Background and purpose The safety of using adult-sized neuroendovascular devices in the smaller pediatric vasculature is not known. In this study we measure vessel diameters in the cervical and cranial circulation in children to characterize when adult-approved devices might be compatible in children.
Methods For 54 children without vasculopathy (mean age 9.5±4.9 years (range 0.02–17.8), 20F/34M) undergoing catheter angiography, the diameters of the large vessels in the cervical and cranial circulation (10 locations, 611 total measurements) were assessed by three radiologists. Mean±SD diameter was calculated for the following age groups: 0–6 months, 1, 2, 3, 4, 5–9, 10–14, and 15–18 years. To compare with adult sizes, each vessel measurement was normalized to the respective region mean diameter in the oldest age group (15–18 years). Normalized measurements were compared with age and fitted to a segmented regression.
Results Vessel diameters increased rapidly from 0 to 5 years of age (slope=0.069/year) but changed minimally beyond that (slope=0.005/year) (R2=0.2). The regression model calculated that, at 5 years of age, vessels would be 94% of the diameter of the oldest age group (compared with 59% at birth). In addition, most vessels in children under 5, while smaller, were still potentially large enough to be compatible with many adult devices.
Conclusions The growth curve of the cervicocerebral vasculature displays rapid growth until age 5, at which point most children's vessels are nearly adult size. By age 5, most neuroendovascular devices are size-compatible, including thrombectomy devices for stroke. Under 5 years of age, some devices might still be compatible.
- Vessel Wall
Statistics from Altmetric.com
Cerebral angiography has been used with increasing frequency in the diagnosis and treatment of pediatric vascular diseases over the last decade.1 In certain vascular diseases such as dural arteriovenous fistulas (AVFs) or vein of Galen malformations, endovascular treatment is considered the primary treatment modality2 ,3 while adjunctive endovascular embolization is also commonly used for arteriovenous malformations (AVMs) and tumors. A number of studies have looked at the safety of both diagnostic catheter angiography and embolization procedures in children,4–7 and the rates of complication are comparable to the published adult safety profile.8–10 Despite the comparable rates of complication, there is concern among neurointerventionalists about the safety of using devices approved for use in adults (ie, those not specifically tested in children) in the smaller pediatric vasculature.11
To date, no studies exist which assess at what age pediatric cervicocerebral vessels reach a caliber similar to that of adults from radiologic or cadaveric studies. Here we present our findings for pediatric vessel size measurements in children seen at our pediatric hospital who underwent digital subtraction angiography between September 2012 and December 2014.
A Philips Allura Xper FD20/20 (Philips Healthcare, Andover, Massachusetts, USA) biplane angiography suite was installed at our institution in 2012. Images acquired from this machine allowed for direct vessel measurements to be taken with our imaging viewer, IMPAX (Agfa Healthcare, Belgium). All patients aged <18 years who underwent craniocervical catheter angiography with or without intervention were identified from billing records. Patients were excluded if they had cervical or cranial vascular pathology known to influence lumen diameter, including fibromuscular dysplasia, moyamoya disease, significant high-flow shunting, or other causes of craniocervical vessel abnormalities/narrowing (ie, radiation-induced vasculopathy). In children who had multiple angiograms, if the interval between angiograms was <1 year, then only the study with the largest number of injected vessels was included for measurement. A total of 54 patients met the inclusion criteria and these images were included for analysis. Additional data collected for these patients included indication for angiography, age at time of angiography, body mass index (BMI), intervention performed (if any), and any procedure-related complications.
Vessel size measurements
Ten locations within the anterior and posterior craniocervical arterial circulation were selected for measurements. These included the common carotid artery, internal carotid artery (ICA), high cervical ICA, horizontal segment of the petrous ICA, ICA terminus, midpoint of the M1 segment of the middle cerebral artery (MCA), high cervical (distal V2) segment of the vertebral artery (VA), intracranial (mid V4) segment of the VA (near the posterior inferior cerebellar artery take-off), the mid basilar artery, and the proximal P1 of the posterior cerebral artery. The rationale for these locations was based on common locations where catheters of various sizes and types may be used.
