Background There is little consensus regarding the use of echocardiography in patent ductus arteriosus (PDA) treatment in preterm infants. The use of troponin T (cTnT) and N-terminal Pro-BNP (NTpBNP) in combination with echocardiography assessment may facilitate the development of a superior predictive model.
Objective To investigate the ability of cTnT, NTpBNP and a PDA scoring system applied at 48 h of life to predict death before discharge and neurodevelopmental outcome at 2 years of age.
Design/Methods Infants <32 weeks and <1500 g were prospectively enrolled. Echocardiography evaluation coupled with cTnT and NTpBNP measurements were done at 48 h. The ductus arteriosus was scored (0–6) according to echocardiography markers of haemodynamic significance. Infants were assessed at 2 years using the Bayley scales and categorised into two groups: Severe Disability/Death before discharge or Normal/Mild Disability.
Results Sixty infants with a median gestation of 27.7 weeks (26.2–29.4) and a median birth weight of 1.01 kg (0.86–1.22) were followed up to 2 years of age. Plasma cTnT and NTpBNP were higher in the Severe Disability/Death compared to the Normal/Mild Disability group (2.30 μg/l vs 0.19 μg/l, p<0.001; 9209 pmol/l vs 1664 pmol/l, p<0.001, respectively). The severe group had a higher PDA score compared to the mild and normal groups (5 vs 2, p<0.001).
Conclusion Blood cTnT, NTpBNP and a PDA scoring system at 48 h may facilitate the identification of those infants with a PDA, who are at greatest risk of poor neurodevelopmental outcome at 2 years of age.
Statistics from Altmetric.com
Premature infants are susceptible to haemodynamic compromise in the first few days of life. The complexity of postnatal adaptation and immaturity of the neonatal myocardium, in addition to the presence of patent ductus arteriosus (PDA) with left to right shunting predispose the infant to organ hypoperfusion. PDA in preterm infants is associated with a higher incidence of intraventricular haemorrhage (IVH) and death.1 2 A haemodynamically significant PDA has a negative effect on cerebral oxygenation in premature infants.3 The relationship of this physiologic disturbance to neurodevelopmental disability later in life remains unknown. Unfortunately, it remains difficult to predict which infants with a PDA are at increased risk of neonatal morbidity and/or poor neurodevelopmental outcome at 2 years of age.4 Conventional echocardiography markers such as ductal diameter and left atrial to aortic ratio (LA:Ao) do not predict neonatal outcome. In addition, treatment based on these parameters has not led to improved neurodevelopmental outcome at 2 years in preterm infants.5,–,8 The goal of PDA treatment is to prevent the potential development of short-term and long-term complications thought to be causally related to the presence of a significant PDA. The development of a comprehensive diagnostic tool applied in the early neonatal life for determining which PDAs warrant treatment remains elusive.
What is already known on this topic
▶ N-terminal Pro-BNP (NTpBNP) and cardiac troponin T (cTnT) rise in the presence of a PDA in preterm infants and fall following successful treatment.
▶ NTpBNP and cTnT are associated with the development of intraventricular haemorrhage in the preterm population with a PDA.
What this study adds
▶ Blood cardiac troponin T (cTnT), N-terminal Pro-BNP (NTpBNP) and a PDA scoring system at 48 h may facilitate the identification of those infants with a PDA, who are at greatest risk of poor neurodevelopmental outcome at 2 years of age.
▶ The use of biochemical markers and a PDA scoring system may provide a basis for trials of a targeted approach to PDA management.
Troponin T is a predictor of myocardial injury in perinatally asphyxiated neonates,9 and levels are higher in premature infants with respiratory distress syndrome (RDS).10 The presence of a PDA is associated with higher cTnT levels in preterm infants compared to controls.11 The nature of the elevated troponin may relate to diastolic steal and lower coronary blood flow leading to ischaemia.12 Cardiac troponin T (cTnT) is released following acute myocardial damage. cTnT assays are not thought to cross react with skeletal forms in the preterm infant.13 B-type natriuretic peptide (BNP) and its inactive by-product N-terminal Pro-BNP (NTpBNP) are released by the stressed neonatal myocardium in response to volume and pressure loading.14 These biochemical markers appear uninfluenced by antenatal and postnatal events including birth weight and gestational age.15 These properties render them ideal markers of myocardial damage and the haemodynamic status in preterm infants. NTpBNP is also associated with a haemodynamically significant PDA; blood levels fall following successful treatment.16 17 Recently, in a cohort of preterm infants, our group demonstrated that cTnT and NTpBNP performed at 48 h of life predict severe IVH and/or death before discharge in preterm infants with a PDA.18
In this study, we evaluated the ability of NTpBNP, cTnT and echocardiography to predict death before discharge or adverse neurodevelopmental outcome at 24 months in the same cohort.
