|Year : 2021 | Volume
| Issue : 1 | Page : 11-18
Comparison of perfusion index and echocardiographic parameters in preterm infants with hemodynamically significant patent ductus arteriosus
Melek Buyukeren1, Şule Yiğit1, Hayrettin Hakan Aykan2, Tevfik Karagöz2, Hasan Tolga Çelik1, Murat Yurdakök1
1 Department of Pediatrics, Division of Neonatology, Faculty of Medicine, Hacettepe University, Ankara, Turkey
2 Department of Pediatrics, Division of Cardiology, Faculty of Medicine, Hacettepe University, Ankara, Turkey
|Date of Submission||03-Jun-2020|
|Date of Decision||30-Dec-2020|
|Date of Acceptance||04-Jan-2021|
|Date of Web Publication||08-Feb-2021|
Dr. Melek Buyukeren
Department of Pediatrics, Division of Neonatology, Faculty of Medicine, Hacettepe University, Sihhiye, 06100, Altindag, Ankara
Source of Support: None, Conflict of Interest: None
Background/Aim: The aim of the study was to compare echocardiographic parameters and the perfusion index (PI) and plethysmographic variability index (PVI) values obtained by routine pulse oximetry in the diagnosis of hemodynamically significant patent ductus arteriosus (hsPDA). Materials and Methods: This prospective study was conducted between 2016 and 2017 at the HacettepeUniversity Neonatal Intensive Care Unit. The study included premature neonates who had a birth weight below 1500 g. Patients were routinely monitored from the right wrist and right foot using a pulse oximeter (Masimo Radical-7® Pulse CO-Oximetry), and PI and PVI values were recorded. The difference between right-hand and right-leg PI values was calculated as the delta PI (ΔPI). A cardiologist blinded to the results evaluated the presence of patent ductus arteriosus (PDA) with echocardiography on postnatal days 1th, 3rd, and 7th. Results: Of the 66 preterm neonates included in the study, 23 had hsPDA. On postnatal day 1, the hsPDA group had a significantly greater ductal diameter, PDA/left pulmonary artery (LPA) ratio, and left ventricle (LA)/aortic (Ao) ratio (P < 0.05). On day 7, the hsPDA group had a significantly higher ductal velocity, PDA/LPA ratio, LA/Ao ratio, antegrade PA and LPA diastolic flow, and LV/Ao ratio (P < 0.05). In hsPDA group, the median ΔPI values were 0.85 (25–75 interquartile range [IQR]; 0.62–1.15) on day 1; 1.03 (25–75 IQR; 0.85–1.26) on day 3; and 0.89 (25–75 IQR; 0.64–1.22) on day 7. The median (25–75 IQR) ΔPI values were higher in the hsPDA group than in the non-hsPDA group on postnatal days 1, 3, and 7 (P < 0.001, P < 0.001, and P < 0.001, respectively). The ΔPI cutoff values for the diagnosis of hsPDA were 0.47 on day 1 (91.3% specificity; 90.5% sensitivity), 0.41 on day 3 (100% specificity; 97.3% sensitivity), and 0.47 on day 7 (90% specificity; 100% sensitivity). Conclusions: Our study shows that the difference between PI values (ΔPI) in the right hand and right leg obtained by pulse oximetry has diagnostic value in hsPDA and can assist diagnosis when echocardiography is not available.
Keywords: Newborn, patent ductus arteriosus, perfusion index, plethysmographic variability index, pulse oximeter
|How to cite this article:|
Buyukeren M, Yiğit &, Aykan HH, Karagöz T, Çelik HT, Yurdakök M. Comparison of perfusion index and echocardiographic parameters in preterm infants with hemodynamically significant patent ductus arteriosus. J Clin Neonatol 2021;10:11-8
|How to cite this URL:|
Buyukeren M, Yiğit &, Aykan HH, Karagöz T, Çelik HT, Yurdakök M. Comparison of perfusion index and echocardiographic parameters in preterm infants with hemodynamically significant patent ductus arteriosus. J Clin Neonatol [serial online] 2021 [cited 2021 Nov 28];10:11-8. Available from: https://www.jcnonweb.com/text.asp?2021/10/1/11/308847
| Introduction|| |
In preterm neonates, failure of ductus arteriosus closure (patent ductus arteriosus [PDA]) can result in substantial blood flow from the aorta to the pulmonary artery (PA). This is referred to as hemodynamically significant PDA (hsPDA). Clinical signs suggestive of hsPDA include the development of cardiac failure in a previously stable preterm neonate, systemic hypoperfusion, and a new or increased need for respiratory support. There are various systemic signs and symptoms, primarily manifesting in the heart, lungs, and gastrointestinal system. Heart failure, pulmonary hemorrhage, pulmonary edema, intracranial hemorrhage (ICH), bronchopulmonary dysplasia (BPD), and necrotizing enterocolitis (NEC) may occur.,
The incidence of PDA in premature infants increases as smaller gestational age and birth weight decrease. It occurs in about 30% of neonates with a birth weight below 1500 g and 70%–75% of those born before 28 weeks of gestation.