Three specialty-trained radiologists (two attending neuroradiologists, one chief radiology resident) performed vessel measurements from the 54 available angiograms. In order to assess vessel measurement reliability, 14 angiograms were measured by all three radiologists. The remaining 40 angiograms were divided equally among the radiologists for vessel measurements. The radiologists were blinded to patient age and indication for angiography. In cases where there was concern for abnormal vessel dilation secondary to identifiable pathology (eg, high flow related to an AVM), these vessels were excluded from measurement. For example, if a right-sided AVM had dilated vessels from the right ICA, these vessels would be excluded from measurement; however, the left ICA and/or posterior circulation vessels may still be included in the measurement and subsequent analysis if deemed by the radiologist to be without pathologic enlargement.
Additionally, the diameter of the common femoral artery (CFA) at the insertion of the groin sheath site was measured for all children when available. The decision to perform a CFA angiogram was at the discretion of the proceduralist if there was a likelihood that a groin closure device may be used; in our practice this is reserved for children older than 12 with larger body habitus. In an effort to minimize radiation, CFA angiograms are not routinely performed in children.
All statistical analysis was performed using SPSS (IBM). Descriptive statistics for the cohort were calculated, reporting continuous variables as mean±SD and categorical variables as number (%). Mean±SD diameter of each vascular region was calculated for the following age groups: 0–6 months, 1, 2, 3, 4, 5–9, 10–14, and 15–18 years. The intraclass correlation coefficient (ICC) for each of the 10 measurements was calculated for the 14 common angiograms measured by all three radiologists.
Each vessel measurement was normalized to the respective region mean diameter in children aged 15–18 years, where vessels were most comparable to adults. Normalized measurements were plotted against age and fitted to a segmented regression, with the breakpoint at 5 years of age. This breakpoint was selected after review of the initial scatterplot. Post hoc, to account for the potential confounding effect of BMI on vessel size, linear regression was performed using BMI and age as independent variables and normalized vessel diameter as the dependent variable. Two analyses were performed: (1) patients <5 years of age; and (2) patients >5 years of age. While dilated vessels from known high-flow vascular malformations were excluded, to ensure that the disease process did not globally affect contralateral vessels, additional analysis was performed. Linear regression was performed using age and high-flow vascular process (AVM, AVF, or Vein of Galen aneurysmal malformation (VGAM); n=21) as independent variables and normalized cerebral vessel diameter as the dependent variable. Also, for patients with CFA data, correlation with CFA diameter and BMI using linear regression analysis was performed.
During the study period, 54 consecutive children who underwent catheter angiography met the inclusion criteria. Their mean age was 9.5±4.9 years (range 0.0–17.8 years). There were 20 (37%) female and 34 (63%) male patients. The mean BMI was 18.8±5.8 kg/m2.
Indications for angiography
Of the 54 angiograms analyzed, 38 were diagnostic angiograms, 6 were Wada tests, and the remaining 10 were interventions (table 1). The most common indication for diagnostic angiography was AVM (n=15, 39%) and the most common intervention procedure was for embolization of a ruptured vascular lesion (n=4, 40%). Five patients presented with acute hemorrhage, four of which underwent embolization (two for AVM, one for pseudoaneurysm, one for dural AVF). Three extracranial lesions were embolized including a lingual artery for tonsillar hemorrhage and two embolizations of facial lesions. Of the tumor interventions, one underwent intra-arterial chemotherapy for retinoblastoma and one underwent sacrifice of the left VA prior to open surgical resection of a chordoma encasing the vessel. One patient was treated with mechanical thrombectomy for acute basilar artery occlusion.
