Association of early skin breaks and neonatal thalamic maturation

Abstract

Objective

To test the hypothesis that a strategy of prolonged arterial line (AL) and central venous line (CVL) use is associated with reduced neonatal invasive procedures and improved growth of the thalamus in extremely preterm neonates (<28 weeks' gestation).

Methods

Two international cohorts of very preterm neonates (n = 143) with prolonged (≥14 days) or restricted (<14 days) use of AL/CVL were scanned serially with MRI. General linear models were used to determine the association between skin breaks and thalamic volumes, accounting for clinical confounders and site differences. Children were assessed at preschool age on standardized tests of motor and cognitive function. Outcome scores were assessed in relation to neonatal thalamic growth.

Results

Prolonged AL/CVL use in neonates (n = 86) was associated with fewer skin breaks (median 34) during the hospital stay compared to restricted AL/CVL use (n = 57, median 91, 95% confidence interval [CI] 60.35–84.89). Neonates with prolonged AL/CVL use with fewer skin breaks had significantly larger thalamic volumes early in life compared to neonates with restricted line use (B = 121.8, p = 0.001, 95% CI 48.48–195.11). Neonatal thalamic growth predicted preschool-age cognitive (B = 0.001, 95% CI 0.0003–0.001, p = 0.002) and motor scores (B = 0.01, 95% CI 0.001–0.10, p = 0.02). Prolonged AL/CVL use was not associated with greater incidence of sepsis or multiple infections.

Conclusions

Prolonged AL/CVL use in preterm neonates may provide an unprecedented opportunity to reduce invasive procedures in preterm neonates. Pain reduction in very preterm neonates is associated with optimal thalamic growth and neurodevelopment.

Very preterm neonates (<32 weeks' gestation) are often exposed to ≥14 invasive procedures/day¹ as part of lifesaving care in the neonatal intensive care unit (NICU). Increasing invasive procedures predict delayed thalamic maturation in very preterm neonates.^2–4 Neonates born at the youngest gestational ages (<28 weeks' gestation) appear to be the most vulnerable to the adverse effects of invasive procedures.^2,3 In very and extremely preterm neonates, early exposure to noxious stimuli has been associated with altered behavioral responses,⁵ adverse neurodevelopmental outcomes,⁶ disrupted functional connectivity,⁷ and alterations in thalamic and thalamocortical development later in childhood.⁸ These findings suggest that the developmental consequence of early exposure to noxious stimuli may be influenced by alterations in early thalamic development.

The management of neonatal pain and discomfort is challenging in the context of providing neonatal intensive care, where venous and arterial access are essential. Arterial lines (ALs) are used for blood sampling and blood pressure monitoring. Central venous lines (CVLs) are used to deliver IV fluids, nutrition, and medication.⁹ While these lines reduce the number of invasive procedures, their use may be associated with an increased incidence of nosocomial infections.¹⁰ As such, there is variability across NICUs in the use of these lines.

In 2 cohorts of extremely preterm neonates,^3,11 one with restricted AL and CVL use (<14 days) and the other with prolonged use of AL and CVL (≥14 days), we examined the association of early skin breaks and neonatal thalamic volumes as determined by MRI in neonates. We then assessed neonatal thalamic growth from the neonatal period until term equivalent age in relation to motor and cognitive functions assessed at preschool age.

Methods

Study population

Two prospective cohorts of extremely preterm neonates (<28 weeks' gestation) were included in the study. At site 1 (British Columbia Women's Hospital, Vancouver, Canada), restricted AL and CVL were used as part of routine clinical care (i.e., central lines for 1–2 days, up to a maximum of 14 days). The neonates were recruited (2006–2013) as part of a prospective research study.^2,3,12,13 Neonates were scanned twice, once early in life (MRI 1) and again at term equivalent age (MRI 2).

At site 2 (Wilhelmina Children's Hospital of the University Medical Center Utrecht, the Netherlands), neonates had prolonged AL/CVL use during the stay in the NICU (i.e., central lines for 6–7 days, replaced by peripheral ALs and peripherally inserted central catheters [PICCs]). The MRI scans were acquired as part of a prospective research study on extremely preterm neonates between 2008 and 2010 (Neobrain study) and since then are acquired as part of routine clinical care. Neonates were also scanned early in life (MRI 1) and again at term equivalent age (MRI 2).

From both sites, we retrospectively included neonates born before 28 weeks' gestation, with an early brain MRI scan (MRI 1) prior to a postmenstrual age (PMA) of 33.3 weeks (within 5 weeks of birth) and a term equivalent age scan (MRI 2).