Diagnosis of hsPDA is based on clinical findings accompanied by evidence of high ductal flow volume in echocardiography. However, an experienced pediatric cardiologist or echocardiography equipment may not always be available in a medical center.
The perfusion index (PI) and plethysmographic variability index (PVI) are relatively new parameters that are noninvasively and automatically recorded by the pulse oximetry devices routinely used to monitor seriously ill infants in a neonatal intensive care unit (NICU).,, The PI is a ratio of the pulsatile to nonpulsatile signals of arterial blood flow and provides peripheral perfusion monitoring. A nonpulsatile signal is generated by static blood flow, skin, and other tissues. The PVI is a measure of the dynamic changes in the PI during a complete respiratory cycle. It may be affected by clinical deterioration, hypotension, volume insufficiency, or other parameters such as ventilation.
The aim of this study was to compare echocardiographic parameters and PI and PVI values obtained by routine pulse oximetry in the diagnosis of hsPDA.
| Materials and Methods|| |
This prospective study was conducted between November 2016 and December 2017 at the Hacettepe University NICU after obtaining approval from the …… University Clinical Research Ethics Committee (GO 14/108-17).
The study included premature neonates who had a birth weight <1500 g and met the study criteria. On postnatal days 1, 3, and 7, a pediatric cardiologist who was unaware of the patients' pulse oximetry results, conducted bedside standard two-dimensional, M-mode, pulsed and color flow Doppler imaging using an echocardiography device (Acuson X300, Siemens Medical Solutions USA, Inc.) with a 10-Hz probe to obtain anatomic and functional measurements.
We employed the echocardiographic markers used by Sehgal et al. in the diagnosis and staging of hsPDA. The hsPDA group comprised neonates with consistent clinical signs and most of the following echocardiographic findings: two-dimensional transductal diameter ≥1.5 mm in the short axis view; left ventricle/aortic (LV/Ao) ratio ≥2.1 in the long axis view; a PDA/left PA (LPA) ratio ≥0.5; an LA/Ao ratio ≥1.4 in the long-axis view; peak ductal velocity ≤2 m/s at the pulmonary end of the duct (measured by pulse-wave Doppler); an antegrade diastolic flow in the main PA >20 cm/s and antegrade diastolic flow in the LPA >30 cm/s (measured by pulse-wave Doppler); an E/A wave ratio ≥1 (measured by transmitral Doppler); and an isovolumic relaxation time (IVRT) ≤45 ms between the mitral and aortic valves (measured by pulse-wave Doppler). We defined hsPDA group by most of echocardiographic marker.
Patients were routinely monitored from the right wrist and right foot using a pulse oximeter (Masimo Radical-7® Pulse CO-Oximetry, Masimo Corp., Irvine, CA, USA) throughout their stay in the NICU. The PI and PVI recordings were obtained by the same neonatal specialist when the neonates were still, in a propped position, and there were no artifacts in the plethysmography pulse wave. Measurements were made 1–3 h before echocardiography. The pulse oximeter signals were recorded for 10 min, transferred to a personal computer, and analyzed using VitaWin 3 (Telematya, Germany, Teltow, Germany) software, which was designed for this system (installed by Masimo technicians). PI value was calculated by the average of several readings. The study involved no invasive interventions or blood collection and the neonates' routine follow-ups in the NICU continued.
The PI was calculated at 6 s intervals as the ratio of the pulsatile component (AC) to the nonpulsatile component (DC) using the following formula: PI = (AC/DC) × 100. The PVI was calculated using the following formula: PVI = ([PIMax–PIMin]/PIMax) × 100. The difference between right-hand and right-leg PI values was calculated as the delta PI (ΔPI = PIright hand– PIright leg).