Safety of angiography in children
There were no cases with vessel dissection or perforation during angiography and there was no procedure-related mortality. There was a single intraprocedural complication (1.9%): transient MCA thrombus formation during lingual artery embolization which resolved with intra-arterial eptifibatide without clinical consequence. There were no complications associated with Wada testing, diagnostic angiography, or other procedures. No complications related to groin access (eg, hematoma, thrombus, loss of pulse) were reported. This rate of complications is comparable to previously reported safety profiles of pediatric angiography.6
The diameters of the large vessels in the cervical and cranial circulation were assessed by three radiologists in 10 anatomic locations, with a total of 611 measurements. The contralateral unaffected vessels were also measured if imaging was available for measurement. ICC ranged from 0.88 (intracranial VA) to 0.99 (horizontal petrous ICA). Vessel diameters increased rapidly between 0 and 5 years of age (slope=0.069/year) but changed minimally beyond that (slope=0.005/year) (figure 1, R2=0.2). The regression model calculated that, at 5 years of age, vessels would be 94% of the diameter of the older age group compared with 59% at birth (table 2).
Univariate analysis initially identified an association between BMI and vessel diameter. However, in multivariate analysis, BMI was not independently associated with normalized vessel diameter for patients >5 years of age (p=0.237) or patients <5 years of age (p=0.785). However, for patients <5 years of age, the variance inflation factor was 3.651 for both age and BMI, suggesting a non-random interaction between these factors.
For the 21 patients with high-flow vascular lesions, when adjusted for age there was no significant relationship between contralateral high-flow vascular lesion and normalized cerebral vessel diameter (B=0.025, p=0.080). This result indicates that an association between high-flow vascular lesions leading to global increases in vessel diameter was not present.
Of the 11 patients in our study with angiographic imaging of the CFA (mean age 14.8±2 years (range 10–17 years); mean BMI 23.7±6.5 kg/m2), the mean CFA diameter was 7.8±0.2 mm. There was no statistically significant relationship between BMI and CFA diameter (p=0.12).
Over the last decade, utilization of endovascular treatment for pediatric vascular disease has been increasing.1 A number of large cases series have reported the safety of both diagnostic angiography4–6 and interventional7 ,12 procedures in children. These studies conclude that, in experienced large centers, pediatric angiography and neurointervention have morbidity and mortality rates similar to adult rates.8–10 Despite concerns about the ability of pediatric vessels to accommodate endovascular devices designed for and approved only for use in the adult vasculature,13 we were unable to find any radiographic or anatomic studies related to pediatric cervicocerebral vessel size.
Our current study indicates that pediatric craniocervical arteries reach 94% of their final size by age 5. There was rapid linear growth of mean arterial vessel size with age from 0 to 5 years; at 5 years the arterial vessel diameter was approximately 94% of the adult diameter; and growth tapered off thereafter with the slope of growth decreasing by a factor of 10. BMI and sex were not found to be statistically significantly associated with vessel diameter size in multivariate regression. These data mirror an earlier cadaveric study which found that brain weight demonstrates a rapid period of growth from birth to 5 years, at which point the brain is 95% of the weight of adults, with only marginal increases during the rest of childhood and adolescence.14
One limitation of this study is that we specifically address only the pediatric vessel diameter and not other factors such as vasospasm. While smaller vessels in children under 5 are large enough to accommodate adult devices in most locations (tables 2 and 3), there is an increased risk of vasospasm. Specifically, Franken et al15 showed that vasospasm increases with decreasing arterial/catheter (A/C) difference, defined as 100 × (arterial diameter − catheter diameter)/artery diameter. Vasospasm can place vessels at injury for dissection, impede blood flow to distal vessels, and increase procedural risk. For example, using a Neuron Max guide catheter (outer diameter 2.67 mm) in a child aged <6 months in the high cervical ICA (mean diameter 3.9 mm) results in A/C difference of 32%, correlating with a higher likelihood of moderate to severe spasm. In comparison, using the same catheter and location in a 5-year-old (mean diameter 7 mm) results in A/C difference of 62%, placing the vessel at low risk for spasm. While the A/C difference was only used to study spasm in the femoral artery, it is reasonable to extrapolate its conclusions to help infer the risk of spasm in other pediatric vessels undergoing placement of catheters.
