Resident operative experience in pediatric neurosurgery across the United States.

OBJECTIVE: Neurosurgical residents receive exposure to the subspecialty of pediatric neurosurgery during training. The authors sought to determine resident operative experience in pediatric neurosurgery across Accreditation Council for Graduate Medical Education (ACGME)-accredited neurosurgical programs. METHODS: During 2018-2019, pediatric neurosurgical case logs for recent graduates or current residents who completed their primary pediatric exposure were collected from US continental ACGME training programs. Using individual resident reports and procedure designations, operative volumes and case diversity were analyzed collectively, according to training site characteristics, and also correlated with the recently described Resident Experience Score (RES). RESULTS: Of the 114 programs, a total of 316 resident case logs (range 1-19 residents per program) were received from 86 (75%) programs. The median cumulative pediatric case volume per resident was 109 (IQR 75-161). Residents at programs with a pediatric fellowship reported a higher median case volume (143, IQR 96-187) than residents at programs without (91, IQR 66-129; p < 0.0001). Residents at programs that outsource their pediatric rotation had a lower median case volume (84, IQR 52-114) compared with those at programs with an in-house experience (117, IQR 79-170; p < 0.0001). The case diversity index among all programs ranged from 0.61 to 0.80, with no statistically significant differences according to the Accreditation Council for Pediatric Neurosurgery Fellowships designation or pediatric experience site (p > 0.05). The RES correlated moderately (r = 0.44) with median operative volumes per program. A program's annual pediatric operative volume and duration of pediatric experience were identified as significant predictive factors for median resident operative volume. CONCLUSIONS: Resident experience in pediatric neurosurgery is variable within and between programs. Case volumes are generally higher for residents at programs with in-house exposure and an accredited fellowship, but case diversity is relatively uniform across all programs. RES provides some insight on anticipated case volume, but other unexplained factors remain.

Association between supratentorial pediatric high-grade gliomas involved with the subventricular zone and decreased survival: a multi-institutional retrospective study.

OBJECTIVE: The subventricular zone (SVZ), housed in the lateral walls of the lateral ventricles, is the largest neurogenic niche in the brain. In adults, high-grade gliomas in contact or involved with the SVZ are associated with decreased survival. Whether this association holds true in the pediatric population remains unexplored. To address this gap in knowledge, the authors conducted this retrospective study in a pediatric population with high-grade gliomas treated at three comprehensive centers in the United States. METHODS: The authors retrospectively identified 63 patients, age ≤ 21 years, with supratentorial WHO grade III-IV gliomas treated at three academic centers. Basic demographic and clinical data regarding presenting signs and symptoms and common treatment variables were obtained. Preoperative MRI studies were evaluated to assess SVZ contact by tumor and to quantify tumor volume. RESULTS: Sixty-three patients, including 34 males (54%), had a median age of 12.3 years (IQR 6.50-16.2) and a median tumor volume of 39.4 ml (IQR 19.4-65.8). Tumors contacting the SVZ (SVZ+) were noted in 34 patients (54%) and overall were larger than those not in contact with the SVZ (SVZ-; 51.1 vs 27.3, p = 0.002). The SVZ+ tumors were also associated with decreased survival. However, age, tumor volume, tumor grade, and treatment with chemotherapy and/or radiation were not associated with survival in the 63 patients. In the univariable analysis, near-total resection, gross-total resection, and seizure presentation were associated with increased survival (HR = 0.23, 95% CI 0.06-0.88, p = 0.03; HR = 0.26, 95% CI 0.09-0.74, p = 0.01; and HR = 0.46, 95% CI 0.22-0.97, p = 0.04, respectively). In a multivariable stepwise Cox regression analysis, only SVZ+ tumors remained significantly associated with decreased survival (HR = 1.94, 95% CI 1.03-3.64, p = 0.04). CONCLUSIONS: High-grade glioma contact with the SVZ neural stem cell niche was associated with a significant decrease in survival in the pediatric population, as it is in the adult population. This result suggests that tumor contact with the SVZ is a general negative prognosticator in high-grade glioma independent of age group and invites biological investigations to understand the SVZ's role in glioma pathobiology.