Infants with suspected early neonatal sepsis, a history of premature rupture of membranes or chorioamnionitis, a need for inotropic medication, hemodynamic instability, hypoxic–ischemic encephalopathy, NEC (Stage 2 or higher), intraventricular hemorrhage (Stage 2 or higher), metabolic disease, major congenital anomaly, and chromosomal disease were excluded from the study. Infants who died within the first 7 days of life and those whose parents did not provide consent were also excluded from the study.
Statistical analyses were done using Statistical Package for the Social Sciences (SPSS) 22.0 (IBM Corp., Armonk, NY, USA). Data showing normal distribution were evaluated with an independent samples t-test, while variables without normal distribution were analyzed using the nonparametric Mann–Whitney U-test. Categorical variables were analyzed using Fisher's exact and Pearson Chi-square tests. Results were expressed as mean ± standard deviation for normal distributions or median and 25th–75th percentile interquartile range (IQR). In addition, a receiver operating characteristic curve analysis was done to determine the cutoff value for ΔPI in the hsPDA diagnosis. P < 0.05 was considered statistically significant.
| Results|| |
Seventy-eight patients were initially included. Of these, seven died within the first postnatal week, 4 were excluded because exact PVI measurements could not be obtained, and one was excluded due to a diagnosis of congenital heart disease. The study was completed with the remaining 66 patients. Demographic data, neonatal characteristics, and study parameters were recorded and analyzed.
Of the 66 preterm neonates included in the study, 23 had hsPDA, and 43 did not. The demographic characteristics of the patients are shown in [Table 1].
There was no difference between the groups in terms of gender or birth weight (P = 0.450 and P= 0.135, respectively), whereas the gestational age was lower in the hsPDA group (P < 0.001). Neonatal resuscitation, respiratory distress syndrome (RDS), NEC, and BPD were significantly more common in the hsPDA group than in the non-hsPDA group (P = 0.021, P < 0.001, P= 0.038, and P < 0.001, respectively). There were no statistical differences between the study groups in terms of small for gestational age prevalence, delivery method, use of assisted conception techniques, or frequency of ICH (P = 0.172, P= 1.000, P= 0.015, and P= 0.172, respectively). In addition, 5-min Apgar scores were lower in the hsPDA group (P = 0.007), while hospital stays were significantly longer in this group (P = 0.003).
The mortality rate was similar in both groups (P = 0.547). There was no statistical difference between hemoglobin or hematocrit values in the groups (P = 0.711 and P= 0.909, respectively). Pulmonary hemorrhage was observed in two patients in the hsPDA group. All infants received fluid therapy according to NICU protocol depending on birth weight and postnatal day. Infants received 70–80 cc/kg fluid on 1st day of life, 90–110 cc/kg on 2nd day of life, and 130–150 cc/kg on 3rd day of life and the other days. Fluid management changed depending on daily measurement of serum sodium levels and weight of infants.
Echocardiography parameters and PI and PVI values are shown in [Table 2]. On postnatal day 1, the hsPDA group had a significantly greater ductal diameter, PDA/LPA ratio, LA/Ao ratio (P = 0.04, P= 0.001, and P= 0.007, respectively) and a statistically equivalent ductal velocity, antegrade PA and LPA diastolic flow, E/A ratio, IVRT, and LV/Ao ratio (P = 0.458, P= 0.120, P= 0.206, P= 0.149, P= 0.225, and P= 0.087, respectively).
On day 3, the hsPDA group had a significantly higher ductal diameter, ductal velocity, PDA/LPA ratio, LA/Ao ratio, and antegrade PA and LPA diastolic flow (P < 0.001, P= 0.042, P < 0.001, P < 0.001, P < 0.001, and P= 0.004, respectively), while the IVRT was lower (P = 0.003). The E/A ratio and LV/Ao ratio were similar between the groups (P = 0.703 and P= 0.064, respectively).
On day 7, the hsPDA group had a significantly higher ductal velocity, PDA/LPA ratio, LA/Ao ratio, antegrade PA and LPA diastolic flow, and LV/Ao ratio (P < 0.001, P < 0.001, P < 0.001, P < 0.001, P < 0.001, P < 0.001, and P= 0.026, respectively). There was no difference in the E/A ratio or IVRT (P = 0.118 and P= 0.078, respectively).
