Introduction

Critically ill neonates often require arterial cannulation for hemodynamic monitoring and blood sampling. Arterial access in neonates can be achieved using central (e.g., umbilical, femoral) or peripheral (e.g., radial, posterior tibial) arteries.1,2,3 Peripheral arterial access is preferred when cannulation of a central artery is not feasible (e.g., anterior abdominal wall defects) or difficult.4 Radial artery is most frequently chosen for peripheral arterial access given its superficial location and a relatively larger diameter compared to other arteries.5,6 Cannulating peripheral arteries in neonates can often be difficult due to their small calibre. Repeated attempts at cannulation may lead to bleeding, spasm and dissection of the peripheral arteries.7,8

Ultrasound-guidance can increase the success of arterial cannulation but requires high level of expertise.9,10,11 In ideal circumstances, first pass success rate with ultrasound guided dynamic tip positioning is 83-86%.12 However, a systematic review of studies (n = 8) assessing pediatric ultrasound guided arterial cannulation has reported a success rate ranging from 14 to 85%.13 This wide variation probably relates to heterogeneity in operator skills and patient characteristics.9,10 Overall, these results suggest that despite the use of ultrasound, there is scope for improving procedural success.

Increasing the internal diameter of the artery and preventing vasospasm are potential strategies for improving successful first attempt peripheral arterial cannulation in neonates. Glyceryl trinitrate (GTN) may be beneficial in this context considering it increases the diameter of peripheral arteries and prevents vasospasm when applied locally.14 GTN is converted to nitric oxide in the vascular smooth muscle, which in turn activates guanylate cyclase and increases the cyclic guanosine monophosphate levels. This relaxes the arterial smooth muscle leading to vasodilatation.15 Studies in adult patients have shown that topical GTN improves the success rate of cannulation by increasing the internal diameter of peripheral arteries and reducing vasospasm and occlusions during cardiac catheterization.14,16,17

Our systematic review exploring this issue in the pediatric population found only two small (N = 153) studies that did not recruit neonates.18 One used topical GTN patch8 and the other used subcutaneous GTN.19 Both reported that GTN improved the success rate and reduced the time required for arterial cannulation. A recently published randomised controlled trial (RCT) found that subcutaneous GTN in pediatric patients prevents radial artery occlusion after catheter removal by increasing radial artery diameter (RAD) and improving blood flow.20

Considering its potential benefits, we aimed to assess the efficacy and safety of topical GTN in dilating the radial artery for peripheral arterial cannulation in neonates. Our hypothesis was that topical GTN in the form of a transdermal patch would increase the internal diameter of the radial artery and facilitate peripheral artery cannulation in neonates.

Methods

Study design

This single-center, randomised controlled trial was conducted in the neonatal intensive care unit (NICU) at the Perth Children’s Hospital, Western Australia from August 2023 to July 2024. The study was approved by the institutional ethics committee (RGS 0000005763) and conducted according to a protocol that has been previously published. The trial was registered with the Australian and New Zealand Clinical Trials Registry (ACTRN 12623000509662p).21

Participants

Hemodynamically stable neonates (gestation >34 weeks) admitted to the NICU were eligible for recruitment. We excluded neonates with previous attempts at radial arterial cannulation, hematoma at the cannulation site, abnormal Allen’s test, hypercoagulable state, coagulopathy, and peripheral vascular disease. Other exclusion criteria were unstable vital signs, including hypotension, shock, significant arrhythmias, increased intracranial pressure, intracranial haemorrhage, and recent use of Sildenafil. We also excluded neonates undergoing cardiac surgery and neonates with visible deformity in the radial artery area.

Randomization, allocation concealment and blinding

Randomisation to GTN patch or placebo group was done by the clinical trial pharmacist using a computer-generated randomisation sequence. Allocation concealment was achieved using serially numbered opaque sealed envelopes. The envelopes were opened only after entering the key neonatal demographic details following informed parental consent. CobanTM adhesive bandage was applied topically at the site of the radial artery to ensure blinding.