Converting Pediatric Patients and Young Adults From a Shunt to a Third Ventriculostomy: A Multicenter Evaluation.

BACKGROUND: Endoscopic third ventriculostomy (ETV) is an effective primary treatment for certain forms of hydrocephalus. However, its use in children with an existing shunt is less well known. OBJECTIVE: To report a multicenter experience in attempting to convert patients from shunt dependence to a third ventriculostomy and to determine predictors of success. METHODS: Three participating centers provided retrospectively collected information on patients with an attempted conversion from a shunt to an ETV between December 1, 2008, and April 1, 2018. Demographic, clinical, and radiological data were recorded. Success was defined as shunt independence at the last follow-up. RESULTS: Eighty patients with an existing ventricular shunt underwent an ETV. The median age at the time of the index ETV was 9.9 yr, and 44 (55%) patients were male. The overall success rate was 64% (51/80), with a median duration of follow-up of 2.0 yr (range, 0.1-9.4 yr). Four patients required a successful repeat ETV at a median of 1.7 yr (range, 0.1-5.7 yr) following the index ETV. Only age was predictive of ETV failure on multivariate analysis (odds ratio 0.86 [95% CI 0.78-0.94], P = .005). No patient less than 6 mo of age underwent an ETV, and of the 5 patients between 6 and 12 mo of age, 4 failed. CONCLUSION: Although not every shunted patient will be a candidate for an ETV, nor will they be successfully converted, an ETV should at least be considered in every child who presents with a shunt malfunction or who has an externalized shunt.

Application of the Preventable Shunt Revision Rate to an Adult Population.

BACKGROUND: The preventable shunt revision rate (PSRR) was recently introduced in pediatric hydrocephalus as a quality metric for shunt surgery. We evaluated the PSRR in an adult hydrocephalus population. METHODS: All ventricular shunt operations (January 1, 2013 to March 31, 2018) performed at a university-based teaching hospital were included. For any index surgery (de novo or revision) resulting in reoperation within 90 days, the index surgery details were collected, and a consensus decision was reached regarding whether the failure had been potentially avoidable. Preventable failure was defined as failure due to infection, malposition, disconnection, migration, or kinking. The 90-day shunt failure rate and PSRR were calculated. Bivariate analyses were performed to evaluate the individual effects of each independent variable on preventable shunt failure. RESULTS: A total of 318 shunt operations had been performed in 245 patients. Most patients were women (62%), with a median age of 48.2 years (interquartile range, 31.2-63.2 years). Most had had ventriculoperitoneal shunts placed (86.5%), and just more than one half were new shunts (51.6%). A total of 53 cases (16.7%) in 42 patients experienced shunt failure within 90 days of the index operation. Of these, 27 failures (8.5% of the total cases; 51% of the failures) were considered potentially preventable. The most common reasons were infection (37%; n = 10) and malposition of the proximal and distal catheters (both 25.9%; n = 7). Age was the only statistically significant difference between the 2 groups, with the patients experiencing preventable shunt failure older than those without preventable shunt failure (51.4 vs. 37.1 years; P = 0.017). CONCLUSIONS: The 90-day PSRR can be applied to an adult population and serve as a quality metric.

Shunt Failure-The First 30 Days.