PVI values were significantly higher in the hsPDA group than in the non-hsPDA group on postnatal days 1 and 7 (P < 0.05), but there was no statistically significant difference between the groups on day 3 (right hand: P= 0.197; right foot: P= 0.214). The median (25–75 IQR) ΔPI values were higher in the hsPDA group than in the non-hsPDA group on postnatal days 1, 3, and 7 (P < 0.001, P < 0.001, and P < 0.001, respectively). The cutoff values for ΔPI in hsPDA are shown in [Figure 1], [Figure 2], [Figure 3]. The ΔPI cutoff values for the diagnosis of hsPDA were 0.47 on day 1 (91.3% specificity and 90.5% sensitivity), 0.41 on day 3 (100% specificity and 97.3% sensitivity), and 0.47 on day 7 (90% specificity and 100% sensitivity).
A positive correlation was found between ΔPI values on the 1st day and ductal diameter and LPA diastolic flow in 23 patients with hemodynamically significant PDA; no correlation was detected in other echocardiographic parameters. The strongest correlation was found between LPA diastolic flow and ΔPI. A positive correlation was found between PI values and echocardiography parameters on the 3rd and 7th days, with the strongest correlation between ductal diameter and ΔPI. There was no correlation between 1st and 3rd day PVI values (both right hand and right foot) and echocardiography measurements, and a positive correlation was detected between PVI (both right hand and right foot) data on day 7 and ductal diameter and PDA/LPA ratios [Table 3].
|Table 3: Correlation analysis of pulse oximetry and echocardiography data in the hemodynamically significant patent ductus arteriosus (+) group|
Click here to view
Patients who had hsPDA on the 1st day of life but whose PDA measurements regressed in the following days were evaluated in the hsPDA group. There were three patients with this feature in the hsPDA group whose PDA spontaneously closed.
During the follow-up, 3 (13%) neonates with hsPDA had spontaneous closure with no treatment. The remaining twenty patients were treated with a course of intravenous ibuprofen (10 mg/kg/dose on day 1; 5 mg/kg/dose on days 2 and 3) during the postnatal 1st week. Nine of those patients exhibited ductal closure after this medical treatment. Seven patients received a second course of intravenous ibuprofen, which resulted in ductal closure in three patients. The remaining eight patients (34%) underwent duct ligation. Statistical analysis could not be performed because the number of patients who underwent ductal ligation was low (n = 8) [Table 4]. The gestational weeks of the patients who underwent surgical ligation therapy (median [25–75p]) were 27.1 (26.1–28.0), their birth weight was 885 (805–1015) g. Only one of these patients (12.5%) died. All of these patients had both RDS and BPD. The ΔPI values of these patients were 0.82 (0.54–1.28) on day 1, 1.08 (0.82–0.32) on day 3, and 0.84 (0.64–1.11) on day 7.
|Table 4: Demographic data and pulse oximetry values of patients who underwent ductus ligation|
Click here to view
| Discussion|| |
As survival rates among preterm infants have improved, rates of hsPDA and associated morbidities have also increased. In particular, lung injury, severe BPD, and pulmonary hypertension due to hsPDA result in prolonged NICU stays and increased health costs. Although anatomic and dynamic measurements obtained with echocardiography are the gold standard in diagnosis, experienced pediatric cardiologists and echocardiography devices are not always available. Single-point and continuous measurements obtained by the pulse oximetry devices routinely used in NICUs for noninvasive monitoring may facilitate diagnosis., Our study shows that the difference between PI values (ΔPI) in the right hand and right leg obtained by pulse oximetry has diagnostic value in hsPDA and can assist diagnosis when echocardiography is not available.
