Association of Cerebrovascular Stability Index and Head Circumference Between Infants With and Without Congenital Heart Disease

Abstract

Congenital heart disease (CHD) is a common birth defect in the United States. CHD infants are more likely to have smaller head circumference and neurodevelopmental delays; however, the cause is unknown. Altered cerebrovascular hemodynamics may contribute to neurologic abnormalities, such as smaller head circumference, thus we created a novel Cerebrovascular Stability Index (CSI), as a surrogate for cerebral autoregulation. We hypothesized that CHD infants would have an association between CSI and head circumference. We performed a prospective, longitudinal study in CHD infants and healthy controls. We measured CSI and head circumference at 4 time points (newborn, 3, 6, 9 months). We calculated CSI by subtracting the average 2-min sitting from supine cerebral oxygenation (rcSO₂) over three consecutive tilts (0–90°), then averaged the change score for each age. Linear regressions quantified the relationship between CSI and head circumference. We performed 177 assessments in total (80 healthy controls, 97 CHD infants). The average head circumference was smaller in CHD infants (39.2 cm) compared to healthy controls (41.6 cm) (p < 0.001) and head circumference increased by 0.27 cm as CSI improved in the sample (p = 0.04) overall when combining all time points. Similarly, head circumference increased by 0.32 cm as CSI improved among CHD infants (p = 0.04). We found CSI significantly associated with head circumference in our sample overall and CHD infants alone, which suggests that impaired CSI may affect brain size in CHD infants. Future studies are needed to better understand the mechanism of interaction between CSI and brain growth.

Introduction

Congenital heart disease (CHD) puts neonates at increased risk for brain abnormalities and neurodevelopmental delays compared to healthy infants [1]. CHD affects approximately 1 in 100 live births in the United States, making it one of the most common birth defects [2, 3]. Roughly, 1.3 million American adults have survived from CHD due to the advancement in medical and surgical management [4–6]. Despite vast improvements in CHD survival, 75% of children with CHD will have cognitive or motor developmental delays during their early school years [7, 8], and 65% will require some form of remedial academic or behavioral assistance [9].

Microcephaly is an indicator for reduced brain growth and is a crucial marker for adverse neurodevelopmental outcomes [10]. Infants with CHD are more likely to have a smaller head circumference [11, 12] and neurodevelopmental delays compared to controls [13]; however, the cause of this is unknown. Impaired cerebral autoregulation may contribute to neurologic abnormalities in infants with CHD [14, 15]. We demonstrated previously that neonates with CHD have cerebrovascular instability compared to healthy controls, as they tended to have decreased cerebral oxygenation in response to a postural tilt, whereas healthy infants had increased brain saturations [16]. Cerebrovascular stability is likely connected to autoregulation as both relate to the brain’s homeostatic response to stabilize cerebral blood flow during changes in body positions [16]. Reduced cerebrovascular stability may cause cerebral hypoxia and decreased perfusion due to altered cerebrovascular hemodynamics, which have been associated with CHD [16]. Thus, we created a cerebral stability index (CSI) [16] and sought to examine its association to head circumference in infants with CHD compared to healthy control infants across four time points. We hypothesized that infants with CHD would have reduced CSI and smaller head circumference compared to healthy controls and that this would persist over time.

Materials and Methods

Study Design

We performed a prospective, longitudinal study in infants with CHD and healthy controls. We measured CSI and head circumference at neonatal, 3, 6, & 9-month ages. We recruited infants and performed assessments between March 2015 [17] and January 2021 at Children’s Hospital Los Angeles (CHLA) and its affiliated AltaMed general pediatrics clinic.

Study Sample

We recruited fetuses diagnosed with CHD from the Fetal Maternal Center and Fetal Cardiology clinics and infants with CHD admitted into the cardiothoracic intensive care and other inpatient units at CHLA. Infants with CHD were eligible to enroll in our study if they were ≥ 37 weeks gestational age at birth and had a documented structural heart defect that required admission to CHLA for additional monitoring or intervention. In addition, we excluded infants with CHD if they had hemodynamic instability or endotracheal intubation, due to inability to perform the study measures.

