WHAT IS ALREADY KNOWN ON THIS TOPICWHAT THIS STUDY ADDS

Despite a transient imbalance of biventricular loading conditions, the cardiac output remains stable throughout the early cardiovascular transition in term vigorous infants receiving delayed cord clamping, especially on the left side, which accounts for cerebral perfusion.

The variation of left ventricular output across the cord clamping is independently influenced by the establishment of left ventricular preload.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

Strategies that aim to optimise the left ventricular preload during the early phases of postnatal life, such as establishing ventilation before clamping the umbilical cord, may prevent or mitigate complications of perinatal haemodynamic imbalance.

Introduction

The transition from fetal to neonatal circulation is a complex physiological phenomenon. Cardiorespiratory changes during the first seconds after birth are perhaps the most dramatic encountered in humans. In utero, the low-resistance placenta receives approximately one-fifth to one-third of the combined fetal ventricular output.1–3 Conversely, the fetal pulmonary circulation is characterised by high vascular resistance and low blood flow,4 and the majority of the right ventricular output (RVO) is diverted away from the lungs via the ductus arteriosus.5 6 Left ventricular (LV) preload is predominantly derived from the placental circulation, via the ductus venosus and the foramen ovale.7 8

The onset of breathing is critical for unimpeded fetal to neonatal transition. This triggers a rapid decline in pulmonary vascular resistance with a consequent increase in pulmonary blood flow,9 crucial for alveolar gas exchange and establishment of pulmonary venous return.10 Removal of the low-resistance placental circuit at the time of umbilical cord clamping (CC) ultimately determines the transition to the neonatal circulation. It results in simultaneous increase in systemic vascular resistance,10 and transient reduction of biventricular venous return following occlusion of the umbilical vein.

Timed delayed CC (DCC) in vigorous infants is the current international standard.11 Although some controversy exists around the timing of the intervention,12 it aims to facilitate placental blood transfusion to the infant, thereby increasing haemoglobin levels at birth and improving iron stores in the first months of life.13 14 In preterm infants, DCC is associated with lower rates of intraventricular haemorrhage,15 presumably due to smoother cardiovascular adaptation. A recent animal study in sedated and intubated lambs receiving positive pressure ventilation demonstrated haemodynamic benefits of clamping the cord after the establishment of pulmonary blood flow.10 Animal models, in particular the lamb model, have provided the foundation for our understanding of fetal-to-neonatal transition,16 although some important differences exist. Brouwer et al recently highlighted the opposite effects of breathing on the umbilical vein blood flow between lambs and humans, presumably due to interspecies anatomical variations.17 18 However, data on the immediate postnatal haemodynamic physiology around CC in humans are lacking. This study aimed to non-invasively measure cardiovascular changes associated with transition from fetal to neonatal circulation in spontaneously breathing term infants born by elective caesarean section, through repeated targeted echocardiographic assessments immediately after birth, before and after DCC.

Methods

This prospective observational study was conducted at Cork University Maternity Hospital, Ireland, a tertiary university maternity hospital, between 5 May 2022 and 6 July 2022.

Mothers of singleton pregnancies with a gestational age above or equal to 37 weeks scheduled for elective caesarean section were eligible. Infants were excluded from the study in case of (1) contraindications to DCC, (2) congenital heart defects and (3) chromosomal abnormalities.

In accordance with local guidelines, the umbilical cord was clamped at least 60 s after birth. Oxytocin was administered to the mother via continuous intravenous infusion at the time of CC. Targeted echocardiograms were performed at three different time points by the same operator (RC): before CC (T1), immediately after CC (T2) and at approximately 5 min after birth (T3) (online supplemental videos 1 and 2). The clinical team, not involved in the research project, provided standard care throughout.