In some studies, no significant correlation was observed between hsPDA and postductal PI values, while others reported that postductal PI values were higher in infants with hsPDA. Vidal et al. investigated the relationship between postductal PI values and PDA but found that it was not significant. In a study of 342 neonates born at <32 weeks of gestation, Alderliesten et al. reported that postductal PI values were higher in the hsPDA group than in the non-hsPDA group. Gomez-Pomar et al. observed that postductal PI values were higher than preductal PI values in the hsPDA group, but echocardiography and pulse oximetry measurements could not always be taken consecutively. For their first 4 postnatal days, Terek et al. evaluated 42 neonates who were born at <36 weeks of gestation; the authors determined that PI values were significantly higher in neonates with hsPDA compared to those without. As in most previous studies, the PI values were higher in the PDA group in our study. Alderliesten et al. speculated that this was due to an increase in the AC component in the formula used to calculate PI associated with greater pulse pressure and hyperdynamic circulation. Despite low systemic perfusion in the HsDPA group, high PI values could be explained by reverse diastolic flow in the descending aorta.
In the study of Alderliesten et al., it was found that PI increased as gestational age increased. In our study, we preferred to use the ΔPI value, not the PI value, as there was a difference between the two groups in terms of gestational age.
Khositseth et al. evaluated thirty preterm neonates at <34-week gestation and observed that postductal PI values were significantly higher than preductal PI values in the hsPDA group. Similar to their study, we also observed higher ΔPI in the hsPDA group compared to the non-hsPDA group on postnatal days 1, 3, and 7 (P < 0.001, P < 0.001, and P < 0.001, respectively). Khositseth et al. determined that ΔPI at postnatal day 1, 3, and 7 was a diagnostic indicator for hsPDA above a threshold value of 1.05 (66.7% sensitivity and 100% specificity). Balla et al. evaluated 27 preterm infants at <34-week gestation on postnatal days 1 and 3 and found that ΔPI values of >0.85 on day 1 (80% sensitivity and 94% specificity) and >0.95 on day 3 (80% sensitivity and 88.2% specificity) had significant diagnostic value for hsPDA. Based on the data obtained in the present study, the ΔPI cutoff values for the diagnosis of hsPDA were 0.47 on day 1 (specificity 91.3% and sensitivity 90.5%), 0.41 on day 3 (specificity 100% and sensitivity 97.3%), and 0.47 on day 7 (specificity 90% and sensitivity 100%).
It has been reported that anemia may affect the PI. In our study, the groups showed no statistically significant differences in hemoglobin or hematocrit levels on day 1 (P = 0.711 and P= 0.909, respectively).
In a study by Vidal et al. including 45 preterm neonates with a median gestational age of 27 weeks, the PI values were not affected by the ductal flow pattern; however, the PVI value was affected by the ductal flow pattern in the growing and pulsatile PDA groups, and this value was lower. In contrast, according to the present study, the PVI values were significantly higher in the hsPDA group compared to the non-hsPDA group on days 1 and 7 (P < 0.05) and were statistically equivalent on day 3 (right hand: P= 0.197 and right foot: P= 0.214). The significant difference in the PVI values of neonates with and without hsPDA on postnatal days 1 and 7 but not on day 3 indicates that the PVI does not have adequate discriminatory power for hsPDA diagnosis.
Consistently, artifact-free noninvasive monitoring of hemodynamic parameters using a pulse oximetry device, transferring the data to a computer, and conducting statistical analyses is technically difficult and personnel dependent. To simplify and accelerate this procedure, patient-based monitoring devices should be developed that are capable of performing these processes automatically to provide a rapid and accurate flow of information to physicians. Such devices would enable closer hemodynamic monitoring and facilitate medical decisions in the NICU. There is also a need for prospective observational clinical studies with longer follow-up times and a larger series of preterm neonates of different gestational ages.
As Nitzan et al. stated in their study, a higher PI value (proportionally lower ΔPI), especially in the postductal region, may be attributed to more powerful PDA shunt during the 2nd week of life. In our study, a correlation was found between PI values and echocardiography measurements on the 3rd and 7th days [Table 3].
In conclusion, our study demonstrates that measuring the differences between right-hand and right-leg PI values (ΔPI) obtained by pulse oximetry can facilitate the diagnosis of hsPDA when echocardiography is not available. In this case, the ΔPI cutoff values for the diagnosis of hsPDA can be used 0.47 on day 1, 0.41 on day 3, and 0.47 on day 7. Our study is a prospective observational study. We believe that intraobserver variability is minimal, as echocardiography evaluations are performed by an experienced pediatric cardiologist.
| Conclusions|| |
Our study shows that the difference between PI values (ΔPI) in the right hand and right leg obtained by pulse oximetry has diagnostic value in hsPDA and can assist diagnosis when echocardiography is not available.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Rolland A, Shankar-Aguilera S, Diomandé D, Zupan-Simunek V, Boileau P. Natural evolution of patent ductus arteriosus in the extremely preterm infant. Arch Dis Child Fetal Neonatal Ed 2015;100:F55-8.