Standard protocol approvals, registrations, and patient consents

At site 1, the study was approved by the Research Ethics Board. Families provided informed consent. At site 2, most infants were enrolled in the Neobrain study, approved by the Research Ethics Board, and informed consent was obtained from parents. In the remaining infants, the Research Ethics Board gave permission for use of the clinical data for research.

Magnetic resonance imaging

Infants at site 1 were scanned unsedated (MRI 1: 30.4 weeks PMA [interquartile range (IQR) 29.7–31.9]; MRI 2: 39.7 weeks PMA [IQR 38.4–41.6]). Most infants at site 2 were scanned unsedated for MRI 1 (30.7 weeks PMA [IQR 30–31.3]; all were sedated for MRI 2: 41.1 weeks PMA [IQR 40.9–41.4]) using chloral hydrate. Volumes of the thalamus and total cerebral volumes were extracted using established pipelines.^3,14,15 In brief, the volumetric segmentations were performed using the Multiple Automatically Generated Templates (MAGeT) Brain^16,17 pipeline. The algorithm requires a small number of manually labeled anatomical structures that are propagated to an intermediate template library to segment all images in a dataset using a nonlinear registration algorithm.^18,19 The volumes of white matter injury (WMI) were calculated based on manual delineations of punctate lesions identified on MRI.¹⁴

Demographic and clinical data collection

At both sites, extensive clinical data were collected from the patient charts, prospectively at site 1 and retrospectively at site 2. We quantified neonatal procedural pain/stress as the number of invasive procedures (e.g., heel lance, peripheral IV or central line insertion, chest tube insertion, nasogastric tube insertion, catheter insertion for urine, pleural taps) as previously reported^2,6,20 during the stay in the NICU, using identical data collection sheets at both sites.

The cumulative dose of midazolam and morphine was calculated (IV dose plus converted oral dose) as the average daily dose adjusted for daily body weight.^2,21 Additional variables of interest included gestational age (GA) at birth, sex, PMA, length of stay, days of intubation, postnatal positive culture infections, necrotizing enterocolitis (NEC), patent ductus arteriosus, Apgar scores at 5 minutes, volumes of WMI,^14,22 and steroids (hydrocortisone and dexamethasone). At site 2, only hydrocortisone was administered. To develop a composite measure of postnatal steroid exposure, values of hydrocortisone (site 2) and dexamethasone (site 1) were log transformed to make the dosages in milligrams equivalent.

Central lines

At site 1, medications were administered through umbilical venous catheters (UVCs) and PICCs. Blood was not drawn from these catheters. At site 2, UVCs/PICCs were used for the delivery of total parenteral nutrition, morphine, and inotropes. At both sites, umbilical artery catheters (UACs) and peripheral ALs were used for blood sampling and blood pressure monitoring. Heparin was only used for the ALs (dose: site 1, 1 unit heparin per milliliter; site 2, 0.5 unit heparin per milliliter).

Line-related sepsis was considered when a positive blood culture was obtained when the UVC or PICC was either still present or within 48 hours after removal.

Neuropsychological testing

Children returned at median ages of 4 years 9 months (site 1) and 5 years 9 months (site 2) of age for motor and cognitive testing. The Wechsler Preschool and Primary Scale of Intelligence, third edition (WPPSI-III), a measure of general intelligence, was administered to all children. The Movement Assessment Battery for Children, second edition (MABC-2)²³ is designed to identify and describe motor impairments.

Statistical analysis

Analyses were carried out using SPSS (v 25; SPSS, Chicago, IL). To account for variations in clinical care, MRI scanner strength, and education, all analyses were adjusted for site. Our first aim was to assess whether prolonged AL/CVL use was associated with fewer invasive procedures and thalamic volumes. Based on our previous work indicating that early invasive procedures in extremely preterm neonates were associated with smaller thalamic volumes assessed in the first few weeks of life,³ we focused our analyses for our first aim on the early acquired MRIs (MRI 1). To examine AL/CVL use in relation to invasive procedures, we used an unpaired t test. We subsequently examined the association between skin breaks and thalamic volumes in a general linear model, adjusting for demographic and clinical confounders such as days of intubation and WMI volume, among others, described in detail below and controlling for site (1 = restricted AL/CVL use; 2 = prolonged AL/CVL use). We had one hypothesis regarding the association of AL/CVL use and thalamic volumes and the α level was set at 0.05.