We recruited healthy controls infants from the newborn clinics at the AltaMed General Pediatrics clinic located within the CHLA campus. Healthy controls infants were eligible to enroll in our study if they were ≥ 37 weeks gestational age at birth, had no major complications during pre-, peri-, and postnatal periods, and had no congenital anomalies.

We excluded infants from both groups if they had any genetic defects, neurologic abnormalities, or intraventricular hemorrhage, were on antibiotics for a known infection, diagnosed with intrauterine growth restriction or small for gestational age, maternal substance abuse, maternal chorioamnionitis, or last trimester maternal or neonatal use of steroids. We assessed the infants ≤ 14 days old and within 2 weeks before or after turning 3, 6, and 9 months. We measured the infants serially at all age time points when possible. We measured head circumference in a standard fashion by placing the measuring tape at the occiput of the infant’s head, then above the ears, and lastly, above the eyebrows.

Cerebrovascular Stability Index

We measured CSI using a novel noninvasive bedside technique incorporating cerebral oxygenation ( rcSO₂) using the INVOS 5100C near-infrared spectroscopy (Somanetics, Troy, MI) during a postural tilt. The Bernoulli data acquisition system (Cardiopulmonary Corporation, Milford, CT) aggregated the data every 5 s. First, we placed the infant in a supine (0°) posture and measured rcSO₂ continuously for 2 min. Then, we put the infant in a sitting (90°) posture and measured r cSO₂ continuously for 2 min. We collected rcSO₂ data during three consecutive postural tilts (0–90°). We calculated CSI by subtracting the mean 2-min sitting from supine cerebral oxygenation over three consecutive tilts (0–90°). Then, we averaged the change in scores for each age (= [(mean sitting r cSO₂ for tilt 1 − mean supine r cSO₂ for tilt 1) + (mean sitting rcSO₂ for tilt 2 − mean supine r cSO₂ for tilt 2) + (mean sitting r cSO₂ for tilt 3 − mean supine r cSO₂ for tilt 3)]/3). The calculation for CSI differed from our statistical modeling for cerebrovascular stability in our groups’ prior work [16]. Our previous analysis of cerebrovascular stability used repeated measures analysis of variance model which accounted for the ~ 48 values of rcSO₂ (2 min of supine and sitting brain oxygenation), which was the outcome variable. This analysis, however, only had one value for our outcome (i.e., head circumference at each age time point), thus we could not use a repeated measures analysis of variance model and thus needed to create a single value for CSI predictor.

Statistical Analyses

We conducted all statistical analyses on IBM SPSS Statistics 27. We conducted and summarized the descriptive analyses based on CHD status as means with standard deviations (± SD) for continuous variables and frequencies with proportions for categorical variables. Two-sample independent T tests determined whether a statistically significant difference existed between the mean head circumference of the two independent groups of infants. Independent sample T tests also compared differences between population head circumference and the CHD and healthy control groups, respectively. Pearson’s chi-square test determined if the infant’s age differed by infant group.

Multiple linear regressions quantified the relationship between CSI and head circumference. We performed five separated linear regression models, (1) one for each of the four age time points to examine the association of CSI on head circumference and (2) one for the interaction term of CSI-by-age to examine CSI effect over time. The four separate models entered into the linear regressions included head circumference as the dependent variable with CSI as the independent variable for each age time point separately. The fifth regression model included head circumference as the dependent variable with the main effect of CSI and age as the independent variables, and the interaction term of CSI-by-age at the time of exam. Lastly, we performed a sensitivity analysis using a regression model that included head circumference as the dependent variable at each age time point with CSI as the independent variable, while covarying for group, sex, and gestational age at the time of exam. All p values ≤ 0.05 were considered significant.

Results

We performed a total of 177 assessments on healthy controls (80 assessments) and infants with CHD (97 assessments). Specifically, we examined 19 healthy controls and 31 infants with CHD at the neonatal time point, 24 healthy controls and 24 infants with CHD at the 3-month time point, 21 healthy controls infants and 23 infants with CHD at the 6-month time point, and 16 healthy controls infants and 19 infants with CHD at the 9-month time point. See Table 1 for the values obtained in both CHD and healthy control groups. See Table 2 for a breakdown of the cardiac subtypes and status of CHD repair at each time point. The average head circumference was significantly smaller in infants with CHD (39.2 cm, SD ± 4.5 cm) than in healthy controls infants (41.6 cm, SD ± 3.5 cm) (t = 3.8, 95CI% 1.1, 3.6, p < 0.001) overall (Table 1). Moreover, infants with CHD had significantly smaller (~ 1 cm) head circumference, when compared to population norms, at all age time points (Table 1). On average, infants with CHD had poorer CSI (− 0.75) compared to healthy controls infants (− 0.57), however did not reach statistical significance (Table 1).

