (≤ 28 days) vs.late (>28 days) PDA ligation groups.a Duration of hospital stay with PDA ligation before or after 28 days; b day of extubation with PDA ligation ≤ 28 days.PDA patent ductus arteriosus

The outcomes of Bayley scales of infant and toddler development

Complete BSID data was available for 72 (54%) infants,38(28.6%) did not attend for their BSID assessment,12 were not eligible as they were still less than 2 years at the time of this report and 11 were dead.With the limited BSID data available,there was no difference in all three domains of BSID between the two groups.There is no difference in characteristics between infants who completed BSDI versus infants who did not complete BSID assessment (Supplementary Table 2).

Table 1 Baseline characteristics of early (≤ 28 days) and late ligation (>28 days) groups

Early (≤ 7 days) vs.late extubation (>7 days)post patent ductus arteriosus ligation

Mean gestational age [mean difference 0.8 weeks (95% CI 0.28–1.3)] and birth weight [mean difference 33 g (95% CI 31–161)] were significantly higher in the early extubation group (Supplementary Table 3).No significant difference was noted in the use of post-natal steroid (PNS) prior to ligation,and medical treatment for ductal closure between the groups.Use of PNS [OR=0.27 (95% CI 0.12–0.61),P=0.001] post-PDA ligation was significantly less in the early extubation group.Infants in early extubation had statistically higher LA:AO ratio as compared to late extubation group.On performing receiver operator curve characteristics(area under the curve of 0.61),LA/AO ratio of ≥ 1.75 prior to duct ligation would give sensitivity of 41% and 80% specificity for early extubation (≤ 7 days).By performing logistic regression with extubation within 7 days of PDA ligation(categorical) as dependent variable and birth gestational age,birth weight,day of PDA ligation (continuous variable) asindependent variables none of the factors were significantly associated with extubation within 7 days of PDA ligation(Hosmer and Lemeshow test:P>0.05).Age at PDA ligation had OR of 1.04 (95% CI 0.99–1.1,P=0.07) for extubation within 7 days of PDA ligation.There was no significant difference in other neonatal morbidities between the early and late extubation group (Supplementary Table 4).The median age for weaning of respiratory support to room air/low flow,median duration of hospital stay,and death were significantly less in the early extubation group.

Table 2 Comparison of outcomes in early (≤ 28 days) and late ligation (>28 days) groups

Discussion

In this large observational data over 10-year period,we report the outcomes of early and late PDA ligation which has been a topic of debate.Although previous studies [17,18] have varied time period definition for early ligation age ranging from 2 to 3 weeks,we pragmatically defined early ligation as ligation performed during neonatal period(≤ 28 days).We had a very small number of infants underwent ligation before 14 days (n=2) hence it was not possible for us to perform analysis based on this early cut off time period.The median age of ligation was 32 days in our study which is similar to a previous reported UK wide prospective study (median age of ligation was 33 days) [19].In our study there was a small proportion of infants (14%)who were managed with non-invasive respiratory support underwent PDA ligation.This could be due to various clinical circumstances,such as inability to wean non-invasive respiratory support,persistent high oxygen requirement and signs of cardiac failure.

Infants in the late ligation group had significantly higher left ventricular dimensions which could be an effect of longstanding PDA or due to their higher age/weight at the time of ligation.Most infants,who had PDA ligation,in this cohort were ≤ 27 weeks (94%) and ≤ 750 g (59.4%),which is similar to the population from previously reported studies [20,21].This is also the population who are most likely not to respond to medical treatment [21].Though we don’t have enough explanations for our high rates of BPD,we presume this could be due to our cohort of smaller gestational age group/birth weight and poor clinical/respiratory conditions needing referral for PDA ligation.This may also be related to lung injury following prolonged ventilation,exposure to a significant duct or the ligation procedure itself.Surgical ligation could be marker of disease severity rather than being an independent risk factor for the development of BPD.However,contrary to some of the reports our study found less severe BPD in the early ligation group [20,22].Those exposed to late ligation were often those with a more severe lung disease,possibly outlining a selection of those newborns with more extreme pulmonary phenotype.However,a recent study showed a significantly increased risk of BPD/death in extreme preterm infants exposed to even a short period (7–13 days) of exposure to a moderate-to-large PDA with no further added risk after continuous exposure beyond this period [23].A reduced number of severe BPD cases could also explain the marginal reduction in hospitalisation days in the early ligation group.Marginal reduction in the duration of hospital must be interpreted with caution as there was overlapping of confidence intervals.

One of the main controversies surrounding surgical PDA ligation has been the timing of closure and there is significant variation in both the numbers of patients referred for closure and the timing of that closure between units [10,11].Often,initially,medical management is preferred with courses of non-steroidal anti-inflammatory drugs (NSAID)to close a PDA.In addition to the perceived invasive nature of surgical ligation as well as its more complex deliverability may lead to a relative delay in seeking surgical closure.However,targeted medical management of PDA with NSAIDs is not benign with a range of adverse effects reported including NEC,gastrointestinal bleeding,renal dysfunction and periventricular leukomalacia [24].Effect of NSAID drugs in PDA closure decreases after the first week of life,and a proportion of infants will have contraindications for medical therapy (e.g.,NEC).NSAID drugs for PDA closure have not shown any other short or long-term benefits [25].Cochrane systematic review comparing surgical versus medical treatment for PDA closure concluded that there is insufficient evidence to recommend medical or surgical treatment as preferred treatment method for PDA closure [26].Whether PDA should be closed in the first place is another major controversial topic.Few recent retrospective studies have reported high rates of spontaneous closure (>90%) even in extreme preterm population without increasing major morbidities [27,28].