BACKGROUND: Incontrovertible predictors of shunt malfunction remain elusive. OBJECTIVE: To determine predictors of shunt failure within 30 d of index surgery. METHODS: This was a single-center retrospective cohort study from January 2010 through November 2016. Using a ventricular shunt surgery research database, clinical and procedural variables were procured. An "index surgery" was defined as implantation of a new shunt or revision or augmentation of an existing shunt system. The primary outcome was shunt failure of any kind within the first 30 days of index surgery. Bivariate models were created, followed by a final multivariable logistic regression model using a backward-forward selection procedure. RESULTS: Our dataset contained 655 unique patients with a total of 1206 operations. The median age for the cohort at the time of first shunt surgery was 4.6 yr (range, 0-28; first and third quartile, .37 and 11.8, respectively). The 30-day failure rates were 12.4% when analyzing the first-index operation only (81/655), and 15.7% when analyzing all-index operations (189/1206). Small or slit ventricles at the time of index surgery and prior ventricular shunt operations were found to be significant covariates in both the "first-index" (P < .01 and P = .05, respectively) and "all-index" (P = .02 and P < .01, respectively) multivariable models. Intraventricular hemorrhage at the time of index surgery was an additional predictor in the all-index model (P = .01). CONCLUSION: This study demonstrates that only 3 variables are predictive of 30-day shunt failure when following established variable selection procedures, 2 of which are potentially under direct control of the surgeon.

Dissecting the Financial Relationship Between Industry and Academic Neurosurgery.

BACKGROUND: Established by the Centers for Medicare and Medicaid Services (CMS), the Open Payments Database (OPD) has reported industry payments to physicians since August 2013. OBJECTIVE: To evaluate the frequency, type, and value of payments received by academic neurosurgeons in the United States over a 5-yr period (2014-2018). METHODS: The OPD was queried for attending neurosurgeons from all neurosurgical training programs in the United States (n = 116). Information from the OPD was analyzed for the entire cohort as well as for comparative subgroup analyses, such as career stage, subspecialty, and geographic location. RESULTS: Of all identified neurosurgeons, 1509 (95.0%) received some payment from industry between 2014 and 2018 for a total of 106 171 payments totaling $266 407 458.33. A bimodal distribution was observed for payment number and total value: 0 to 9 (n = 438) vs > 50 (n = 563) and 0-$1000 (n = 418) vs >$10 000 (n = 653), respectively. Royalty/License was the most common type of payment overall (59.6%; $158 723 550.57). The median number (40) and value ($8958.95) of payments were highest for mid-career surgeons. The South-Central region received the most money ($117 970 036.39) while New England received the greatest number of payments (29 423). Spine surgeons had the greatest median number (60) and dollar value ($20 551.27) of payments, while pediatric neurosurgeons received the least (8; $1108.29). Male neurosurgeons received a greater number (31) and value ($6395.80) of payments than their female counterparts (11, $1643.72). CONCLUSION: From 2014 to 2018, payments to academic neurosurgeons have increased in number and value. Dollars received were dependent on geography, career stage, subspecialty and gender.

Image Guidance for Ventricular Shunt Surgery: An Analysis of Ventricular Size and Proximal Revision Rates.

BACKGROUND: Image guidance is a promising technology that could lead to lower rates of premature shunt failure by decreasing the rate of inaccurate proximal catheter placement. OBJECTIVE: To perform a detailed radiographic analysis of ventricular size using 3 well-described methods and compare proximal revision rates. METHODS: Our shunt surgery research database was queried to identify procedures (new placement or revision) where frameless stereotactic electromagnetic neuronavigation was used (January 2010-June 2016). A randomly selected cohort of surgeries done without image guidance during the same time period served as the comparison group. A radiographic analysis utilizing the following indices was used to classify ventricular size: bifrontal, bicaudate, and frontal-occipital horn ratio. The primary outcome was shunt failure due specifically to proximal catheter malfunction at 90 and 180 days. RESULTS: A total of 108 stereotactic and 95 free-hand cases were identified. Overall, there was no difference in ventricular size between the 2 groups. Neuronavigation yielded improved accuracy rates (73% grade 1; P < .001). Although there was no statistically significant difference in proximal revision rates when all patients were analyzed, there was a clinically beneficial reduction in the 90- and 180-day failure rates across all radiographic indices in children with small-to-moderate ventricular sizes when using image guidance. CONCLUSION: Electromagnetic neuronavigation results in more accurate placement of catheters, but did not result in an overall reduction in proximal shunt failure at 90 and 180 days after the index surgery. However, subgroup analysis suggests a clinically important benefit in those patients with harder to cannulate ventricles.

