Background
A patent ductus arteriosus (PDA) is one of the most frequent complications of the premature neonate, especially those with accompanying lung disease.(1,2) Studies have been reported examining outcomes of very low birth weight (VLBW) neonates at Charlotte Maxeke Johannesburg Academic Hospital (CMJAH).(3) However, no study has explored factors associated with symptomatic PDA in VLBW neonates at CMJAH.
The National Collaborative study conducted in the United States in the early 1980s, reported an incidence of symptomatic PDA ranging from 11% to 36% in 13 different neonatal units, with a higher incidence in neonates with hyaline membrane disease (HMD) and those <30 weeks of gestation.(4) Dice et al. in a review article reports an incidence of 48% in neonates <1000 g.(5) Clinical diagnosis of symptomatic PDA (such as bounding pulses, active precordium, wide pulse pressure, systolic murmur on left sternal border, increase in ventilator requirements despite surfactant therapy, congestive cardiac failure and apnoea) lacks accuracy, particularly in the early post-natal days.(5,6) Both false negative and false positive clinical findings are common in the first post-natal week. As a result, echocardiography has become an important non-invasive method for confirming ductal patency and determining its haemodynamic significance.(6)
In addition to prematurity and HMD, there are other risk factors associated with persistent ductal patency in VLBW neonates and these include female gender and sepsis.(7) Complications of prematurity that have also been associated with a symptomatic PDA include necrotizing enterocolitis (NEC), intraventricular haemorrhage (IVH), cystic periventricular leukomalacia (PVL), chronic lung disease (CLD) and retinopathy of prematurity (ROP).(8)
Not all VLBW neonates with PDAs need treatment, and trying to define a target population becomes difficult.(8) Some studies use clinical, echocardiographic, pharmacokinetic and biochemical parameters to identify the target population that needs treatment.(8) Time of treatment becomes debatable, whether prophylactically, pre-symptomatically, or when the neonate has symptoms.(8) In addition, individual variability exists as to the drug and dose choice.(8) At CMJAH, VLBW neonates are treated once symptomatic, after confirmation by echocardiography.
Over the past few years, the approach to PDA has changed from early diagnosis and medical treatment (using indomethacin or ibuprofen) to conservative treatment.(8) The PDA in neonates is increasingly being viewed not as a pathological condition but as a physiological bystander.(8) There is further evidence that a PDA has no significant effects on major morbidity and mortality in VLBW neonates.(8) However, a previous study done at CMJAH reported a significant increase in mortality in VLBW neonates with a PDA compared to those without PDA.(3)
The current study aims to highlight clinical conditions significantly associated with a symptomatic PDA, assesses outcome and analyses current practices in the management of symptomatic PDA compared to other studies.
Methods
Study design
This study was a cross-sectional, retrospective, descriptive study of symptomatic PDAs in VLBW neonates weighing between 500 and 1500 g, admitted to CMJAH, a tertiary public hospital in the Gauteng Province in South Africa, between January 2013 and June 2016. Symptomatic PDA is defined as showing a clinical impact (respiratory or circulatory) on the neonate. The terminology ‘symptomatic’ PDA and ‘haemodynamically’ significant PDA are often used interchangeably as there are no standard definitions.(9,10) Entry criteria included all VLBW neonates admitted for longer than 72 h after birth. Exclusion criteria included unobtainable records, major birth defects as per Vermont Oxford Network (www.vtoxford.org) and neonates with cardiac conditions co-existing with a PDA.
VLBW neonates with a clinically symptomatic PDA confirmed on echocardiogram were compared to those without PDA. Echocardiography was obtained within 24 to 48 h after a clinical suspicion of a symptomatic PDA. Echocardiographic criteria of symptomatic PDA included: a left atrial/aortic ratio >1.4 and a PDA diameter >1.4 mm.