Echocardiography measurements

Targeted echocardiography was performed using a Vivid E95 ultrasound machine (GE Medical) with a high-frequency neonatal probe. The scans were stored in an online archiving system. Off-line analysis was performed afterwards. Briefly, the pulsed wave Doppler measurements were obtained at the level of the aortic and pulmonary valve annuli. Pulmonary artery acceleration time (PAAT) was measured from the pulsed wave Doppler through the pulmonary valve and indexed to the right ventricle ejection time (RVET): PAATi=PAAT/RVET. PAATi is a reliable measure of RV afterload. Tissue Doppler imaging (TDI) velocities were obtained from the apical four-chamber view. Peak myocardial systolic velocity (S’), early (E’) and late (A’) diastolic velocities were measured from the average of at least three consecutive cycles. Isovolumic contraction time, isovolumic relaxation time, ejection time and heart rate (HR) were measured on the TDI recordings. Myocardial performance index (MPI) was calculated as follows: (isovolumic contraction time+isovolumic relaxation time)/ejection time. MPI is a measurement of global ventricular performance. More details on echocardiography evaluation are provided in the online supplemental material.

Statistical analysis

Categorical variables were described using frequency and percentage. Continuous variables were tested for normality with histogram illustrations and the Shapiro-Wilk test. Continuous variables were described using mean (SD) when normally distributed or median (IQR) otherwise. Normally distributed echocardiographic parameters were compared across the three time points using repeated-measures analysis of variance (ANOVA). The Greenhouse-Geisser adjusted p value was reported when the assumption of sphericity was violated. Post hoc pairwise comparisons were performed with Bonferroni adjustment. Skewed echocardiographic parameters were compared across the time points using the Friedman test followed by post hoc pairwise Dunn-Bonferroni tests. The HR trend across 15 time points (5 at each of T1, T2 and T3) was tested for linearity using repeated measurements ANOVA. The cumulative sum (CUSUM) test for structural breaks analysis was conducted on the progression of mean HR across the 15 time points. Forward stepwise multiple linear regression analyses were performed to investigate variables associated with variation (Δ) of LVO and RVO across the first two time points (T1–T2), expressed as the difference between the values before and after CC. The variables were selected a priori, based on physiological principles. Variables were inserted in the model if p was <0.05. For the ΔRVO, the following variables were chosen: ΔS’, ΔA’ and ΔMPI measured at the lateral tricuspid annulus; and ΔPAATi. For the ΔLVO, the following variables were chosen: ΔS’, ΔA’ and ΔMPI measured at the lateral and septal mitral annuli. The TDI-derived E’ waves measured on tricuspid and mitral annuli were excluded from the analysis, as the diastolic waves often were fused due to tachycardia. Statistical analysis was performed using IBM SPSS Statistics V.28 and Prism V.9. All tests were two-tailed, and a p value of <0.05 was considered statistically significant. More details on sample size calculations are provided in the online supplemental material.

Results

The clinical characteristics of infants included in the analysis are outlined in online supplemental table S1 and reported in the online supplemental material. Forty-six infants were included in the analysis (figure 1). Echocardiography measurements are outlined in table 1 and figure 2. HR progressively decreased after birth, following a linear trend (p<0.001) (figure 3). The CUSUM test revealed a structural break in this trend after CC (p=0.013). RVO decreased initially after CC and then stabilised at T3. The increase of RV stroke volume between T2 and T3 (mean (SD): 1.67 (0.30) mL/kg and 1.76 (0.31) mL/kg, respectively), although not statistically significant, contributed to the stabilisation of RVO, despite the decreasing HR. In contrast to this, the LVO significantly increased in the first 5 min after birth (mean (SD): 239.7 (33.6) mL/kg/min at T3 vs 222.4 (32.5) mL/kg/min at T1, p=0.01), due to the augmentation in the LV stroke volume, increasing significantly at each echo time point.

Figure 1

Study flow chart.

Figure 2

Non-invasive haemodynamic monitoring in transitioning term vigorous infants. Evolution of echocardiography measurements across the study time points: before cord clamping (T1), after cord clamping (T2) and at 5 min after birth (T3). The dark red line joins the mean values across the time points, while the 95% CIs are connected by the upper and lower line. HR, heart rate; IVS, interventricular septum; LVO, left ventricular output; LVSV, left ventricular stroke volume; LW, lateral wall; MPI, myocardial performance index; PAATi, pulmonary artery acceleration time indexed to right ventricle ejection time; RVO, right ventricular output; RVSV, right ventricular stroke volume; RV, right ventricle. *P<0.05, **p<0.001.

Figure 3

Heart rate (HR) trend after the birth of term vigorous infants and the influence of cord clamping. Detailed evolution of HR over the study time points: before cord clamping (T1), after cord clamping (T2) and at 5 min after birth (T3). The dark red line joins the mean values across the time points, while the 95% CIs were displayed with the error bars.