Raw scores on the MABC-2 were log transformed. The MABC-2 and WPPSI-III (full-scale IQ [FSIQ], performance IQ, verbal IQ, processing speed) were assessed in relation to neonatal thalamic growth, adjusting for site, birth GA, sex, maternal education, and volumes of early WMI.¹⁴ Thalamic growth was calculated by dividing the difference in volumes acquired at the term-equivalent scan compared to the early scan time point by the difference in PMA at the time of the scans (volume 2 – volume 1/PMA scan 2 – PMA 1). As our hypothesis addressed 2 outcomes regarding the association of motor and cognitive function relative to early invasive procedures, a p value < 0.025 was considered statistically significant.

Data availability

Study data are not available for download in accordance with local data sharing policies stipulated by the research ethics boards. Families were not consented to share the data publicly.

Results

Clinical and demographic variables

The participant flow is illustrated in figure 1. At site 1, a total of 57 neonates (median GA 26.6 weeks, IQR 25.9–27.6 weeks, range 24.3–28.9 weeks, median birthweight 940 g) who had restricted AL and CVL use (i.e., central lines for 1–2 days, up to a maximum of 14 days) were included in the study.

Figure 1. Participant flow chart.

At site 2, 86 neonates (median GA 26.9 weeks, IQR 26–27.4 weeks, range 24.4–27.9 weeks, median birthweight 915 g) had prolonged use of AL and CVL (i.e., central lines for 6–7 days and were replaced by peripheral arterial and PICC lines) during their NICU stay and were included in the study. At site 1, morphine and midazolam were administered to the neonates to manage pain and discomfort. Morphine was administered during acute painful procedures, while midazolam was primarily given to neonates to manage agitation during chronic ventilation. At site 2, morphine was administered during acute painful procedures and was given to neonates to manage agitation during chronic ventilation.

At site 1, 79% (n = 45) of neonates had an umbilical vein or umbilical artery line inserted on the first or second day of life for a minimum of 2–3 days for a maximum of 14 days, and on average for 6–7 days. At site 2, all neonates had central lines inserted that were removed after a maximum of 6–7 days and replaced by peripheral ALs until repetitive daily blood samples were no longer needed and PICC lines were replaced only when less than 50% of enteral feeding was achieved.

At site 1, restricted use of AL/CVL was associated with a greater number of invasive procedures (median 91, IQR 65–153) in comparison to neonates at site 2 who had prolonged use of AL/CVL during the NICU stay (median 34, IQR 26–42, p < 0.001, 95% confidence interval [CI] 60.35–84.89; table 1). At site 1, early-onset sepsis (<72 hours of birth) was present in 1 infant; late-onset culture-positive blood infections (>72 hours after birth) were present in 1 infant after a PICC line and a peripheral infusion in 9 infants. At site 2, early-onset sepsis was present in 2 babies; late-onset sepsis was present in 8 infants who had a UVC line inserted, 9 infants with PICC lines, and 6 infants who had a peripheral infusion. Further details on the associations of line use and health outcomes are provided in table 1.

Table 1.

Participant demographics and clinical variables for infants with restricted use of an arterial/central line (site 1, n = 57) or with prolonged arterial/central line (site 2, n = 86)

Neonates at site 1 who had restricted use of AL/CVL had comparable incidence of culture-positive infection relative to neonates at site 2 with prolonged AL/CVL use (p = 0.2; table 1). At site 1, the neonate with an early-onset infection had Coagulase-negative staphylococci (CoNS) bacteria.

For neonates at site 2, the 2 neonates with early infections had Escherichia coli and group B streptococcus.

At site 1, in neonates with late-onset infection, the organisms were primarily CoNS (n = 9 [90%]) and one infant had E coli bacteria detected (10%). At site 2, 18 neonates had CoNS bacteria detected (78%), 3 (13%) had E coli, and the remaining 2 neonates had Staphylococcus warneri and Staphylococcus aureus bacteria detected. The types of organisms for late-onset sepsis did not differ significantly between sites (χ² = 1.4, p = 0.8).

Cross-sectional models: Invasive procedures predict smaller thalamic volumes

In a basic model, greater numbers of invasive procedures were significantly associated with decreased thalamic volumes at the first scan acquired early in life (B = −0.88, CI −1.51 to −0.25, p = 0.006; table 2), adjusted for site, birth GA, PMA at scan, male sex, days of intubation, total cerebral volumes, and steroids. Figure 2 demonstrates smaller thalamic volumes relative to a higher number of invasive procedures, including the data from both cohorts.