Patients and methods
This was a hypothesis generating analysis from a pilot prospective observational study conducted at the National Maternity Hospital, Dublin, Ireland between July 2006 and June 2007.18 Infants less than 32 weeks gestation and less than 1500 g were eligible. The study was approved a priori by the National Maternity Hospital's Ethics Committee and informed parental consent was requested within the first 24 h of life. Infants with major congenital abnormalities and structural heart disease were excluded. Maternal details including pre-eclampsia, histological chorioamnionitis, antenatal steroid administration, the presence of antepartum haemorrhage were recorded. A complete course of antenatal steroids was defined as betamethasone 12.5 mg given twice, 12 h apart and a single dose was termed partial treatment. Clinical details including RDS were noted. The definition of RDS was based on clinical signs of respiratory distress and a rising pCO2 on an arterial blood gas sample. In addition, radiological diagnosis was used to further support clinical evidence.19
Cranial ultrasound scans were performed at 12 and 48 h, and days 5 and 7 of life by the same radiologist. All ultrasound studies were classified by the radiologist according to the highest grade of IVH. IVH was classified according to the Papile et al20 method as no IVH, IVH grades I and II, and IVH grades III and IV. Echocardiography examination was performed by a single echocardiographer (AFEL-K) at 48 h of age. Samples of blood for NTpBNP and troponin T estimation were collected at the same time points. Both the radiologist and echocardiographer were blinded to all NTpBNP and cTnT results.
NTpBNP and cTnT measurements were taken immediately following the echocardiogram. A sample of 1 ml of blood was collected in a lithium–heparin bottle, centrifuged and then plasma frozen at −20°C for batch analysis using the Elecsys 2010 quantitative assay (http://www.roche.com). The Roche test is a non-competitive chemiluminescent technology for the determination of NTpBNP in human serum and plasma. This is a two-site (sandwich) assay incorporating a polyclonal NTpBNP-specific antibody and a polyclonal NTpBNP-specific antibody labelled with a ruthenium complex. The variance is 4.6% (5.6 pmol/l) and 1.9% (107.5 pmol/l), respectively, for low- and high-concentration patient samples, and the respective day-to-day variance is 5.5% (6.4 pmol/l) and 2.6% (113.6 pmol/l).21 The cTnT assay is an electrochemiluminescent sandwich ELISA has a lower limit of detection of 0.01 pg/l, with minimal cross reactivity with cardiac troponin I (0.002%) and skeletal troponin T (0.001%). This third generation assay is unaffected by bilirubin levels up to exchange values, sample haemolysis or renal insufficiency. The repeatability co-efficient for a paired sample is 10%, and the variability co-efficient for precision analysis is 6.4% (Anon 1999, Troponin T STAT data sheet; Roche Diagnostics, http://www.roche.com).
All studies were performed at 48 h of age by a single sonographer (AFEL-K) using the Siemens Acuson Sequoia Ultrasound machine (Southern California) and a 10v4 cardiology multi-frequency probe. Two-dimensional, M-mode, pulse and colour flow Doppler imaging were performed. A PDA scoring system was devised based on six established criteria which are determinants of haemodynamic significance. Haemodynamic significance is based on echocardiographic evidence of pulmonary overcirculation: a high LA:Ao, mitral valve E wave to A wave ratio greater than 1, and a rising left ventricular output (LVO). In addition, a falling celiac artery blood flow (CAF) despite a rising LVO and absent or reversed end diastolic flow in the descending aorta signify systemic hypoperfusion.4 22,–,24 Each abnormal measurement was allocated 1 mark if the value was deemed significant in the setting of a PDA. A non-significant value was given 0 marks. The maximum score possible was 6. The following echocardiography variables were determined in each study. One mark was allocated to each of the following:
E. Mitral valve early phase to atrial phase velocity ratio (E:A). A number equal to or greater than 1 signifies increased pulmonary venous return and left atrial pressure loading.24
F. CAF to LVO ratio (CAF:LVO): normal CAF is 20% of LVO. A reduced ratio of 15% or less indicates increased pulmonary venous return and reduced systemic blood flow.4
Treatment of a PDA was at discretion of the consultant on service. Treatment commenced on or after day 3 of life. The treatment course was comprised of three doses of ibuprofen 24 h apart (10 mg/kg, followed by two doses of 5 mg/kg). Prophylactic indomethacin was not used during the study period.
Developmental assessment including the Mental Developmental Index and the Psychomotor Developmental Index of the Bayley Scales of Infant Development II was carried out at 24 months corrected age by a single psychologist (MS). Infants were classified into severe and mild disabilities in addition to a normal group according to their cognitive score. Severe Disability was defined as a score <70. Mild disability was defined as a score between 71 and 85. Normal development was defined as a score greater than 85.