Table 1

Echocardiographic measurements during three time points: before cord clamping (T1), after cord clamping (T2) and at 5 min after birth (T3), n=46

There was a reduction in PAATi from 0.26 (0.04) to 0.21 (0.04) following CC (p<0.001), which recovered to basal values at T3 (0.29 (0.07)). A significant decline occurred in the left and right ventricular TDI systolic velocities immediately after CC, improving at 5 min after birth. A similar trend in TDI diastolic velocities of the tricuspid and septal mitral annuli was observed, while the lateral mitral annulus was not affected by CC (p=0.054). CC was associated with a transient and ubiquitous prolongation of the isovolumic contraction and relaxation times, which returned to baselines at T3. These variations resulted in a peak of left and right ventricular MPI after CC, decreasing at T3. The stepwise regression revealed a positive correlation between ΔRVO and ΔPAATi (p=0.02, r=0.34), and between ΔLVO and ΔA’ of the septal mitral annulus (p=0.03, r=0.32) (figure 4).

Figure 4

Association between variations of cardiac output across the cord clamping and variation of loading conditions over the same period. (A) Positive relationship between right ventricular output (RVO) variation across the cord clamping and the variation of pulmonary artery acceleration time indexed to ejection time (PAATi) over the same time points. The stepwise regression model: ΔRVO=13.29+402.51×ΔPAATi. (B) Positive relationship between variation of left ventricular output (LVO) across the cord clamping and the variation of septal mitral annulus diastolic late velocity (A’ interventricular septum (IVS)) over the same time points. The stepwise regression model: ΔLVO=−13.57+6.48×ΔA’. Continuous lines represent the regression equation, while the dotted lines represent the 95% CIs.

Discussion

This study describes haemodynamic changes that occur immediately following birth in spontaneously breathing term infants delivered by caesarean section. This is the first description of cardiovascular transition immediately after birth and details haemodynamic alterations following CC in human neonates. We observed that (1) HR progressively decreases after birth, following a linear trend transiently interrupted by CC; (2) LVO and LV stroke volume increase throughout the first minutes after birth, while RVO and RV stroke volume fall after CC; (3) RV afterload rises after CC, significantly decreasing in the following minutes; (4) TDI-derived myocardial velocities decline after CC, recovering at 5 min after birth; (5) global ventricular performance temporarily falls after CC; (6) the variation of LVO across CC is directly proportional to the variation of the septal mitral annulus A’ velocity (as a marker of LV preload) in the same timespan and (7) the variation of RVO across CC is directly proportional to the variation of the RV afterload.

Our description of HR changes immediately after birth is somewhat novel. We determined the HR from the direct visualisation of the cardiac cycles on the Doppler ultrasound, as normally performed in fetal echocardiography.6 Unlike others, we consistently observed an HR peak in the first minute (at the first echocardiography assessment), with subsequent reduction thereafter.19 Another novel finding was the impact of CC on HR, which caused a brief and transient HR increase. We speculate that this is an indirect sign of an acute reduction in biventricular preload following CC, triggering a temporary response to pseudo-hypovolemia.

We observed increasing LVO during the early haemodynamic transition, mainly determined by the significant and constant LV stroke volume increase. This is consistent with the report by van Vonderen et al,20 in which they assessed LV function with echocardiography at 2, 5 and 10 min after birth in term neonates delivered by caesarean section and receiving DCC (30–60 s). Similar results were shown by Katheria et al,21 who used electrical cardiometry to measure the changes in HR and cardiac output during DCC (5 min) in a cohort of term infants. These consistent findings underlie the crucial role of increasing LV stroke volume during the cardiovascular transition to neonatal circulation. RVO was negatively affected by CC, primarily due to reduced RV stroke volume coupled with a decreasing HR. Our findings on the trends of HR and ventricular outputs over the early haemodynamic transition and following CC are consistent with the physiological data already derived from previous animal work on transition.10