GTN groups

A 5 mg GTN patch was applied at the radial arterial cannulation site for 30 min (Group A: n = 15) or 45 min (Group B: n = 15) or 60 min (Group C: n = 15). The GTN patch was immediately covered with a CobanTM bandage to ensure blinding. Dose selection was based on pharmacokinetic data from previous studies.8,16,18,21

Control group (Group D: N = 15)

A CobanTM adhesive bandage without GTN was applied at the site of radial artery cannulation for 30 min (n = 5) or 45 min (n = 5), or 60 min (n = 5).

To our knowledge, Hasanin et al. is the only pediatric (not neonatal) study using GTN patch for radial arterial cannulation. The median (quartiles) GTN dose in their study was 5.6 (4.4–6.3) microgram/kg/h, whereas the dose we used was 30–60 μg/kg/h. We used the full 5 mg patch to maintain patch integrity and avoid uncertain drug delivery kinetics. Hasanin et al. have suggested that since the dose in their study was not associated with hypotensive episodes, future studies evaluating higher doses of GTN could be considered.8 As a higher dose of GTN was used in our study, we implemented robust safety measures to monitor adverse effects of GTN. An independent Data and Safety Monitoring committee (DSMC) reviewed the study data following recruitment of 20 and 40 participants for the key safety outcomes and advised continuation of the trial.

Blinding

Clinicians and study investigators were blinded to the group allocation. The ultrasound measurements of the radial artery were obtained by an operator who was blinded to the intervention. The bedside nurse who measured blood pressure every 15 min until removal of the intervention was also blinded.

A nurse not involved in neonatal management was responsible only for opening the envelopes, group assignment, and application and removal of GTN patch or placebo.

Measurements and study endpoint

The RAD was measured using a linear, high-frequency hockey stick probe (6–13 MHz) on the Philips EPIQ CVx Diagnostic Ultrasound system. Measurements were standardized and taken in the short-axis plane, in the distal forearm, medial to the border of the styloid process of the radius, with the wrist in a neutral position.22 The operators were experienced neonatologists who routinely perform ultrasounds. Measurements were taken from both radial arteries at baseline before application of the GTN patch/no patch and again at 30-min or 45-min or 60-min after application. Three separate measurements were recorded, and the average of anteroposterior (AP) and transverse diameters was calculated separately. The primary endpoint of the study was the RAD.

The mean and systolic blood pressure were recorded every 15 min until 15 min after the endpoint. Development of clinically significant hypotension (defined as >30% drop in systolic blood pressure from the baseline) and/or sudden increase in oxygen requirement with a rise in methemoglobin levels on arterial blood gas, was the indication for withdrawal from the study.

Primary and secondary outcomes

The primary outcome was the change in RAD in the ipsilateral limb. Secondary outcomes were post-intervention cross-sectional RAD and percentage change in RAD in the ipsilateral limb. Post-intervention absolute and percentage change from baseline RAD in the contralateral limb was also assessed. The safety of the topical GTN patch was assessed by monitoring for hypotension and methemoglobinemia.

Data collection

Data was collected from randomisation until the study endpoint, captured into an electronic database. The trial involved the collection of clinical data and radial artery images by ultrasound measurement. They were stored in a password-protected database on the hospital computer. An independent DSMC was established to monitor patient safety and data quality.

Statistical analysis

Based on previous studies, we assumed a baseline RAD of 0.76 mm and a common within-group standard deviation of 0.18 mm. Recruitment of 60 neonates was estimated to be adequate for ensuring >80% power for a one-way Analysis of variance (ANOVA) (at alpha=0.05) to detect at least a 25% increase from baseline in mean RAD across the GTN groups. Such an increase in the RAD is of the order previously observed in children aged 0–2 years receiving subcutaneous injections of GTN19 and substantially less than that observed after patch application in children aged 2–8.8

Statistical analysis was performed using R (https://cran.rstudio.com/) within the RStudio environment (https://posit.co/products/open-source/rstudio/). Cross-sectional RAD, change in baseline and safety outcomes are presented as a mean and standard deviation. A one-way ANOVA was performed for the principal analyzes of GTN patch effectiveness, with post-hoc comparisons undertaken using Tukey’s HSD test for multiple comparisons. Linear regression was used to extend the ANOVA analyzes to investigate the association with duration of application and to include model adjustments for demographic variables (sex, weight, corrected age and medical/surgical admission).

