The dilemma of feeding during the treatment of patent ductus arteriosus with oral ibuprofen in preterm infants ≤30 weeks of gestation—a randomized controlled trial

To evaluate the effect of minimal enteral feeding (MEN) versus withholding feeding on time to reach full feeds during treatment of hs-PDA with oral ibuprofen in infants ≤30 weeks. We performed a single-center, randomized control trial of 126 premature infants born ≤30 weeks gestation, <7 days of age with hs-PDA comparing continuation of MEN (n = 64) vs no feeding (n = 62) during treatment. The primary outcome was time to reach a feed volume of 150 ml/kg/day. Secondary outcomes included were episodes of feed intolerance, GI bleed, NEC and other comorbidities. There was no difference in the time to reach full feeds - median age of 16 days in both groups (p = 0.573). Incidence of feed intolerance, NEC and other secondary outcomes were also similar in both groups. Continuing MEN during treatment of hs-PDA with oral ibuprofen does not decrease time to reach full enteral feeds in very preterm infants.


INTRODUCTION
Patent ductus arteriosus (PDA), can result in significant morbidity and mortality especially in very and extremely preterm infants. The clinical consequences depend on the degree of left-to-right shunting, known as steal phenomenon, which increases the risk of renal dysfunction, necrotizing enterocolitis (NEC), intraventricular hemorrhage (IVH), postnatal growth faltering in addition to deterioration of respiratory status [1].
The systemic hypoperfusion that occurs secondary to the steal phenomenon compromises mesenteric blood flow often leading to adverse gastrointestinal effects like feed intolerance and NEC [2]. The cyclooxygenase inhibitor, ibuprofen, used for the treatment of hemodynamically significant patent ductus arteriosus (hs-PDA) in preterm infants also causes gastrointestinal side effects by alterations in gastrointestinal permeability leading to concerns about enteral feeding when this drug is used [3].
Some clinicians tend to discontinue or reduce enteral feed volume during treatment of PDA, in order to reduce the gastrointestinal complications, particularly NEC, while others continue enteral feeds unchanged. Currently, there is limited evidence to support either approach [4]. Withholding enteral nutrition for even short periods of time may result in gut atrophy and impair digestive processes.
With this background, we conducted this study with a hypothesis that infants who received minimal enteral nutrition (MEN) while on oral ibuprofen for early treatment of hs-PDA during the first week of life would reach full enteral feeds earlier compared to those who were nil by mouth during the period of treatment.

MATERIALS AND METHODS
This was an open label parallel group randomized controlled study conducted over 2 years from January 2020 to December 2022 in a tertiary NICU at a major teaching hospital in India, after approval of institutional ethics committee(IEC/90/19) and CTRI registration (2020/03/02394). The study was performed in accordance with declaration of Helsinki.

Randomization
The infants receiving oral ibuprofen for treatment of hs-PDA were randomized into two groups: Group 1 -Feeding group to receive minimal enteral nutrition (MEN) and Group 2 -No feeding group to be kept nil per oral.
As a unit policy, we used to withhold enteral feeds in infants on treatment with oral ibuprofen for hs-PDA till completion of treatment.

Sample size
Based on our observations in the preceding year, the average age at achieving enteral feed volumes of 150 ml/kg/day in infants ≤30 weeks when feeds were withheld during ibuprofen treatment was 15 ± 4 days. Based on the results of DAFII trial which showed 2.5 days reduction in time to reach feeding volume of 120 ml/kg/day in the feeding group, we hypothesized that we would need to enroll 126 infants, with 63 infants in each group to achieve a 15% (2.2 days) reduction in time to reach enteral feed volume of 150 ml/kg/day, (with a power of 80 percent and two sided alpha error of 0.05) with continuation of MEN feeds during oral ibuprofen for early treatment of hs-PDA.

Inclusion and exclusion criteria
We included all intramural preterm infants ≤30 weeks of gestation on any form of respiratory support diagnosed to have a hemodynamically significant patent ductus arteriosus (hs-PDA) needing early treatment with oral ibuprofen after a written informed consent from the parents. We excluded infants with (a) Contraindications for the use of ibuprofen before the start of drug such as active bleeding, gastrointestinal bleeding, intracranial hemorrhage (grade II or high) or, thrombocytopenia (platelet count <60 × 10 9 /L), significant impairment of renal function with serum creatinine >1.5 mg/dL. (b) Congenital heart disease with duct dependent pulmonary/ systemic circulation. (c) Major congenital, chromosomal & gastrointestinal anomalies d) Severe shock requiring 2 or more inotropes for hypotension (e) Treatment of PDA after first week of life. (f) Infants already on feed volumes >60 ml/kg/day. (g) Outborn infants.
The infants were randomized after written informed consent from the parents into two groups using a computer generated random number sequence and allocation concealment was done using serially numbered opaque sealed envelopes.