Table 1.

Participant characteristics stratified by group: infants with congenital heart disease (CHD) and healthy controls

	Infants with CHD Mean (± SD)	Healthy controls Mean (± SD)	p value
Number of infant’s at each age	97	80	0.66
Neonate	31	19
3-months	24	24
6-months	23	21
9-months	19	16
Head circumference (cm), all ages	39.18 (± 4.49)	41.57 (± 3.53)	< 0.001**
Neonate	33.73 (± 1.42)	35.04 (± 1.95)	0.04*
3-months	39.17 (± 1.95)	40.22 (± 1.78)	0.06
6-months	42.52 (± 2.30)	43.18 (± 1.52)	0.29
9-months	44.05 (± 1.44)	45.25 (± 1.25)	0.03*
Comparison with population norms for head circumference (cm) at each age
Neonate (35.81 cm)	33.73 (± 1.42)	35.04 (± 1.95)	< 0.0001**/0.23
3-months (41.21 cm)	36.17 (± 1.95)	40.22 (± 1.78)	< 0.0001**/ < 0.0001**
6-months (43.72 cm)	42.52 (± 2.30)	43.18 (± 1.52)	0.005*/0.05
9-months(45.27 cm)	44.05 (± 1.44)	45.25 (± 1.64)	< 0.0001**/0.95
Cerebrovascular Stability Index ( rcSO2), all ages	− 0.75 (± 2.92)	− 0.57 (± 1.99)	0.65
Neonate	− 1.87 (± 2.23)	− 1.17 (± 2.75)	0.35
3-months	− 0.39 (± 4.44)	− 0.58 (± 1.25)	0.85
6-months	− 0.32 (± 1.74)	− 0.36 (± 1.17)	0.94
9-months	− 0.02 (± 2.15)	− 0.27 (± 2.69)	0.90

SD standard deviation

^*p < 0.05;

^**p < 0.01

Table 2.

Type of congenital heart defects by assessment time points

Congenital heart defects subtypes	N (%)
Neonatal
N = 31
Cardiac defect
Arch anomaly + VSD	5 (16.1)
d-TGA	5 (16.1)
HLHS	4 (12.9)
Isolated arch anomaly	2 (6.5)
l-TGA/Coarctation	1 (3.2)
Single ventricle with pulmonary stenosis/atresia	6 (19.4)
TAPVR	2 (6.5)
TOF	3 (9.7)
Truncus arteriosus	3 (9.7)
Cyanotic
No	8 (25.8)
Yes	23 (74.2)
Cardiac defect repaired
No	31 (100)
Yes	0 (0)
3 Months
N = 24
Cardiac defect
Arch anomaly + VSD	2 (8.3)
d-TGA	4 (16.7)
HLHS	3 (12.5)
Isolated arch anomaly	2 (8.3)
l-TGA/Coarctation	1 (4.2)
Single ventricle with pulmonary stenosis/atresia	7 (29.2)
TAPVR	2 (8.3)
TOF	1 (4.2)
Truncus arteriosus	2 (8.3)
Cardiac defect repaired
No (remain cyanotic)	11 (45.8)
Yes	13 (54.2)
6 Months
N = 23
Cardiac defect
Arch anomaly + VSD	3 (13.0)
d-TGA	5 (21.7)
HLHS	3 (13.0)
Isolated arch anomaly	2 (8.7)
l-TGA/Coarctation	1 (4.3)
Single ventricle with pulmonary stenosis/atresia	4 (17.4)
TAPVR	3 (13.0)
TOF	2 (8.7)
Cardiac defect repaired
No (remain cyanotic)	9 (39.1)
Yes	14 (60.8)
9 Months
N = 19
Cardiac defect
Arch anomaly + VSD	2 (10.5)
d-TGA	5 (26.3)
HLHS	3 (15.8)
Isolated arch anomaly	1 (5.3)
l-TGA/Coarctation	1 (5.3)
Single ventricle with pulmonary stenosis/atresia	3 (15.8)
TAPVR	3 (15.8)
Truncus arteriosus	1 (5.3)
Cardiac defect repaired
No (remain cyanotic)	6 (31.6)
Yes	13 (68.4)