Table 2.

Results of a general linear model of early neonatal thalamic volumes

Figure 2. Early thalamic volumes divided by total cerebral volumes plotted against the number of invasive procedures at site 1 and 2.

In an analgesic and sedative model adjusting for WMI volumes, exposure to morphine, midazolam, steroids, Apgar scores, culture-positive infection, and invasive procedures remained significant and a strong predictor of decreased thalamic volumes (B = −1.7, CI −3.3 to −0.2, p = 0.03; table 2). In an additional model examining patent ductus arteriosus, no association was found with thalamic volumes (χ² = 1.87, p = 0.6); however, invasive procedures remained a significant negative predictor of thalamic growth (B = −1.04, CI −1.69 to −0.39, p = 0.002). Similarly, in a model examining the association of NEC with thalamic volumes (B = 48.37, CI −58.03 to 154.77, p = 0.4), invasive procedures remained significant (B = −2.38, CI −4.27 to −0.50, p = 0.01).

In an extended model, adjusting for analgesics, sedatives, and steroids, we examined thalamic volumes based on the median number of invasive procedures for the sites (site 1: ≥91 [median]; site 2: ≤34 [median]) as well as the range of invasive procedures between the sites (35–90). When examining the data in the 3 invasive procedures groups (group 1: ≥91; group 2: range 35–90; group 3: ≤34), thalamic volumes differed significantly (χ² = 11.2, p = 0.004). Group 3 (16–34 skin breaks) had larger thalamic volumes (B = 121.8, CI 48.48–195.11, p = 0.001) relative to group 1. Group 2 (34–90 [60% site 1], B = 87.58, CI 31.02–144.14, p = 0.002) also had larger volumes relative to group 1, adjusting for site, birth GA, sex, days of intubation, PMA at scan, morphine, midazolam, steroids, and total cerebral volumes (figure 3).

Figure 3. Preterm neonates exposed to fewer numbers of early invasive procedures have larger thalamic volumes (B = 121.8, p = 0.001).

Estimated marginal means adjusted for covariates: birth gestational age, age at scan, total cerebral volume, site, male sex, days of intubation, and doses of midazolam, morphine, and steroids. **p < 0.01.

Longitudinal models: relationship between neurodevelopmental outcome and thalamic maturation

Children (n = 118, 83%) returned for assessments at a median age of 4 years and 9 months (site 1, n = 38 [67%]) and 5 years and 9 months (site 2, n = 80 [93%]) years of age. At site 1, 2 (3.5%) infants died, and 17 children (28%) were lost to follow-up. At site 2, 6 children (7%) were lost to follow-up. Children's motor outcomes spanned a wide range of percentiles from the 1st to percentiles in the 90th range in both cohorts. The median percentiles for the MABC-2 were 31 (IQR 11–72) for site 1 and 25 (IQR 9–63) for site 2. Children spanned a wide range of cognitive outcomes; however, the median scores for the WPPSI-III FSIQ were within the typical range: 104 (IQR 95–112, range 78–131) for site 1 and 96 (IQR 87–109, range 57–124) for site 2.

In a general linear model, motor raw scores (log transformed) on the MABC-2 were positively associated with thalamic volumetric growth (B = 0.01, CI 0.001–0.10, p = 0.02), adjusted for birth GA, site, sex, and WMI volumes and maternal education.

Processing speed on the WPPSI-III was positively predicted by thalamic growth (B = 0.001, CI 0.0003–0.001, p = 0.002) adjusting for clinical care factors (birth GA, site 2, male sex, maternal education, WMI volume). No associations of FSIQ (B = 0.0004, CI −0.0009–0.001, p = 0.1), performance IQ (B = 0.001, CI 0.00008–0.001, p = 0.025), or verbal IQ (B = 0.00008, CI −0.0004–0.001, p = 0.8) with thalamic growth were evident in these multivariable models.

Discussion

This multicenter study with extremely preterm neonates (<28 weeks' gestation) found that neonates with extended use of AL/CVL had fewer invasive procedures and larger thalamic volumes, which were predictive of preschool age cognitive (processing speed) and motor outcomes.