Database
Data was collected from an existing neonatal database on the Research Electronic Data Capture platform.(11) Diagnosis of a PDA was substantiated by records captured electronically in the Department of Paediatric Cardiology. Any missing information in both systems was verified through retrieval of patient files stored in the hospital records department. Data collected included (i) maternal data – antenatal steroids, antenatal care, multiple gestation; (ii) neonatal data – gender, age on admission, age on discharge, birth weight; (iii) neonatal ventilator support – resuscitation at birth, nasal continuous positive airway pressure (NCPAP), intermittent positive pressure ventilation (IPPV), surfactant at any point; (iv) complications of prematurity – HMD, NEC stages 2 and 3 as per modified Bell's staging criteria,(12) blood transfusion, CLD, steroids for CLD, late-onset sepsis (LOS), IVH grades 3 and 4 as per sonographic grading system by Papile et al.,(13) pneumothorax, pulmonary haemorrhage, ROP stages 3 and 4 (14) and cystic PVL. Management of PDA including conservative treatment, indomethacin, ibuprofen and surgical ligation was recorded. Conservative treatment included fluid restriction to 150 ml/kg/day beyond day 5 of life, treatment of cardiac failure with furosemide and optimizing haemoglobin levels to 14 g/dl. VLBW neonates who were discharged and those who were transferred out were grouped together as survivors for the purpose of this analysis.
Statistical analysis
Data was captured into MS Excel (Microsoft, USA) and exported into a statistical software program SPSS 23 (IBM, USA) for analysis. Continuous variables were normally distributed and were described using mean and standard deviation (SD). Categorical variables were described using frequencies and percentages. Comparison was made between VLBW neonates with symptomatic PDA vs those without PDA for clinical characteristics and outcome. Continuous variables were normally distributed and compared using unpaired t-test. Categorical variables were compares using χ2 tests or Fisher's exact test, as appropriate. A p-value <0.05 was considered significant. Variables with a p-value <0.1 on univariate analysis were entered into a binary logistic regression model to determine the most significant association with a symptomatic PDA. Only valid data was reported (i.e. missing variables were excluded).
This study was approved by the Human Research Ethic Committee of the University of the Witwatersrand, Johannesburg.
Results
There were 1844 eligible VLBW neonates. Three hundred and eight were excluded, leaving 1536 neonates in the final sample. One hundred and eighty-four (11.9%) had echocardiographically confirmed PDA and the remaining 1352 (88.0%) had no PDA (Fig 1). The mean birth weight of all eligible neonates was 1157.0 g (SD 222.1), the mean age on admission was 1.2 days (SD 7.0) and the mean age at discharge was 34.8 days (SD 21.9).
VLBW neonates who had PDAs were of a lower birth weight than those without PDA (p < 0.001) (Table 1). Duration of hospitalization was significantly longer in neonates with PDA as compared to those without PDA (p < 0.001).
Comparison between neonates with PDA and without PDA for continuous variables.
| Variable | PDA (mean SD) | No PDA (mean SD) | p-Value |
|---|---|---|---|
| Birth weight (g) | 1062.3 (118.0) | 1169.9 (224.1) | <0.001 |
| Age on admission (days) | 1.4 (7.2) | 1.2 (7.0) | 0.76 |
| Age on discharge (days) | 54.2 (28.9) | 32.2 (19.3) | <0.001 |
Clinical and demographic characteristics in VLBW neonates with a symptomatic PDA as compared to those without a PDA are shown in Table 2. Neonates with a symptomatic PDA as compared to those without a PDA were significantly less likely to have received antenatal steroids (p < 0.001) or to have a mother who attended antenatal care (p = 0.002). Resuscitation at birth (p = 0.001), HMD (p < 0.001), NCPAP (p < 0.001), IPPV (p < 0.001), surfactant therapy (p < 0.001), NEC (p < 0.001) and blood transfusion were significantly more common in VLBW neonates with symptomatic PDA than those without PDA [155/183 (84.7%) vs 479/1321 (36.3%), p < 0.001]. Other complications of prematurity such as CLD (p < 0.001), LOS (p < 0.001), IVH grades 3 and 4 (p < 0.001), pneumothorax (p = 0.006) and pulmonary haemorrhage (p = 0.026) were all significantly more common in VLBW neonates with a symptomatic PDA. The overall rate of LOS in these VLBW neonates was 33.2%.
Comparison between neonates with PDA and without PDA for clinical and demographic characteristics.