Pediatric Abusive Head Trauma: Return to Hospital System in the First Year Post Injury.

BACKGROUND: Abusive head trauma (AHT) may result in costly, long-term sequelae. OBJECTIVE: To describe the burden of AHT on the hospital system within the first year of injury. METHODS: Single institution retrospective evaluation of AHT cases from January 2009 to August 2016. Demographic, clinical (including injury severity graded I-III), and charge data associated with both initial and return hospital visits within 1 yr of injury were extracted. RESULTS: A total of 278 cases of AHT were identified: 60% male, 76% infant, and 54% African-American. Of these 278 cases, 162 (60%) returned to the hospital within the first year, resulting in 676 total visits (an average of 4.2 returns/patient). Grade I injuries were less likely to return than more serious injuries (II and III). The majority were outpatient services (n = 430, 64%); of the inpatient readmissions, neurosurgery was the most likely service to be involved (44%). Neurosurgical procedures accounted for the majority of surgeries performed during both initial admission and readmission (85% and 68%, respectively). Increasing injury severity positively correlated with charges for both the initial admission and returns (P < .001 for both). Total calculated charges, including initial admission and returns, were over $25 million USD. CONCLUSION: AHT has a high potential for return to the hospital system within the first year. Inpatient charges dominate and account for the vast majority of hospital returns and overall charges. A more severe initial injury correlates with increased charges on initial admission and on subsequent hospital return.

Pediatric Supratentorial Ependymoma: Surgical, Clinical, and Molecular Analysis.

BACKGROUND: Pediatric supratentorial ependymomas (SEs) have distinct molecular and behavioral differences from their infratentorial counterparts. OBJECTIVE: To present our experience with pediatric SEs over a 24-yr period. METHODS: Clinical, operative, and radiographic information was abstracted retrospectively. Our primary outcomes were progression-free survival (PFS) and overall survival (OS). Detection of C11orf95-RELA rearrangement was performed using interphase fluorescence in situ hybridization (iFISH). RESULTS: Seventy-three patients were identified (41 female, 32 male); median age was 6.7 yrs (range, 1 mo-18.8 yr); median follow-up was 8.3 yrs (range, 2.0-26.3). Fifty-eight (79.5%) of 73 patients underwent gross total resection (GTR); no patient with subtotal resection had greater than 1 cm3 of residual tumor; 42 patients (57.5%) experienced subsequent disease progression with 17 patients ultimately dying of their disease. Median PFS was 3.7 yrs. Molecular analysis was available for 51 patients (70%). On bivariate analysis, PFS and OS were not statistically affected by age, tumor grade, or extent of resection, although there was a clinically significant trend for the latter in favor of aggressive resection on PFS (P = .061). Children with RELA fusion had significantly higher PFS (P = .013) than those without, although there was no difference in OS when compared with those with no C11orf95-RELA fusion or C11orf95 gene rearrangement alone. CONCLUSION: In our series, GTR may be associated with better PFS, but did not impact OS. Surprisingly, RELA fusion was not found to be a negative prognostic factor, raising the possibility that the deleterious effects may be overcome by aggressive resection.

The Preventable Shunt Revision Rate: A Multicenter Evaluation.