Vascular access is necessary for the care of extremely and very preterm neonates. A UVC provides easy and quick vascular access in very low birthweight neonates and a UAC allows continuous blood pressure measurements and repeated blood sampling with more accurate CO₂ and O₂ measurements than capillary samples or transcutaneous monitoring. Many centers, including site 1 in the current study, will insert umbilical lines at birth but will remove these after a few days because of the risk of infection or thrombosis. In site 2, umbilical catheters were removed on day 6 or 7 and the UVC replaced by a PICC line when oral feeding was still limited and a UAC by a radial or tibial artery line when the infant still needed ventilation or frequent blood sampling.²⁴ Some studies have reported that UVC, especially when not removed within 7 days following insertion, is associated with an increased risk of sepsis, but this was not confirmed by others or in a recent Cochrane review.^25–28 A recent prospective randomized controlled trial examining UVC and PICC use in neonates reported no difference in the incidence of infection.²⁹ In the current study, neonates with restricted and prolonged use of AL/CVL had comparable incidences of sepsis. In addition, the incidence of sepsis in the neonates with restricted AL/CVL showed no association with thalamic volumes.

Besides the risk of infection, a central line carries a risk of thrombosis. In a review of 26 articles, the incidence was 9.2%, with increased length of stay, malpositioning of the central catheter, and infusion of blood products across the line identified as risk factors.³⁰ In a recent study using prospective ultrasound screening, 40% of infants with a UVC were diagnosed with a mostly asymptomatic thrombosis by day 6–8. Regression of the thrombus was seen without anticoagulation therapy.³¹ Functional loss of one kidney occurred in one of the infants at site 2, a rare but recognized complication of an umbilical AL in situ.³² The thrombosis was diagnosed on day 5, which is within the window we allow the UAC to be in place. The risks associated with AL/CVL use in neonates highlights the feasibility of conducting randomized controlled trials in this population and the need for multicenter studies to examine differences in AL/CVL use on brain development and functional outcomes.

A key finding from our study is that we stratified our analysis of thalamic volumes by invasive procedures and found that regardless of site the total number of invasive procedures was associated with smaller early thalamic volumes. The diencephalon appears during the 5th week of gestation, shortly after the closure of the neural tube in the human embryo.³³ The thalamus develops exponentially through embryonic and fetal life with thalamocortical connections forming in weeks 12–16 of gestation.³⁴ During the second and third trimesters, the thalamus nearly doubles in volume in typical fetal development.³⁵ Our results demonstrating smaller thalamic volumes with increasing number of invasive procedures are in line with previous findings in neonates.^2,36 These findings are particularly relevant given the association of thalamic volumes with neurodevelopmental outcome in early childhood³ and the recognition of thalamo-cortical connectivity as a critical contributor to neurodevelopmental function in preterm born children and adults.³⁷ Moreover, our results highlight that a reduction in the total number of invasive procedures through continuous AL/CVL use is associated with thalamic growth and optimal functional outcomes.

Whether the benefit of reducing the number of invasive procedures by extending the use of AL/CVL indeed outweighs the complications that may occur needs to be further studied in future multicenter studies. Of note, invasive procedures were collected retrospectively at site 2, which may inadvertently introduce bias into the dataset. The results of the current study, whereby we combined invasive procedure data from neonates at 2 NICUs where the use of AL/CVL varied from a few days (site 1) to several weeks (site 2), indicate that extended AL/CVL use is associated with fewer invasive procedures and the overall benefits for thalamic development and functional outcomes may outweigh the risks associated with a central line.

Pain management practices in extremely preterm neonates should be refocused on minimizing exposure to noxious stimuli. Optimal use of AL/CVL in preterm neonates, balancing pain reduction and infection prevention, needs to be determined in future multicenter studies.

Glossary

AL

arterial line

CI

confidence interval

CoNS

Coagulase-negative staphylococci

CVL

central venous line

FSIQ

full-scale IQ

GA

gestational age

IQR

interquartile range

MABC-2

Movement Assessment Battery for Children, second edition

NEC

necrotizing enterocolitis

NICU

neonatal intensive care unit

PICC

peripherally inserted central catheter

PMA

postmenstrual age

UAC

umbilical artery catheter

UVC

umbilical venous catheter

WMI

white matter injury

WPPSI-III

Wechsler Preschool and Primary Scale of Intelligence, third edition

Appendix. Authors

Study funding

This research has been funded by the Canadian Institutes of Health Research MOP 79262 to S.P.M. and MOP 86489 to R.E.G. and the Kid's Brain Health Network to S.P.M. S.P.M. is supported by the Bloorview Children's Hospital Chair in Paediatric Neuroscience. This work includes infants participating in the Neobrain study (LSHM-CT-2006-036534, contract 036534).

Disclosure

The authors report no disclosures relevant to the manuscript. Go to Neurology.org/N for full disclosures.