PAATi has previously been validated as a reliable non-invasive marker of RV afterload.22 Despite the expected fall in pulmonary vascular resistance due to lung expansion and aeration, we observed a transient increase in RV afterload immediately after CC, subsequently decreasing at 5 min after birth. In term infants, the direction of flow across the ductus arteriosus immediately after birth has been shown to resemble that of fetal circulation.23 Indeed, the ductal diameter may even approximate that of the descending aorta in the first minutes of postnatal life.23 The increased RV afterload is most likely secondary to transductal propagation of the abrupt increase in systemic vascular resistance produced by clamping the umbilical arteries. This is the first report showing the propagation of afterload oscillation produced by CC from the systemic to the pulmonary circulations, suggesting a close interplay between the RV and placental circulation in postnatal transition. We observed a significant drop in the RV afterload at T3, consistent with progressively decreasing pulmonary vascular resistance with ongoing lung aeration and enhanced pulmonary blood flow as suggested by progressively increasing LV stroke volume during this time span.

TDI velocities particularly reflect acute changes in preload.24 25 These acutely declined in both ventricles following CC and recovered by 5 min after birth, illustrating that, even with established ventilation, biventricular preload is negatively affected by CC and suggests that RV and LV preload are still supported by placental circulation at approximately 60 s after birth. Again, this interplay between CC and ventricular preload is consistent with previous animal studies.10 At 5 min after birth, the TDI velocities of both ventricles returned to pre-CC values, underlining the transient nature of this phenomenon.

Increased MPI has been identified as a prognostic parameter characterising patients with conditions such as pulmonary hypertension, cardiac failure and myocardial infarction.26 27 While relatively independent of HR,28 29 previous animal studies reported that MPI is directly affected by acute changes in preload and afterload.30 31 We documented temporal variations of LV and RV MPI following CC. At T2, we observed MPI higher than reported normative values on the first day of postnatal life.32 Reassuringly, these changes returned to baseline values at T3.

In a multilinear regression model, RVO variation across CC was independently associated with the variation of RV afterload. LVO variation between T1 and T2 was independently associated with the variation of the septal mitral annulus A’ velocity. This suggests that establishment of LV preload is crucial in the postnatal transition of the left heart and highlights the importance of clamping the cord after lung aeration to increase the pulmonary blood flow and ensure sufficient pulmonary venous return to replace the umbilical venous return as the primary source of LV preload. Facilitating a smooth transition for the LV, and minimising LVO variation, may be of particular clinical importance for delivery room stabilisation of very preterm infants, who lack mature cerebral autoregulation,33 and have high risk for intraventricular haemorrhage.

Current recommendations in non-vigorous infants are to perform immediate CC.11 A recent randomised controlled trial suggests that this practice may not be beneficial.34 We suggest that immediate CC in non-vigorous infants may have more pronounced changes in cardiac performance than those described here with DCC in vigorous spontaneously breathing infants. Thus, performing ventilation while the infant remains attached to the umbilical cord may facilitate a smoother transition. Clinical trials demonstrate that initiating ventilation with an intact umbilical cord in such infants is feasible.35 36

We acknowledge limitations of the present study. Due to time restrictions, our echocardiography assessment was targeted and did not include initial interatrial or ductal shunting. Therefore, we could not directly document the interaction between systemic and pulmonary circulations or the interatrial shunting before and after CC. Considering the evidence supporting DCC, we felt it unethical to design a study comparing immediate and delayed CC. To maintain homogeneous CC timing, vaginally born infants were excluded from the present study as CC is typically performed at a later time period for these infants.37 While all newborns were spontaneously breathing at the time of echocardiogram, the level of respiratory effort and subjective assessment of breathing establishment was not tracked throughout and may have varied, although none required respiratory support subsequently. Further studies are warranted to investigate the cardiovascular impact of CC in preterm infants and/or non-vigorous infants and to assess the interaction between ventilation and CC in human neonates.

To conclude, we have described important aspects of haemodynamic transition in healthy-term human infants born by caesarean section. These transient physiological changes following CC may have a significant impact for the non-vigorous term infant. We believe our findings highlight the importance of establishing effective ventilation before CC. The current recommendation of immediate CC in non-vigorous infants warrants further consideration.

Data availability statement

Data are available on reasonable request. The data that support the findings of this study are available on request from the authors.

Ethics statementsPatient consent for publication

Consent obtained from parent(s)/guardian(s).

Ethics approval

This study was approved by Clinical Research Ethics Committee—University College Cork (reference number ID: CM 4 (z) 11/1/2022). Participants gave informed consent to participate in the study before taking part.