BACKGROUND: The Preventable Shunt Revision Rate (PSRR) was recently introduced as a novel quality metric. OBJECTIVE: To evaluate the PSRR across multiple centers and determine associated variables. METHODS: Nine participating centers in North America provided at least 2 years of consecutive shunt operations. Index surgery was defined as new shunt implantation, or revision of an existing shunt. For any index surgery that resulted in a reoperation within 90-days, index surgery information (demographic, clinical, and procedural) was collected and a decision made whether the failure was potentially preventable. The 90-day shunt failure rate and PSRR were calculated per institution and combined. Bivariate analyses were performed to evaluate individual effects of each independent variable on preventable shunt failure followed by a final multivariable model using a backward model selection approach. RESULTS: A total of 5092 shunt operations were performed; 861 failed within 90 days of index operation, resulting in a 16.9% combined 90-day shunt failure rate and 17.6% median failure rate (range, 8.7%-26.9%). Of the failures, 307 were potentially preventable (overall and median 90-day PSRR, 35.7% and 33.9%, respectively; range, 16.1%-55.4%). The most common etiologies of avoidable failure were infection (n = 134, 44%) and proximal catheter malposition (n = 83, 27%). Independent predictors of preventable failure (P < .05) were lack of endoscopy (odds ratio [OR] = 2.26), recent shunt infection (OR = 3.65), shunt type (OR = 2.06) and center. CONCLUSION: PSRR is variable across institutions, but can be 50% or higher. While the PSRR may never reach zero, this study demonstrates that overall about a third of early failures are potentially preventable.

Postgraduate publishing output in pediatric neurosurgery: correlation with fellowship site and individual scholars.

OBJECTIVE: The objective of this study was to analyze the publication output of postgraduate pediatric neurosurgery fellows for a 10-year period as well as identify 25 individual highly productive pediatric neurosurgeons. The correlation between academic productivity and the site of fellowship training was studied. METHODS: Programs certified by the Accreditation Council for Pediatric Neurosurgery Fellowships that had 5 or more graduating fellows from 2006 to 2015 were included for analysis. Fellows were queried using Scopus for publications during those 10 years with citation data through 2017. Pearson correlation coefficients were calculated, comparing program rankings of faculty against fellows using the revised Hirsch index (r-index; primary) and Hirsch index (h-index; secondary). A list of 25 highly accomplished individual academicians and their fellowship training locations was compiled. RESULTS: Sixteen programs qualified with 152 fellows from 2006 to 2015; 136 of these surgeons published a total of 2009 articles with 23,735 citations. Most publications were pediatric-specific (66.7%) clinical articles (93.1%), with middle authorship (55%). Co-investigators were more likely from residency than fellowship. There was a clustering of the top 7 programs each having total publications of around 120 or greater, publications per fellow greater than 12, more than 1200 citations, and adjusted ir10 (revised 10-year institutional h-index) and ih10 (10-year institutional h-index) values of approximately 2 or higher. Correlating faculty and fellowship program rankings yielded correlation coefficients ranging from 0.53 to 0.80. Fifteen individuals (60%) in the top 25 (by r5 index) list completed their fellowship at 1 of these 7 institutions. CONCLUSIONS: Approximately 90% of fellowship-trained pediatric neurosurgeons have 1 or more publications, but the spectrum of output is broad. There is a strong correlation between where surgeons complete their fellowships and postgraduate publications.

The Impact of a Pediatric Shunt Surgery Checklist on Infection Rate at a Single Institution.

BACKGROUND: Shunt infections remain a significant challenge in pediatric neurosurgery. Numerous surgical checklists have been introduced to reduce infection rates. OBJECTIVE: To introduce an evidence-based shunt surgery checklist and its impact on our shunt infection rate. METHODS: Between January 1, 2008 and December 31, 2015, pediatric patients who underwent shunt surgery at our institution were indexed in a prospectively maintained database. All definitive shunt procedures were included. Shunt infection was defined according to the Center for Disease Control and Prevention's National Hospital Safety Network surveillance definition for surgical site infection. Clinical and procedural variables were abstracted per procedure. Infection data were compared for the 4 year before and 4 year after protocol implementation. Compliance was calculated from retrospective review of our checklists. RESULTS: Over the 8-year study period, 1813 procedures met inclusion criteria with a total of 37 shunt infections (2%). Prechecklist (2008-2011) infection rate was 3.03% (28/924) and decreased to 1.01% (9/889; P = .003) postchecklist (2012-2015), representing an absolute risk reduction of 2.02% and relative risk reduction of 66.6%. One shunt infection was prevented for every 50 times the checklist was used. Those patients who developed an infection after protocol implementation were younger (0.95 years vs 3.40 years (P = .027)), but there were no other clinical or procedural variables, including time to infection, that were significantly different between the cohorts. Average compliance rate among required checklist components was 97% (range 85%-100%). CONCLUSION: Shunt surgery checklist implementation correlated with lower infection rates that persisted in the 4 years after implementation.