Statistical analysis was performed using SPSS 16.0 for Windows. The cohort was divided into two groups: infants with normal neurodevelopmental outcome at 2 years of age or mild disability (termed Normal/Mild disability group) and infants with Severe Disability at 2 years or death before discharge (termed Severe Disability/Death group). There were only three infants in the mild disability category. These infants had similar clinical and biochemical markers as the normal infants. A separate analysis was performed between the two groups with the infants that died excluded. The echocardiographic and plasma NTpBNP/cTnT measurements were non-parametric continuous variables. Medians were compared using the Mann–Whitney U test for non-parametric data. The Kruskal–Wallis test was used for multiple groups, and the χ2 test was used for categorical data. Correlation between the studied variables and the absolute cognitive score in the survivors was also performed. Results were expressed as median inter-quartile range (IQR) unless stated otherwise. We considered a p value of <0.05 as significant.
Eighty infants were enrolled in the original study of echocardiographic findings in preterm infants including infants less than or equal to 1500 g and/or <32 weeks gestation.18 Seven infants were excluded as they had a birth weight of 1500 g and no developmental follow-up was performed. Seventy-three infants were followed prospectively during the study period. Five infants did not have complete cTnT and NTpBNP levels. Five families refused attendance and three were lost to follow-up. These infants had a median gestation of 28.7 weeks (26.1–30.2) and a median birth weight of 1.10 kg (0.910–1.37). A total of 60 infants were enrolled in this study. Twelve infants died before discharge and 48 infants completed their neurodevelopmental assessments at 24 months corrected age. The median gestational age of the cohort was 27.7 weeks (26.2–29.4); the median birth weight was 1.01 kg (0.86–1.22). Forty infants had a normal neurodevelopmental outcome. Three infants had Mild Disability and five infants had Severe Disability. The infants were divided into two groups: Normal/Mild disability and Severe Disability/Death. Infants in the Severe Disability/Death group have a statistically significant lower birth weight and gestational age (table 1). Seventy-two per cent of infants in the Normal/Mild Disability group received a complete course of antenatal steroids compared 41% of the Severe Disability/Death group (p=0.03). There was no significant difference in the Apgar scores, and the mode of delivery.
There was no statistically significant difference between the groups in the use of inotropes, RDS, BPD or NEC (table 2). A higher incidence of severe IVH was noted in the Severe Disability/Death group (71% vs 9%, p < 0.001).
At 48 h NTpBNP was higher in the Severe Disability/Death group compared to the Normal/Mild Disability group (table 3, figure 1). The difference in NTpBNP level between infants with mild disability and normal infants was not statistically significant. Similarly, cTnT levels at 48 h were significantly higher in the Severe Disability/Death group compared to the two other groups (table 3, figure 1). There was no difference in the cTnT levels between infants with mild disability and normal infants. Infants in the Severe Disability/Death group had a higher PDA score compared to infants with mild disability and normal outcome (5 vs 2, p<0.001, table 3). There was no difference in the scores between infants with mild disability and normal infants. There was a negative correlation between the studies markers and the absolute cognitive score in the surviving infants. This did not reach statistical significance.
We compared the biomarkers and the PDA scoring system results between the surviving infants with Severe Disability (excluding the deaths, n=5) and the Normal/Mild Disability group. Surviving infant with Severe Disability had a lower gestational age (25.8 (25.5–26.8) weeks vs 28.7 (27.2–29.7) weeks, p=0.028). Survivors with Severe Disability had a lower birth weight (0.980 (0.860–1.000) kg vs 1.120 (0.910–1.250) kg), however, this did not reach statistical significance. The difference in the PDA score between the Severe Disability infants compared to the Normal/Mild Disability infants remained significant (5 (3–5) vs 2 (1–4), p=0.032). cTnT and NTpBNP levels were both higher in the survivors with disability. This did not reach statistical significance (0.54 (0.52–2.48) µg/l vs 0.20 (0.12–0.46) µg/l, p=0.056, and 5344 (2098–17048) pmol/l vs 2107 (594–5798) pmol/l, p=0.095, respectively).
In this study, we demonstrated a possible association between cTnT, NTpBNP, a haemodynamically significant PDA and poor neurodevelopmental outcome or death in a cohort of preterm infants. There was no statistical difference in cTnT, NTpBNP, or PDA score between the mild disability and the normal outcome groups. This may have resulted from the small number of infants (n=3) in the mild disability range in our cohort. In addition, there seems to be a relationship between the PDA scoring system and Severe Disability at 2 years of age in the surviving infants.