Kort EJ. Patent ductus arteriosus in the preterm infant: An update on morbidity and mortality. Curr Pediatr Rev 2016;12:98-105.
Evans N. Preterm patent ductus arteriosus: A continuing conundrum for the neonatologist? Semin Fetal Neonatal Med 2015;20:272-7.
Narayanan M, Cooper B, Weiss H, Clyman RI. Prophylactic indomethacin: Factors determining permanent ductus arteriosus closure. J Pediatr 2000;136:330-7.
Hamrick SE, Hansmann G. Patent ductus arteriosus of the preterm infant. Pediatrics 2010;125:1020-30.
Jegatheesan P, Nudelman M, Goel K, Song D, Govindaswami B. Perfusion index in healthy newborns during critical congenital heart disease screening at 24 hours: Retrospective observational study from the USA. BMJ Open 2017;7:e017580.
Workie FA, Rais-Bahrami K, Short BL. Clinical use of new-generation pulse oximeters in the neonatal intensive care unit. Am J Perinatol 2005;22:357-60.
Zaramella P, Freato F, Quaresima V, Ferrari M, Vianello A, Giongo D, et al.
Foot pulse oximeter perfusion index correlates with calf muscle perfusion measured by near-infrared spectroscopy in healthy neonates. J Perinatol 2005;25:417-22.
Lima A, Bakker J. Noninvasive monitoring of peripheral perfusion. Intensive Care Med 2005;31:1316-26.
Cannesson M, Desebbe O, Rosamel P, Delannoy B, Robin J, Bastien O, et al
. Pleth variability index to monitor the respiratory variations in the pulse oximeter plethysmographic waveform amplitude and predict fluid responsiveness in the operating theatre. Br J Anaesth 2008;101:200-6.
Sehgal A, Paul E, Menahem S. Functional echocardiography in staging for ductal disease severity: Role in predicting outcomes. Eur J Pediatr 2013;172:179-84.
Khositseth A, Muangyod N, Nuntnarumit P. Perfusion index as a diagnostic tool for patent ductus arteriosus in preterm infants. Neonatology 2013;104:250-4.
Vidal M, Ferragu F, Durand S, Baleine J, Batista-Novais AR, Cambonie G. Perfusion index and its dynamic changes in preterm neonates with patent ductus arteriosus. Acta Paediatr 2013;102:373-8.
Alderliesten T, Lemmers PM, Baerts W, Groenendaal F, van Bel F. Perfusion index in preterm infants during the first 3 days of life: Reference values and relation with clinical variables. Neonatology 2015;107:258-65.
Gomez-Pomar E, Makhoul M, Westgate PM, Ibonia KT, Patwardhan A, Giannone PJ, et al
. Relationship between perfusion index and patent ductus arteriosus in preterm infants. Pediatr Res 2017;81:775-9.
Terek D, Yalaz M, Ulger Z, Koroglu OA, Kultursay N. Medical closure of patent ductus arteriosus does not reduce mortality and development of bronchopulmonary dysplasia in preterm infants. J Res Med Sci 2014;19:1074-9.
Fink D, El-Khuffash A, McNamara PJ, Nitzan I, Hammerman C. Tale of two patent ductus arteriosus severity scores: Similarities and differences. Am J Perinatol 2018;35:55-8.
Balla KC, John V, Rao Pn S, Varghese K. Perfusion index-bedside diagnosis of hemodynamically significant patent ductus arteriosus. J Trop Pediatr 2016;62:263-8.
Kanmaz HG, Sarikabadayi YU, Canpolat E, Altug N, Oguz SS, Dilmen U. Effects of red cell transfusion on cardiac output and perfusion index in preterm infants. Early Hum Dev 2013;89:683-6.
Nitzan I, Hammerman C, Fink D, Nitzan M, Koppel R, Bromiker R. The effect of patent ductus arteriosus on pre-ductal and post-ductal perfusion index in preterm neonates. Physiol Meas 2018;39:075006.
[Figure 1], [Figure 2], [Figure 3]
[Table 1], [Table 2], [Table 3], [Table 4]