A Review of Industry Funding in Randomized Controlled Trials Published in the Neurosurgical Literature-The Elephant in the Room.

OBJECTIVE: To analyze the role of industry sponsorship of randomized controlled trials (RCTs) published exclusively in 3 major North American neurosurgical journals. METHODS: Our primary objective was to determine whether an association exists between study conclusion(s) in favor of industry sponsored drugs, devices/implants, or surgical techniques and industry sponsorship. The secondary objective was to describe the quality/quantity of these neurosurgical RCTs. RESULTS: A total of 110 RCTs were analyzed, the majority were published in the Journal of Neurosurgery (85%) and were international in origin (55%). The most common subspecialty was spine (n = 29) and drug study was the most common type (n = 49). Overall quality was good with median Jadad and Detsky scores of 4 (range, 1-5) and 18 (range, 8-21), respectively. There was a statistically significant difference in RCTs with industry funding (31/40, 78%) versus those without (9/70, 13%) that published a favorable conclusion of the new drug, device/implant, or surgical technique (odds ratio [OR], 23.35; P < .0001). Multiple binomial logistic regression analysis identified "number of authors" as mildly protective (OR, 0.79; 95% confidence interval, 0.69-0.91; P = .001) and "industry funding" strongly predictive (OR, 12.34; 95% confidence interval, 2.97-51.29; P = .001) of a positive trial. CONCLUSION: Industry funding was associated with a much greater chance of positive findings in RCTs published in neurosurgical journals. Further efforts are needed to define the relationship between the authors and financial sponsors of neurosurgical research and explore the reasons for this finding.

Taking the next step in publication productivity analysis in pediatric neurosurgery.

OBJECTIVE There has been an increasing interest in the quantitative analysis of publishing within the field of neurosurgery at the individual, group, and institutional levels. The authors present an updated analysis of accredited pediatric neurosurgery training programs. METHODS All 28 Accreditation Council for Pediatric Neurosurgery Fellowship programs were contacted for the names of pediatric neurosurgeons who were present each year from 2011 through 2015. Faculty names were queried in Scopus for publications and citations during this time period. The 5-year institutional Hirsch index [i h(5)-index] and revised 5-year institutional h-index [i r(5)-index] were calculated to rank programs. Each publication was reviewed to determine authorship value, tier of research, clinical versus basic science research, subject matter, and whether it was pediatrics-specific. A unique 3-tier article classification system was introduced to stratify clinical articles by quality and complexity, with tier 3 being the lowest tier of publication (e.g., case reports) and tier 1 being the highest (e.g., randomized controlled trials). RESULTS Among 2060 unique publications, 1378 (67%) were pediatrics-specific. The pediatrics-specific articles had a mean of 15.2 citations per publication (median 6), whereas the non-pediatrics-specific articles had a mean of 23.0 citations per publication (median 8; p < 0.0001). For the 46% of papers that had a pediatric neurosurgeon as first or last author, the mean number of citations per publication was 12.1 (median 5.0) compared with 22.5 (median 8.0) for those in which a pediatric neurosurgeon was a middle author (p < 0.0001). Seventy-nine percent of articles were clinical research and 21% were basic science or translational research; however, basic science and translational articles had a mean of 36.9 citations per publication (median 15) compared with 12.6 for clinical publications (median 5.0; p < 0.0001). Among clinical articles, tier 1 papers had a mean of 15.0 citations per publication (median 8.0), tier 2 papers had a mean of 18.7 (median 8.0), and tier 3 papers had a mean of 7.8 (median 3.0). Neuro-oncology papers received the highest number of citations per publication (mean 25.7). The most common journal was the Journal of Neurosurgery: Pediatrics (20%). MD/PhD faculty members had significantly more citations per publication than MD faculty members (mean 26.7 vs 14.0; p < 0.0001) and also a higher number of publications per author (mean 38.6 vs 20.8). The median i h(5)- and i r(5)-indices per program were 14 (range 5-48) and 10 (range 5.6-37.2), respectively. The mean i r(5)/i h(5)-index ratio was 0.8. The top 5 fellowship programs (in descending order) as ranked by the i h(5)-index corrected for number of faculty members were The Hospital for Sick Children, Toronto; Children's Hospital of Pittsburgh; University of California, San Francisco Benioff Children's Hospital; Seattle Children's Hospital; and St. Louis Children's Hospital. CONCLUSIONS About two-thirds of publications authored by pediatric neurosurgeons are pediatrics-specific, although non-pediatrics-specific articles averaged more citations. Most of the articles authored by pediatric neurosurgeons are clinical, with basic and translational articles averaging more citations. Neurosurgeons with PhD degrees averaged more total publications and more citations per publication. In all, this is the most advanced and informative analysis of publication productivity in pediatric neurosurgery to date.