Results

A total of 100 neonates were screened for eligibility between August 2023 and July 2024. Consent for participation was obtained from 60 of these 100 cases. Forty-five of these 60 neonates were randomly assigned to the GTN group (15 neonates per 30, 45 and 60-min subgroups) and 15 were assigned to the control group (Fig. 1). All 60 were included in the final analysis. Their baseline characteristics are summarised in Table 1.

Fig. 1
Fig. 1The alternative text for this image may have been generated using AI.
Full size image

The CONSORT flow diagram.

Table 1 Patient and clinical characteristics.
Full size table

Primary outcomes

The absolute change from baseline in ipsilateral mean (SD) AP RAD was 0.06 mm (0.08) in the placebo group vs 0.30 mm (0.16) to 0.31 mm (0.17) in the GTN groups (overall GTN mean (SD) 0.31 mm (0.15)), with significant difference in means across the groups by a one-way ANOVA (F = 11.2, p < 0.001, Table 2, Supplementary Fig. 1). Similarly, the absolute change from baseline in ipsilateral mean (SD) transverse RAD was −0.01 (0.07) in the placebo group vs 0.41 (0.23) to 0.45 (0.13) in the GTN groups (ANOVA F = 37.5, p < 0.001, GTN patch overall mean (SD) 0.43 (0.15)) (Supplementary Table 1). For both AP RAD and transverse RAD measures, Tukey’s HSD Test for multiple comparisons showed that the mean change from baseline in the placebo group was significantly less than the increase observed in each of the GTN patch groups considered separately (all p < 0.001). No significant differences were found between any of the mean changes in the GTN 30-min, 45-min or 60-min groups (p > 0.8 for all pairwise comparisons). A lack of association with duration of application was confirmed by linear regression analysis (p > 0.9). The finding of increased vasodilation with GTN patch application remained robust after accounting for potential confounders (sex, weight, corrected age, and medical vs surgical admission) (Table 3).

Table 2 Changes from baseline in radial arterial diameter.
Full size table
Table 3 Regression modelling of absolute change from baseline in RAD outcomes against treatment arm (GTN patch application versus placebo) and duration of patch application, with and without adjustment for demographic variables.
Full size table

Secondary outcomes

Following GTN patch application, the ipsilateral limb showed an average 37% increase in AP-RAD and 35% increase in transverse-RAD from baseline. In the contralateral limb, the increases relative to baseline were 26% for AP-RAD and 28% for transverse-RAD (Supplementary Table 1).

baseline, RAD was similar across the four intervention groups (Supplementary Table 2). However, post-intervention, a significant increase in diameter was observed in neonates allocated to the GTN patch, regardless of the duration of application (Supplementary Fig. 1, and Tables 2 and 3). A similar increase in diameter was noted in the contralateral limb (Supplementary Fig. 2, and Tables 2 and 3). The increase from baseline in contralateral RAD was significantly greater in both the AP and transverse diameters compared to placebo (Table 2). This was similar for all patch durations (Supplementary Fig. 2 and linear regression analyzes, Table 3).

Safety outcomes

Data on methemoglobin, blood pressure, heart rate and oxygen saturation recorded at baseline and post-intervention is shown in Table 4. The baseline diastolic and mean blood pressure were significantly lower in the placebo versus the three GTN groups. The post-intervention blood pressures were not significantly different between the four groups. There was a mean drop of up to 7 mm Hg in diastolic and mean blood pressures from the baseline after 30–60 min of GTN application. None of the neonates in the GTN patch group developed clinically significant hypotension or methemoglobinemia during the study.