Our results showed that timing of ligation does not influence the timing of the extubation,and other factors,e.g.,birth weight and gestation,might play a significant role in facilitating early extubation.In addition,in the absence of uniform weaning/extubation guidelines,individual physician preference would have influenced extubation.In other infants with multiple problems including severe lung disease,PDA ligation would have been carried out as a “treatable” lesion to try and progress the baby despite a relatively insignificant duct.

In a meta-analysis of 397 very low birth weight infants,early ligation group had better respiratory and nutritional outcomes than those undergoing late ligation [19].There was no difference in mortality or postoperative complications between the two groups.Another study with 78 infants concluded that early ligation (median age of ligation was 12 vs.36 days) was associated with a lower prevalence of late onset sepsis and oxygen at discharge [29].If PDA ligation was performed with extubation as the primary goal,whether using PNS would have helped in extubation rather than PDA ligation needs to be answered in future studies.Approximately 40% of infants received PNS in our whole study cohort to help with extubation post ligation and remaining extubated with ligation as the only major intervention.This probably explains that in a certain group of infants,PDA ligation might play a role in aiding extubation.

The main strengths of this study are its sample size,and assessment of long-term neurodevelopmental outcomes.In our whole study cohort only two infants had vocal cord palsy(1.5%) which is less than reported in the literature [30].Major limitations of this study include the lack of data on ventilation parameters,lack of detailed echocardiographic parameters (ductal steal,cardiac output,strain analysis,etc.)and missing BSID data.To a certain extent,lack of data on ventilation parameters provides generalisability of the results irrespective of their respiratory status.One third of our cohort were out born and the lack of uniform criteria(echocardiographic parameters and clinical–respiratory parameters) for PDA ligation could have introduced selection bias.We did not have data on 5-minute Apgar score,Clinical risk index for babies-II score,cord gas,fluid status and post-PDA ligation syndrome.We did not have weight and length at the time of PDA ligation to calculateZscores for echocardiographic parameters.Our study had a power of 63% only,which could limit the applicability of the results.In our study,we only compared early vs.late ligation and did not include the control of medical or no treatment of PDA which limits the extension of findings to all neonates with a PDA.We did not account for changes in the clinical practices of PDA management in the 10 years with the secular trend of deferring surgical ligation even if medical management fails [31,32] which may have biased some results.In addition,all the PDAs were ligated using the traditional surgical approach,which entails a thoracotomy.There were no device closures in our population cohort,which is a limitation for the generalisation of results.

While there remain,ongoing trials exploring the optimal timing and dosage schedules of NSAIDs to medically treat PDA there are is very little evidence comparing medical management to surgical ligations [33].With the increasing use of catheter occlusion as an alternative to surgery or repeated courses of medical therapy further exploration of both the timing and type of PDA treatment is urgently required.Until then,from our study and other evidence,we could understand that in high performing cardiac surgical centres infants undergoing early ligation are less likely to develop severe BPD and those who are exposed to later ligation will often also develop severe BPD due to the possible selection of more severe respiratory cases.For future prospective studies,using our study data for sample size calculation we would need 33 infants in each group with 90% power and 5% type I error to show a reduction in severe BPD (baseline incidence of 73%) by 20%.

Though causation could not be proven from this retrospective study and other existing studies,well designed prospective randomised trial with standardised PDA ligation/closure criteria (using standardised echocardiographic parameters and clinical–respiratory parameters) would provide a definitive answer.

In conclusion,we observed decreased incidence of severe BPD and marginal reduction in duration of hospital stay in infants with early PDA ligation.Timing of PDA ligation did not facilitate early extubation.Apart from early extubation,future studies should focus on other important clinical outcomes,such as BPD.

Supplementary InformationThe online version contains supplementary material available at https:// doi.org/ 10.1007/ s12519-022-00527-3.

AcknowledgementsWe would like to acknowledge the contributions and support of all the parents and their babies participated in our study.We would like to thank all the staff from our hospital.

Author contributionsAG was responsible for concept,design,data collection,interpretation of data,and drafting of the initial manuscript.NA was responsible for concept,design,and drafting of the initial manuscript.EJ and DC was responsible for ECHO data collection and drafting of the initial manuscript.PKL was responsible for concept,design,interpretation of data,data analysis,and drafting of the initial manuscript.All authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.

FundingNone.

Data availabilityOriginal data would be available upon request.Please email your request to:pkannanloganathan@nhs.net.

Declarations

Ethical approvalThe project was registered as a service evaluation with local hospital authorities.

Conflict of interestAuthors have no conflicts of interest relevant to this article to disclose.No financial or non-financial benefits have been received or will be received from any party related directly or indirectly to the subject of this article.None of the authors serves as a current Editorial Team member.