| Variable | Total sample (1536) n (% of total) | PDA (184) n (%) | No PDA (1352) n (%) | p-Value |
|---|---|---|---|---|
| Antenatal steroids | 671/1363 (49.2) | 54/151 (35.8) | 617/1212 (50.9) | <0.001 |
| Antenatal care | 1154/1441 (80.1) | 117/165 (70.9) | 1037/276 (81.3) | 0.002 |
| Resuscitation at birth | 556/1411 (39.4) | 82/159 (51.6) | 474/1252 (37.9) | 0.001 |
| NCPAP | 1015/1419 (71.5) | 154/171 (90.1) | 861/1248 (69.0) | <0.001 |
| IPPV | 317/1386 (22.9) | 101/177 (57.1) | 216/1209 (17.9) | <0.001 |
| Surfactant therapy | 1054/1491 (70.7) | 164/172 (95.3) | 890/1319 (67.5) | <0.001 |
| HMD | 1352/1518 (89.1) | 178/183 (97.3) | 1174/1335 (87.9) | <0.001 |
| NEC 2 and 3 | 161/1533 (10.5) | 33/183 (18.0) | 128/1350 (9.5) | <0.001 |
| Blood transfusion | 634/1504 (42.2) | 155/182 (84.7) | 479/1321 (36.3) | <0.001 |
| CLD | 418/1508 (27.7) | 119/183 (65.0) | 299/1325 (22.6) | <0.001 |
| Steroids for CLD | 230/1485 (15.5) | 70/174 (40.2) | 160/1311 (12.2) | <0.001 |
| LOS | 505/1522 (33.2) | 106/182 (58.2) | 399/1340 (29.8) | <0.001 |
| IVH 3 and 4 | 67/929 (7.2) | 18/158 (11.4) | 49/771 (6.4) | <0.001 |
| Pneumothorax | 10/1531 (0.7) | 4/183 (2.2) | 6/1348 (0.4) | 0.006 |
| Pulmonary haemorrhage | 22/1536 (1.4) | 6/184 (3.3) | 16/1352 (1.2) | 0.026 |
| Multiple gestation | 271/1503 (18.0) | 20/180 (11.1) | 251/1323 (19.0) | 0.010 |
| ROP 3 and 4 | 16/467 (3.4) | 4/82 (4.9) | 12/385 (3.1) | 0.120 |
| Cystic PVL | 14/1188 (1.2) | 3/167 (1.8) | 11/1021 (1.1) | 0.425 |
| Gender (male) | 719/1533 (46.9) | 85/183 (46.4) | 634/1350 (47.0) | 0.896 |
| Death | 270/1536 (17.6) | 41/184 (22.3) | 229/1352 (16.9) | 0.074 |
There was no significant difference in multiple gestation, ROP, cystic PVL and gender between VLBW neonates with symptomatic PDA and those without a PDA. The overall mortality rate between VLBW neonates with PDA and those without PDA was not statistically significant [41/184 (22.3%) vs 229/1352 (16.9%), respectively, p = 0.074].
Variables with p < 0.1 were entered into binary logistic regression to determine which are most significantly associated with a symptomatic PDA. Blood transfusion (p = 0.001), surfactant therapy (p = 0.019), IPPV (p < 0.001), age on discharge (p < 0.001) and antenatal steroids (p = 0.001) (see Table 3) were most significantly associated with a symptomatic PDA.
Binary logistic regression for factors associated with PDA in VLBW neonates at CMJAH.
| Variable | OR | 95% CI | p-Value |
|---|---|---|---|
| Blood transfusion | 3.180 | 1.578–6.405 | 0.001 |
| Surfactant | 3.190 | 1.215–8.381 | 0.019 |
| IPPV | 2.746 | 1.703–4.429 | <0.001 |
| Age on discharge | 1.028 | 1.017–1.039 | <0.001 |
| Antenatal steroids | 0.459 | 0.285–0.738 | 0.001 |
Ibuprofen was administered to 110/184 (59.8%) of the VLBW neonates with symptomatic PDA, while indomethacin was administered to 5/184 (2.7%). The PDA was ligated in 9/184 (4.9%) of the VLBW neonates with symptomatic PDA. The remaining VLBW neonates were treated conservatively, i.e. 60/184 (32.6%).