The blood vessels of children aged >5 years should be able to accommodate the use of adult-sized 6 Fr guide catheters. This guide catheter access allows the use of almost all neuroendovascular devices currently used in adults. For treatment of acute ischemic strokes, stent-retriever devices such as the Solitaire (Covidien, California, USA) are rated for use in arteries as small as 2 mm and its use has been previously described in children as young as 7 years.16 For treatment of aneurysms, adjunctive techniques including balloon remodeling and stent-assisted coiling should also be possible in children aged >5 years.17–19 Treatment of arteriovenous shunts with Onyx or n-Butyl Cyanoacrylate (n-BCA) via compatible microcatheters and/or intermediate catheters can be undertaken as one would in adults.20 ,21
In children under 5 years, the use of smaller guide catheters is necessary. Techniques using 4 Fr and 5 Fr guide catheters for endovascular treatment in young children have been previously described.22 Given the smaller inner diameter of these guide catheters, intermediate catheters may not be compatible for use and access to distal vasculature may be obtained directly with microcatheters. In line with this, treatment of aneurysms must be accomplished with primary coil embolization, or in a sequential manner should stent assistance be necessary. For treatment of arteriovenous shunts, use of dual-lumen balloon microcatheters can still be used as previously described.22
Sonographic data of CFA diameter suggest that the artery inner diameter increases with age. Sandgren et al23 found that at around the age of 12 years it is approximately 7 mm in diameter compared with approximately 8 mm in 25-year-olds. While these authors did not obtain measurements in children aged <8 years, extrapolation of estimates from their logarithmic regression would indicate that at 5 years of age the CFA diameter should be approximately 6 mm, and at birth approximately 5 mm. In the context of CFA vasospasm risk, the CFA A/C difference at birth is 47%, assuming use of a NeuronMax guide catheter (largest OD catheter) in a CFA estimated to be 5 mm in diameter (extreme example); this confers a risk of mild-to-moderate vasospasm.15 Using smaller 4 Fr catheters in these small children, as is convention, may reduce vasospasm risks as this hypothetically increases the A/C ratio to the 73–78% range which is associated with minimal risk of vasospasm.
In our population of pediatric patients with femoral arterial angiography, the mean age was 14.8 years and the mean femoral artery diameter was 7.8 mm, which is consistent with the findings of Sandgren et al.23 Those authors additionally found a positive relationship between femoral artery diameter and body surface area, which suggests that this measure might also be a useful surrogate for predicting CFA size in order to select a safe guide catheter size. We did not observe any femoral artery complications in our series of pediatric patients.
From a size perspective, neurointerventionalists should consider patients aged ≥5 years as having vessels comparable to adults. Beyond the age of 5 years the majority of adult endovascular devices are compatible with arterial sizes. Under 5 years of age smaller guide catheters may be necessary to minimize the risk of vasospasm; however, microcatheters and endovascular devices are usually still compatible for use even in the more distal vasculature. BMI may not provide additional information beyond age when predicting cervicocerebral vessel sizes in children.
Contributors LH designed the data collection tools, monitored data collection for the study, wrote the statistical analysis plan, cleaned and analysed the data, and drafted and revised the paper; she is guarantor. TRL designed the data collection tools, monitored data collection for the study, wrote the statistical analysis plan, cleaned and analysed the data, and drafted and revised the paper. SP, MAD and AAD collected data points and approved the final manuscript. LCJ monitored data collection for the study, revised the paper, and approved the final manuscript. MTF designed the data collection tools, critically revised and approved the final paper, and oversaw the project; he is guarantor.
Competing interests None declared.
Ethics approval Institutional Review Board of Vanderbilt University Medical Center.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.