Cyanotic defects included single ventricles, unrepaired d-TGA, TOF, truncus arteriosus, and TAPVR

d-TGA dextro-transposition of the great arteries; l-TGA levo-transposition of the great arteries; HLHS hypoplastic left heart syndrome; TAPVR total anomalous pulmonary venous return; VSD ventricular septal defect; TOF tetralogy of fallot

As CSI improved (averaged change score in rcSO₂ becomes more positive by 1-unit), head circumference increased by 0.27 cm in the sample including all age time points (β = 0.27, 95%CI 0.01, 0.52, p = 0.044) (Table 3). Additionally, age did not significantly moderate the association of CSI on head circumference (β = − 0.024, 95%CI − 0.15, 0.11, p = 0.71) (Table 3). In a subgroup analysis including all age time points, infants with CHD had comparable results (Fig. 1); as CSI improved among infants with CHD, head circumference increased by 0.32 cm (β = 0.32, 95%CI 0.01, 0.63, p = 0.042) (Table 3). However, CSI did not significantly associate with head circumference in healthy controls (β = − 0.031, 95%CI − 0.49, 0.43, p = 0.89) (Table 3).

Table 3.

Univariate linear regression analyses on the association of head circumference (dependent variable) and Cerebrovascular Stability Index (CSI) by group at all age time points

	𝛽	95% Confidence interval	p value
Overall sample	0.27	0.01, 0.52	0.04*
Healthy controls	− 0.03	− 0.49, 0.43	0.89
Infants with CHD	0.32	0.01, 0.63	0.04*
Interaction of CSI × age	− 0.02	− 0.15, 0.11	0.71

^*p < 0.05

Fig. 1.

Linear relationships between Cerebrovascular Stability Index (CSI) and head circumference. The green line demonstrates that as CSI improves the head circumference increases in the infants with congenital heart disease compared to the blue line which is flat for the healthy controls

Discussion

We found that our novel measure of cerebrovascular stability, calculated by cerebral oxygenation changes in response to a postural tilt, correlated with smaller head circumference in CHD infants when compared to a control group over time. Overall, infants with CHD had smaller head circumference and poorer CSI compared to healthy controls, though the CSI differences did not reach statistical significance. Similarly, as CSI improved, head circumference also increased when combining all age time points. In a subgroup analysis, infants with CHD showed similar findings to the overall sample.

As our technique is novel, we do not have other reports to compare findings for CSI and explain mechanisms. We based our definition of CSI on other pediatric and adult responses of r cSO₂ to gravitational challenges [18–20]. Individuals with medical preconditions exhibited poorer cerebrovascular hemodynamics through decreased r cSO₂ after the postural changes or tilts. The mean CSI for the control infants at 3 and 9 months was lower than the group with CHD. We do not have a clear explanation for the CSI findings in the healthy control group. We theorize, however, that the infants with CHD have more variability than the healthy control group, as demonstrated by the larger standard deviations in the CHD group compared to controls (Table 1). The variability may suggest poorer cerebral hemodynamics in the CHD group. Although the CSI in the CHD group seemed to improve from the neonatal to the 3-month time point, the improvement in CSI may be explained by the first palliative surgery or corrective surgery at the 3-month time compared to the unrepaired neonatal time. Cerebrovascular instability may be related to the underlying type of CHD (cyanotic vs acyanotic) or delayed maturation of the brain in neonates with CHD. We do not know the usual change or progression of CSI over time as this is a novel measure, however we are in the process of examining this concept. Although it may appear that head circumference and CSI are collinear, we believe these variables are independent as prior reports demonstrate variations in r cSO₂ during postural changes in vulnerable infants such as those born prematurely or those with periventricular leukomalacia [21–24]. Moreover, children with CHD continue to display neurocognitive and developmental delays throughout early childhood, as such, dysregulated brain blood flow may be contributing to this problem.