Pediatric abusive head trauma and stroke.

OBJECTIVE Despite established risk factors, abusive head trauma (AHT) continues to plague our communities. Cerebrovascular accident (CVA), depicted as areas of hypodensity on CT scans or diffusion restriction on MR images, is a well-known consequence of AHT, but its etiology remains elusive. The authors hypothesize that a CVA, in isolation or in conjunction with other intracranial injuries, compounds the severity of a child's injury, which in turn leads to greater health care utilization, including surgical services, and an increased risk of death. METHODS The authors conducted a retrospective observational study to evaluate data obtained in all children with AHT who presented to Le Bonheur Children's Hospital (LBCH) from January 2009 through August 2016. Demographic, hospital course, radiological, cost, and readmission information was collected. Children with one or more CVA were compared with those without a CVA. RESULTS The authors identified 282 children with AHT, of whom 79 (28%) had one or more CVA. Compared with individuals without a CVA, children with a stroke were of similar overall age (6 months), sex (61% male), and race (56% African-American) and had similar insurance status (81% public). Just under half of all children with a stroke (38/79, 48%) were between 1-6 months of age. Thirty-five stroke patients (44%) had a Grade II injury, and 44 (56%) had a Grade III injury. The majority of stroke cases were bilateral (78%), multifocal (85%), associated with an overlying subdural hematoma (86%), and were watershed/hypoperfusion in morphology (73%). Thirty-six children (46%) had a hemispheric stroke. There were a total of 48 neurosurgical procedures performed on 28 stroke patients. Overall median hospital length of stay (11 vs 3 days), total hospital charges ($13.8 vs $6.6 million), and mean charges per patient ($174,700 vs $32,500) were significantly higher in the stroke cohort as a whole, as well as by injury grade (II and III). Twenty children in the stroke cohort (25%) died as a direct result of their AHT, whereas only 2 children in the nonstroke cohort died (1%). There was a 30% readmission rate within the first 180-day postinjury period for patients in the stroke cohort, and of these, approximately 50% required additional neurosurgical intervention(s). CONCLUSIONS One or more strokes in a child with AHT indicate a particularly severe injury. These children have longer hospital stays, greater hospital charges, and a greater likelihood of needing a neurosurgical intervention (i.e., bedside procedure or surgery). Stroke is such an important predictor of health care utilization and outcome that it warrants a subcategory for both Grade II and Grade III injuries. It should be noted that the word "stroke" or "CVA" should not automatically imply arterial compromise in this population.