Procedure
PDA diagnosis and treatment protocol. All eligible infants ≤30 weeks on any form of respiratory support were screened for hs-PDA between 12-24 h and if needed, beyond 24 h till 7 days (if infants persisted to need respiratory support) with point of care echocardiography and doppler studies. The transductal diameter was measured in parasternal short axis and high parasternal ductal views at the narrowest point of the ductus at the pulmonary end without putting colour doppler. An average of three measurements with the same settings was taken. The ratio of left atrium to aortic root diameter was measured in parasternal long axis by putting M-mode perpendicular to the aorta at the level of aortic valve. Left pulmonary artery end diastolic velocity (LPA-EDV) in cm/sec was recorded using a pulse wave Doppler. Doppler was put in left pulmonary artery making sure that angle of insonation always was <20°. Criteria for diagnosis of hs-PDA requiring treatment in this study were presence of all of the following [5,6]: (1) PDA size >1.5 mm (2) Left atrium (LA)/ Aortic (Ao) ratio ≥1.4, and (3) Left pulmonary artery End diastolic velocity ≥30 cm/s. hs-PDA was treated with 3 doses of oral ibuprofen @ 10, 5 and 5 mg/kg/ dose at 0, 24 and 48 h respectively [7]. An echocardiogram and Doppler study was performed within 24 h of the last dose of ibuprofen to determine residual ductal patency. A repeat course of oral ibuprofen was administered at the same dose (10,5,5) if echocardiographic criteria for hs-PDA were met after rescreening. If hs-PDA persisted despite two courses of ibuprofen, use of paracetamol or surgical ligation was considered for PDA treatment by the treating clinician. If oral ibuprofen was required to be withheld due to adverse effects at any point during the treatment, IV paracetamol at a dose of 15 mg/kg/dose 6 hourly for 5 days was administered to treat the hs-PDA [8].
Feeding protocol. Infants randomized to the 'feeding group' received minimal enteral nutrition (MEN) @ 20 ml/kg/day. In 'no feeding' group infants were kept nil by mouth throughout the period of treatment with oral ibuprofen for hs-PDA. If any infant in the 'no feeding' group was on some feeds before starting of ibuprofen, feeds were stopped till completion of the treatment. If the infant required second course of ibuprofen, they continued to receive MEN or no feeding as per their original group allocation. If ibuprofen was substituted with paracetamol due to gastrointestinal adverse effects of ibuprofen, infants in both groups were kept NPO till the issue settled, and then continued in their original group till completion of PDA treatment. Infants in the feeding group were fed mothers own milk (MOM) or pasteurised donor human milk (PDHM), if MOM was unavailable [9].
Feeding advancement regimen. Once treatment with oral ibuprofen (first or second course) or paracetamol was complete, and the rescreening 2d-Echo confirmed absence of hs-PDA, feeds were increased at the rate of 20 ml/kg/d with MOM/PDHM, fed every 3 hourly in the feeding group. In the no feeding group, MEN was provided for 48 h after completion of oral ibuprofen treatment, thereafter feeds were advanced in the same manner as the feeding group the rate of 20 ml/kg/d [9] (Table 1).
Feed intolerance criteria and its management [9]. Feed intolerance was defined as presence of any one of the following necessitating stopping feeds.
• Increase in abdominal girth (AG) of more than 2 cm with gastric residuals >50% of the previous feed volume.
• Abdominal signs such as abdominal distension, visible bowel loops, abdominal discoloration or tenderness.
While feeding all infants, pre-feed abdominal girth was measured at the level of umbilicus. If abdominal girth increased by >2 cm, gastric residual volumes were measured and managed as follows: • If gastric residual volume was ≤5 mL/kg or <50% of the previous feed volume then the same feed volume was continued [9].
• If gastric residual volume was >5 mL/kg or >50% of the previous feed volume then one feed was withheld. If it recurred for two consecutive feeds, then feeds were stopped for 12 h and restarted on the same feed volume after 12 h if there were no aspirates. The monitoring of abdominal girth was continued.
• If infant had any features suggestive of NEC then feeds were stopped for 48 h or longer and supportive treatment was provided. Feeds were restarted at 20 ml/kg/day and increased @20 ml/kg/day once infant was found to be clinically stable by the clinical team and features of NEC had resolved.