Cerebral hypoxia may be a potential mechanism to cause cerebrovascular instability, which in turn causes cerebral blood oxygen levels to fluctuate [16] and may explain the CSI effects on head circumference growth in infants with CHD. The brain may be at an increased risk for structural damage and functional impairment due to the reduced cerebrovascular stability and inability to properly maintain adequate blood oxygen levels, nutrient delivery, and blood flow to the brain [16]. However, we observed perturbations in CSI even in infants without cyanotic CHD and in older infants with repaired CHD. Impaired head growth has been associated with poorer neurodevelopmental outcomes in children with hypoplastic left heart syndrome (HLHS) [25]. Longitudinal neurodevelopmental evaluations are ongoing in our cohort. Additionally, head size alone does not describe the entire story, as alterations in brain composition [26] as well as altered brain metabolism [27] and white matter microstructure [28] have been demonstrated in CHD subjects; future studies will aim to assess brain MRI analysis in association with CSI.

Multiple studies found that infants with CHD had smaller mean head circumference than expected for their gestational age [10, 12, 29] in line with our study findings [10, 12]. Those reports, however, used different techniques to measure cerebrovascular hemodynamics, such as cerebroplacental ratios, or cerebral blood flow measured by magnetic resonance imaging. In contrast, our study utilized a novel, noninvasive technique to examine cerebral hemodynamics, by measuring cerebral oxygenation changes in response to a tilt, and thus CSI. We identified an association between head circumference and altered cerebrovascular hemodynamics. Moreover, we used a prospective control group for comparison, strengthening our findings. Only one of the reports above used prospective control groups.

Children with CHD likely have prenatal origins of neurological abnormalities and issues with brain growth that may persist during childhood development. Previous studies found abnormal head biometry changes in the second trimester among mothers carrying fetuses with CHD, with head circumference measurements already below the 5th percentile [30]. CHD lesions are often associated with alteration of fetal cerebral substrate delivery, including fetal intracardiac mixing and change from the usual fetal circulation aimed at sending the blood with highest oxygen content to the brain, such as d-transposition of the great arteries and single ventricle defects, which may impact fetal cerebral saturation and potentially the brain growth [31]. Studies found that abnormal head biometry changes in fetuses with cyanotic CHD as early as the first trimester [22]. Then in the second trimester, head circumference measurements fall below the 5th percentile [30], in addition to a reduction of total brain volumes [32] in fetuses with CHD. Since fetuses with CHD have smaller head biometry and altered cerebral oxygen delivery, it is probable that the newborns with CHD also have distorted cerebrovascular hemodynamics and smaller head circumference.

Clinical implications for infants at risk for reduced CSI are to perform more careful or gentle body position changes and avoiding fluctuations in systemic blood pressure. Longitudinal neurodevelopmental evaluations are ongoing in our cohort. Additionally, head size alone does not describe the entire story, as alterations in brain composition [26] as well as altered brain metabolism [27] and white matter microstructure [28] have been demonstrated in CHD subjects; future studies will aim to assess brain MRI analysis in association with CSI. It is unknown if reduced CSI as a neonate or at any stage < 1 year of age results in worse future neurodevelopmental outcomes; however, we are currently collecting longitudinal data on the correlation between CSI with neurodevelopmental outcomes. Since our CSI is novel, we are still investigating whether there is a minimum threshold for CSI which significant impairment in neurodevelopmental outcomes can be predicted.

Limitations

This study contains a few limitations. First, we could not examine if CHD type affected the association between head circumference and CSI as we could not stratify by CHD type due to our small sample size. Second, while we standardized our method for head circumference measurement, it was obtained by two separate study members. Third, our results’ generalizability may be limited as most of the CHD participants recruited were of Latin decent and from the Los Angeles Metropolitan area.

Conclusion

Findings from this study revealed a relationship between cerebrovascular abnormalities (indicated by a reduced CSI compared to healthy controls) and head circumference in infants with CHD. Head circumference increased as CSI improved in our CHD group. Further investigations examining the association of CSI with brain structure, developmental delays, and brain imaging will help to provide insight into this relationship.