Table 4 Safety outcomes.
Full size table

Discussion

The results of our RCT showed that application of a topical GTN patch increased the RAD in both ipsilateral and contralateral limbs in hemodynamically stable neonates born at >34 weeks of gestation, without significant adverse effects. Additionally, the increase in RAD remained consistent irrespective of the duration of patch application. These findings suggest the potential of GTN for increasing the success of first attempt cannulation of the radial artery in this population. The increase in RAD on the contralateral side could be due to the vasodilatory effect of GTN following its systemic absorption or reflex dilatation. These findings may translate into successful cannulation of arteries at other sites not directly treated with topical GTN when attempts at the primary site are unsuccessful. It is reassuring that we did not observe any significant adverse effects despite the use of higher doses of GTN compared to previous studies.8,19,21,23

Our results have several clinical implications, considering the difficulties in cannulating the small-calibre arteries in neonates. Application of the GTN patch has the potential to complement ultrasound-guided arterial cannulation in this population. Furthermore, GTN may prevent or reduce the risk of limb ischemia associated with peripheral arterial cannulation. It has been successfully used in neonates to manage limb ischemia resulting from catheter-induced peripheral arterial injury.23,24 Topical GTN patch provides a low-cost and potentially effective strategy for assisting peripheral arterial cannulation in resource-limited setups.

Our findings are consistent with the recent studies in pediatric and adult populations reporting benefits of GTN in improving RAD and cannulation success without significant hypotension.18,20,25 Given the need for reliable arterial access in critically ill neonates and the challenges associated with its insertion, topical GTN may be a promising strategy in enhancing the success of peripheral arterial catheterisation and improving clinical outcomes, with careful monitoring of its systemic effects. Our findings will be helpful in guiding further research and clinical practice. Future trials could evaluate first-attempt arterial cannulation success rates, increased longevity of arterial cannulas by preventing vasospasm, and reduced risk of complications associated with peripheral arterial cannulation. The efficacy of the GTN patch can be assessed as a standalone intervention or as an adjunct to ultrasound guidance or transillumination.

The limitations of our study need to be acknowledged. First, we observed an imbalance in baseline diastolic and mean blood pressures between the study groups, perhaps as a chance finding. Such an imbalance in baseline characteristics is not uncommon in small RCTs.26,27 The fact that none of the participants in our RCT developed clinically significant hypotension is reassuring. Second, considering many clinicians were involved in acquiring the ultrasound images, some variation in the exact point of image capture is possible despite clearly defined anatomical landmarks. Third, the absence of formal assessment of inter- and intra-observer variability introduces the risk of interpretation drift, potentially weakening the reliability of the data. However, we attempted to mitigate this risk by providing standardized training to all investigators handling the radial artery scans and assuring strict adherence to the protocol for standardized measurements. Fourth, while our results in hemodynamically stable late preterm and term neonates are encouraging, their validity in critically ill and hemodynamically unstable neonates remains to be studied. There was a mean drop of 7 mm Hg from the baseline in diastolic and mean blood pressures 30–60 min after GTN application in the infants in our study, who were hemodynamically stable. A similar reduction in blood pressure in critically ill neonates could have adverse consequences. Furthermore, changes in systemic and regional perfusion may be underestimated when assessed solely by blood pressure and heart rate, a limitation that is particularly relevant in critically ill neonates. This safety consideration warrants further evaluation in future studies. Additionally, we did not include preterm infants (<34 weeks), the population where the success rates of peripheral arterial cannulation are lower and the risk of complications is higher.28 The use of GTN to enhance arterial diameter or improve arterial cannulation success has not been formally studied in preterm infants. However, its safety and efficacy in the treatment of peripheral limb ischemia following arterial cannulation support the feasibility of RCTs of the topical GTN patch in this high-risk population.23,24,29,30 The strengths of our study include its robust design as an RCT powered adequately for the primary outcome. To the best of our knowledge, this is the first RCT in the neonatal population to assess the efficacy of GTN in improving RAD for facilitating peripheral arterial cannulation in neonates.

Conclusion

To our knowledge, this is the first RCT evaluating the effects of topical GTN application on RAD in neonates. The results of our RCT show that application of a GTN patch increases the RAD on the ipsilateral as well as contralateral side and appears to be safe in hemodynamically stable neonates born >34 weeks’ gestation. However, given the potential for systemic effects, the use of topical GTN in hemodynamically unstable infants warrants caution and further evaluation.