Outcomes
The primary outcome of our study was the time to reach full feeds defined as enteral feed volume of 150 ml/kg/day for 24 h. Secondary outcomes were ascertained by the clinical team as per standard definitions and included episodes of feed intolerance, NEC stage II or higher as per modified Bell's staging) or spontaneous intestinal perforation (SIP) any time during hospitalization, culture proven late onset sepsis (LOS), intraventricular hemorrhage (grade >II according to Volpe classification), any periventricular leukomalacia (as per De Vries classification), bronchopulmonary dysplasia (BPD) as per Jensen definition, retinopathy of prematurity (ROP) requiring treatment, duration of hospital stay among surviving infants & mortality [11][12][13][14]. We recorded adverse effects of ibuprofen like gastrointestinal bleeding, raised creatinine and thrombocytopenia. We monitored for side effects of ibuprofen by measuring baseline urine output, platelet count and serum creatinine before start of ibuprofen and thereafter monitored urine output daily, renal function test (RFT) & platelet counts on alternate days till completion of treatment. Platelet counts were measured daily if baseline platelet counts were less than 150 × 10 6 /L [8]. All patients were followed up till discharge or death.
Statistical analysis was carried out using the SPSS software, version 24.0 for Windows (SPSS, Chicago, IL, USA). Data was summarized by routine descriptive statistics, namely mean and standard deviation for numerical variables that were normally distributed, median and interquartile range for skewed numerical variables, and counts and percentages for categorical variables. Numerical variables were compared by Student's Kaplan Meier model between the two groups after censoring the data for infants who did not meet the primary outcome due to death was also performed.

RESULTS
We assessed 210 infants for eligibility over 2 years. 84 infants were excluded: 49 did not meet the eligibility criteria due to various causes, 4 refused consent and 31 were outborn infants (Fig. 1). Out of 126 very preterm infants enrolled in the study, 64 were randomized to feeding group and 62 to the no feeding group.
Baseline demographic characteristics were comparable between the two groups. Median gestational age and birth weight was 29 weeks and 1200 gm respectively in both the groups. Mean age at enrollment was 16 h in feeding group and 14 h in no feeding group ( Table 2). Fifty four infants in feeding group and 46 infants in no feeding group reached the primary outcome and were analysed. A total of 16 infants in the feeding group and 20 infants in the no feeding group died, of which 10 infants in the feeding group and 16 infants in the no feeding group died before reaching the primary outcome of full feeds. Two and one infant in each group respectively died within 24 h; hence we could not evaluate any outcome in these.
There was no significant difference in the primary outcome between the two groups. Median time to reach full feeds was 16 days in both the groups (p = 0.573) ( Table 3). Survival analysis by Kaplan-Meier model after censoring the data for infants who did not meet the primary outcome due to death also showed no difference between the two groups in time to reach full feeds [log Consort flow diagram    (Fig. 2). There was no significant difference in secondary outcomes such as feed intolerance before reaching full feeds (24% vs 29%). Incidence of feed interruption due to causes other than feed intolerance such as sepsis, recurrent apneas, respiratory deterioration, hypotension requiring 2 or more inotropes, or blood transfusions was comparable (30% vs 32%) in both groups. Incidence of NEC stage II or more/SIP any time during hospitalization, culture proven LOS, total duration of respiratory  support, duration of hospital stay, weight at discharge, IVH, BPD, ROP and mortality were similar in both groups (Table 3). Six infants in feeding group and 5 in no feeding group required two courses of ibuprofen for closure of ductus. All infants who tolerated ibuprofen had complete response and no infant required additional therapy with paracetamol. Ibuprofen had to be substituted with paracetamol due to GI adverse effects of ibuprofen in 16 infants in feeding group and 18 infants in no feeding group. However, there was no significant difference in incidence of adverse effects of ibuprofen like GI bleeding, thrombocytopenia or renal impairment between the two groups.
Overall the PDA closure rate was 100% with medical management in both the groups and none of the infants required surgical ligation.