Discussion
The PDA is a vascular structure allowing a left-to-right shunt from the proximal descending aorta to the main pulmonary artery.(6,15) It becomes persistent once there is failure of the ductus arteriosus to close within 72 h after birth.(6,15) In this study, the incidence of symptomatic PDAs was 11.9%, which is higher compared to previous studies at the same institution, where it was reported to be 5.6% in 2006/2007 and 9.6% in 2013.(3) The increase in incidence may in part be due to increasing survival of extremely low birth weight (ELBW) neonates at CMJAH.(3) The incidence in the current study is comparable to the National Collaborative study in the United States in the early 1980s.(4)
There was a significant association between surfactant therapy and a symptomatic PDA in our study. This finding is in agreement with a study done by Kumar.(16) It is vital that VLBW neonates who receive surfactant therapy for HMD, be monitored clinically and by echocardiography for the appearance of a haemodynamically significant PDA.(16) The effect of surfactant therapy on PDA is unclear. Exogenous surfactant therapy may increase shunting across PDAs following rapid decrease in pulmonary vascular resistance.(17) However, it has been hypothesized that therapies which result in less mechanical ventilation and improvement of lung function (such as surfactant therapy, fluid restriction, NCPAP) could be associated with lower incidence of symptomatic PDAs.(17) Our study supports this finding because a significantly lower number of VLBW neonates with symptomatic PDAs required mechanical ventilation, this may be due to the rise in use of surfactant therapy and NCPAP in ELBW neonates at CMJAH.(3)
The current study is in agreement with reports, which found the closure of symptomatic PDA to be less likely to occur in neonates who did not receive antenatal steroids.(6,18) The present study demonstrated a decreased incidence of PDA with antenatal steroid use. Despite desperate measures to optimize respiratory function, there is still a strong association between symptomatic PDA and development of CLD.(19) This is further emphasized by a recent study done by Mphaphuli on CLD.(20) The presence of a PDA with sepsis substantially increases the risk of CLD.(19) Previous studies of VLBW neonates at CMJAH revealed an increasing number of LOS in 2013 (19%) compared to 2006/2007 (12.5%).(3) The rate of sepsis in the current study was even higher at 33.2%. Therefore, the prevention of sepsis should be prioritized so that the adverse effects of ductal patency can be altered.(19)
Blood transfusion was significantly associated with PDA in our current study. This is unsurprising as a lower haematocrit after birth places VLBW neonates at greater risk of PDA.(21) All preterm neonates develop physiological anaemia of prematurity and this is aggravated by iatrogenic blood loss.(22) However, PDA and blood transfusion in VLBW neonates may place the neonate at even greater risk of NEC than each of these factors alone.(21) This is further emphasized in our study which found a significantly greater number of VLBW neonates with symptomatic PDA developing NEC compared to those without PDA.
Over the years, the haemodynamic and pulmonary consequences of a persistent PDA justified closing of the duct.(23) Neonates <28 weeks gestation or with birth weights ≤1000 g are the target group for controversy since only 60% will develop significant debilitating PDA-related symptoms, but only 30% will spontaneously close their PDA.(18) Observational studies demonstrate that the majority of VLBW neonates with persistent PDA after the third day of life, especially those with ≥1000 g birth weight, do well without any treatment.(23) Current recommendations prefer conservative treatment.(24) It is interesting to note that very few VLBW neonates in our current study received conservative treatment compared to medical treatment. In the current study, ibuprofen was more popular than indomethacin. Only oral preparations of both drugs are available at CMJAH. Intravenous indomethacin is not registered in South Africa. Oral ibuprofen is available as a syrup and is favoured because of its fewer side effects compared to indomethacin.(24) Intravenous ibuprofen is very expensive. A randomized pilot study suggests that oral ibuprofen is as good as intravenous ibuprofen in terms of ductal closure and is associated with fewer side effects.(24)
Only 4.9% of VLBW neonates with symptomatic PDA underwent ductal ligation. It has been found that delaying ligation (by >30 days) causes a reduction in several morbidities associated with ligation.(18,25)
The overall mortality rate of VLBW neonates in our study was 17.6%. There was no significant difference in the mortality rates between neonates with symptomatic PDA and those without PDA. This is encouraging because it is in agreement with studies across the world which show a failure of long-term benefits of ductal closure from several adverse outcomes that are strongly associated with persistent PDA such as death, HMD, NEC and IVH.(23,25) Tools need to be developed to identify VLBW neonates at risk of detrimental outcomes related to a symptomatic PDA such as those with large left to right shunts.(19,23) This must be followed by intervention trials to conclude whether closing the PDA can avoid such outcomes.(19) Until then, choosing to aim to close a PDA of a VLBW neonate who is ill actually remains subjective.(19)
Study Limitations
Our study was limited as this was a secondary analysis of a database. Timing of treatment or whether multiple courses of medication were given could not be established. The dose of medication and timing of complications could not be evaluated.
Conclusion
Our study described the incidence and clinical characteristics significantly associated with a symptomatic PDA. It highlighted the most common method of treatment used for symptomatic PDAs in VLBW neonates at CMJAH. Although management of a symptomatic PDA remains highly controversial, there is a move towards a more conservative approach at the CMJAH neonatal unit.