Left to right shunting across a PDA is associated with systemic hypoperfusion. Echocardiography assessment of the celiac and superior mesenteric arteries demonstrates reduced blood flow in the presence of a PDA compared to controls.4 28 29 This may explain the association between a PDA and NEC. Similarly, cerebral perfusion, assessed using near-infrared spectroscopy is reduced in the presence of haemodynamically significant ductus.3 In addition, cranial Doppler ultrasound demonstrated reduced cerebral perfusion in the presence of a PDA. This effect on the cerebral circulation of the vulnerable preterm brain may explain the association between PDA and IVH, and subsequently poor neurodevelopmental outcome. We have demonstrated in our cohort that a significant PDA with objective evidence of systemic hypoperfusion is associated with severe IVH and poor neurodevelopmental outcome at 2 years of age. This association warrants further study. Neuro-imaging modalities such as MRI may enhance our understanding of any putative link between the presence of a PDA and poor neurodevelopmental outcome.
The use of biochemical markers to assess the impact of the PDA on the myocardium is gaining interest. These markers are potentially useful for neonatologists working in centres with poor access to echocardiography. NTpBNP is the inactive by-product of BNP, which is released by the ventricular myocardium in response to volume and pressure loading. Although it is an inactive precursor of BNP, it is produced in equi-molar quantities and has a more stable half life.14 cTnT is released following acute myocardial damage. cTnT assays are not thought to cross react with skeletal forms in the preterm infant.13 These biochemical markers appear uninfluenced by antenatal factors and unrelated to birth weight and gestational age.15 These properties render them ideal markers of myocardial damage and the haemodynamic status in preterm infants. NTpBNP levels rise in the presence of a haemodynamically significant PDA and fall following successful treatment. NTpBNP levels correlate with echocardiographic markers of ductal significance.17 The magnitude of rise is also associated with the development of severe IVH and/or death before discharge. NTpBNP yields an area under the curve (AUC) of 0.84 (95% CI 0.72 to 0.96, p<0.001) and a level of 5500 pmol/l had a sensitivity of 80% and a specificity of 80% for severe IVH or death before discharge.18 Similarly, cTnT displays a similar pattern with the presence of a PDA and following treatment.11 For the development of severe IVH or death before discharge, the AUC for cTnT was 0.81 (95% CI 0.69 to 0.92, p=0.001), with a level of 0.42 µg/l having a sensitivity of 70% and a specificity of 75%.
NTpBNP is associated with an unfavourable outcome following stroke in adults despite a normal ejection fraction.30 31 A similar association exists between stroke and troponin T, with elevated levels being an independent predictor of mortality.32 The release of catecholamines associated with adult stroke may lead to myocardial stress and foci of necrosis resulting in a rise in cTnT and NTpBNP after the occurrence of the event.33 34 The mechanism of the associated rise in this preterm population may be different. The presence of a PDA is known to affect both systemic perfusion as well as coronary artery perfusion.12 The rise of cTnT may reflect this phenomenon and in turn identify infants at risk of poor cerebral perfusion.
Twenty-four per cent of infants in the poor outcome group underwent a PDA ligation compared to 2% in the Normal/Mild Disability group. PDA ligation in premature infants has been shown to have low surgical morbidity,35 but there is substantial late mortality and a high incidence of morbidity in the survivors.36 Low cardiac output syndrome is now a recognised postoperative complication with potentially harmful effects. After PDA ligation, there is an initial acute deterioration followed by an improvement in global cardiac function.37 Recently, PDA ligation has been associated with poorer neurosensory outcomes compared with medical treatment alone.36
This study has some limitations. The number of infants in the mild and Severe Disability groups is small. In addition, the rise of cTnT and NTpBNP may have been influenced by other confounders unaccounted for in this small cohort. This association needs to be explored in a large number of infants. In addition, there are many confounders that may have influenced the biomarker levels that also have an impact on neurodevelopmental outcome. Although no statistical differences in the antenatal and many of the postnatal parameters existed between the groups (tables 1 and 2), this may have resulted from the small number of infants leading to a type II error. In addition, due to the small event rate in terms of severe neurodisability, constructing a multi-linear regression analysis model was not feasible.
The use of these markers in clinical practise is not established yet. They may be of benefit in screening preterm infants with a PDA at 48 h of life and using the levels to provide a more targeted approach to PDA treatment. Treating a PDA based on the potential to develop adverse outcomes needs to be explored further. A larger cohort of infants needs to be examined to confirm these associations.
There is no consensus regarding PDA treatment of preterm infants. Medical therapy for PDA has well recognised adverse effects and neither prophylaxis nor treatment on the basis of clinical and limited echocardiographic signs has been shown to improve long-term outcomes. The use of cTnT, NTpBNP in combination with echocardiography assessment should be studied in a trial examining the short-term and long-term effects of using this diagnostic model in targeted PDA treatment.
Funding Funded by the National Maternity Hospital Research Fund.
Competing interests None.
Ethics approval This study was conducted with the approval of the National Maternity Hospital Ethics Committee.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.