DISCUSSION
Despite limited evidence, clinicians tend to discontinue or reduce enteral feeds during treatment of hs-PDA with oral ibuprofen. This is mainly due to fear of gastrointestinal morbidities like feed intolerance, NEC & bleeding -either as a complication of hs-PDA or side effects of therapy.
Studies have shown an association between a persistent PDA and feeding intolerance in preterm infants. Hs-PDA has a profound impact on mesenteric perfusion. Term infants are able to maintain their mesenteric blood flow by vasodilating their splanchnic circulation, both at rest and during feeding. However preterm infants have a limited ability to increase their mesenteric blood flow and oxygen extraction at rest [15]. A study done in preterm baboons with moderate PDA showed that they have a limited ability to increase their postprandial mesenteric blood flow velocity which may interfere with their ability to meet the increased intestinal metabolic demands. As a result, hs-PDA may contribute to feed intolerance [16].
Oral ibuprofen also has vasoconstrictive effects on the splanchnic circulation, especially in infants who are NPO. Additionally, there are concerns about the association of oral ibuprofen and the occurrence of GI bleeding, NEC, and bowel perforation due to its high osmolality [17]. A recent meta-analysis showed a reduced risk of NEC with use of oral ibuprofen compared to intravenous ibuprofen; however there was no difference in the risk of SIP and GI bleeding. But the quality of evidence was low [18].
Evidence suggests that starting early minimal enteral nutrition helps in promoting intestinal maturation by improving GI motility, enhancing microbiome development & reducing inflammation. MEN also prevents villous atrophy decreases intestinal permeability and bacterial colonization of the gut thereby improving feed tolerance [19]. In our study MEN was provided for 48 h to ensure that the gut was primed for enough time before feed enhancements so that the risk of feed intolerance due to fasting and ibuprofen was reduced.
The management of enteral feeds during pharmacological treatment of PDA in preterm neonates is still a common clinical dilemma. Therefore, we planned this study to evaluate the effect of feeding versus withholding feeds on time to reach full feeds during early treatment of hs-PDA with oral ibuprofen in very preterm infants.
In this open label randomized controlled trial, we did not find any statistically significant difference in the time to reach full enteral feeds irrespective of whether infants were fed or not during treatment of hs-PDA, even on conducting a survival analysis by Kaplan-Meier model. Infants in both groups reached full feeds at a median age of 16 days (p = 0.573). Both groups experienced feed interruptions not only due to episodes of feed intolerance (24% vs 29%) but also due to sepsis, recurrent apneas, respiratory deterioration, hypotension requiring 2 or more inotropes and need for blood transfusions (30% vs 32%). We speculate that the infants in the feeding group might have had longer duration of feed interruptions for the aforementioned reasons.
Our findings are in contrast to the DAFFII study, which is the first randomised controlled trial that studied effect of enteral feeding during ibuprofen and indomethacin therapy. The authors observed a significant reduction of 2.4 days in time to reach full feeds in feeding group [3]. However, 78% of infants received indomethacin in the DAFFII trial, and only 22% received ibuprofen. There is no other study that has evaluated the effect of feeding or no feeding when oral ibuprofen is used for treatment of PDA. Ibuprofen is known to cause less vasoconstriction in splanchnic circulation in comparison to indomethacin; this may be a reason why we observed no difference in time to reach full feeds whether infants were fed or not during ibuprofen treatment.
In our study, number of episodes of feed intolerance before reaching full feeds was similar in the two groups (24% vs 29%). However the DAFII study found otherwise; they reported fewer episodes of feed intolerance in infants in the feeding arm, though it was not statistically significant. In contrast to the DAFII study, Louis et al. observed lower incidence of feed intolerance in infants who were kept nil by mouth [19]. The incidence of NEC anytime during hospitalization was similar in both the groups.(1.6% vs 3.2%). These observations are similar to the DAFII study and the study by Louis et al., where no significant difference in the rate of NEC was observed.
Similar to the DAFFII study, we did not find any difference in other secondary outcomes like mortality, duration of hospital stay, weight at discharge or any other major morbidities like IVH, BPD, PVL or ROP between the two groups. There were no significant differences in adverse effects due to oral ibuprofen like GI bleeding, thrombocytopenia or raised creatinine levels in our study in both groups.
Our study had certain limitations. We did not adjust for the deaths before reaching the primary outcome which might have resulted in inadequate power to detect significant difference in the primary outcome between the two groups. Blinding of the "feeding" intervention was not possible which had the potential to influence the treating clinical team's decisions about feeding advances and more importantly the lengths of periods of NPO during episodes of feeding intolerance. However we tried to minimize this bias by following a standardized feeding advancement regimen and feeding intolerance protocol.
To conclude, our study shows that during oral ibuprofen therapy for hs-PDA in the first week of life for very preterm infants, MEN did not reduce the time to reach full enteral feeds compared to keeping the infant nil per oral. More studies are required to establish a feeding protocol during oral ibuprofen therapy for hs-PDA due to lack of benefit of either strategy.