Differences Between Viral Meningitis and Abusive Head Trauma Free

We performed a retrospective case-control study of subjects aged <2 years at the authors’ institution between January 2014 and December 2018. Controls consisted of children with viral meningitis confirmed by nucleic acid amplification test of cerebral spinal fluid from lumbar puncture. Viral infections identified by nucleic acid amplification test were parechovirus, enterovirus, and HSV. Cases were identified through the child abuse pediatrics (CAP) team database and subdivided into 2 groups: subjects with SDH plus CSIs, serving as the AHT group because of the presence of extracranial CSIs not explained by viral meningitis; and subjects with SDH without CSIs (SDH only). The CAP consultation often captures more detailed information than emergency department or admission notes. To allow more equal comparison among study groups, data from the CAP consultation regarding HPI and clinical presentation (the elements under study) were not included in this study because viral meningitis subjects did not have a CAP evaluation. Subjects were excluded from the study if they had preexisting abnormalities of the brain, including hydrocephalus, intracranial hemorrhage, infection, surgery, tumor/cancer, or anatomic/vascular malformations diagnosed before the study encounter. Subjects with an isolated, small SDH underlying an impact site were excluded, as were those hospitalized for >24 hours at an outside institution before transfer to the study hospital.

Demographic information, medical history, physical examination, laboratory and radiologic findings, and neurologic outcomes were obtained through chart review. Presenting HPI and clinical exam features were recorded exclusively from the initial emergency department and inpatient admission notes.

Definitions of CSIs were based on those used by Hansen et al (Table 1).^24–29  CSIs were used to categorize subjects with SDH into our 2 case groups, therefore were not features under study. Bruises, petechiae, and burns were included as CSIs if they were identified in the initial emergency department or inpatient admission note only. Other CSIs were included if identified at any time during the subject’s evaluation.

Study groups were compared across 3 domains.

Study authors abstracted subjects’ age, sex, race, and insurance status. Information regarding the relationship of the caretakers residing in the home to the child was obtained from the social history documented in emergency department and admission notes or social work notes when available.

HPI features were symptoms reported by the caretaker(s) that prompted the medical evaluation as documented in the emergency department and/or inpatient admission notes. Measures included acute neurologic and respiratory symptoms of seizures, decreased tone, difficulty breathing, apnea, acute mental status change and other ill symptoms of measured or subjective fever, vomiting, cough or congestion, poor feeding, and fussiness. Acute symptoms were defined as those with a sudden, immediate/near immediate onset in a previously well child. Mental status change included any reported change from neurologic baseline such as drowsiness, lethargy, unconsciousness, or irritability. If a history of injury related to the current presentation was documented, this was recorded. “No history of injury” was only recorded if specifically stated.

Clinical features included initial examination findings, laboratory and radiograph features, requirement of intensive care, and discharge neurologic status. Measures included macrocephaly, fever (based on first set of vitals), cough or congestion, difficulty breathing, intubation, witnessed seizure, hypotonia, and neurologic status, as well as initial white blood cell count (WBC). Neurologic status of the child was categorized as normal, abnormal, death, or unknown. Abnormal was defined as active seizures, pupil abnormalities, focal neurologic abnormalities, depressed mental status, or Glasgow coma score <8. Children with new neurologic deficits, including tone and feeding issues or seizures requiring ongoing medication at discharge, were considered to have abnormal neurologic status. Laboratory data were obtained from the first documented laboratories during the hospital encounter. Features of SDH or other intracranial findings were obtained from the radiology report of the first head computed tomography and/or MRI after presentation to the hospital.

Using χ² test, Fisher’s exact test, and Mann-Whitney U test (for age), groups were compared across a total of 25 measures for the 3 domains: the demographics domain (5 measures), HPI domain (11 measures), and clinical domain (9 measures).

Figure 1 depicts study subject identification and classification. Viral meningitis etiologies included enterovirus (N = 254), parechovirus (N = 137) and HSV (N = 6). Table 2 shows details of testing for occult trauma and viral meningitis. By definition, no SDH-only subjects had retinal hemorrhages characteristic of abuse (Table 1); however, 6 did have retinal hemorrhages that were few in number and did not extend to the periphery. Skin findings and oral injury concerning for abuse were documented in 72 (61.0%) and 9 (7.6%) AHT subjects, respectively. A single viral meningitis subject had petechiae documented that was not described as iatrogenic.

Table 3 characterizes the demographics domain. Viral meningitis subjects were significantly younger than AHT and SDH-only subjects. AHT subjects differed from viral meningitis subjects in sex, race/ethnicity, and insurance status. AHT subjects were more likely than viral meningitis and SDH-only subjects to have an unrelated caregiver living in the home. This information was unknown for 18 viral meningitis subjects. The percentage of subjects living with an unrelated caregiver was calculated out of the total number of subjects in each group, including those with an unknown caregiver relationship.

Table 4 details the HPI domain. Compared with viral meningitis subjects, AHT subjects were >300 times more likely to be afebrile, >150 times more likely to have a mental status change, and nearly 60 times more likely to have at least 1 or more acute symptom. Similar but less pronounced differences were seen when comparing SDH-only and viral meningitis subjects. The AHT and SDH-only groups had fewer other ill symptoms, except vomiting, than the viral meningitis group. Finally, few differences existed between the SDH-only and the AHT subjects.

Table 5 details the clinical domain. AHT subjects were more likely to have neurologic and respiratory symptoms and fewer infectious findings when compared with the viral meningitis group. AHT subjects were nearly 20 times more likely than viral meningitis subjects to require intensive care and were less likely to have normal mental status at discharge (Table 6). Though less acutely ill than the AHT group, the SDH-only group was still 12 times more likely to require intubation and nearly 5 times more likely to require intensive care than the viral meningitis group. Fever was significantly less common in the AHT and SDH-only groups than the viral meningitis group. Among subjects with an abnormal WBC, AHT and SDH-only subjects were more likely to have a high WBC than viral meningitis subjects. The presence of combined features of a history of acute mental status change with lack of clinical fever on presentation had a positive predictive value for abuse of 95%.

Nineteen viral meningitis subjects received head imaging (Supplemental Table 7). Of these, 2 subjects aged <10 days old with parechovirus meningitis had posteriorly located SDH. Both presented with fevers and acute neurologic or respiratory symptoms. One had additional findings of subarachnoid hemorrhage and cytotoxic edema with no retinal hemorrhage identified on ophthalmologic exam. The other subject had hypoxic ischemic injury and diffuse leptomeningeal enhancement and did not have an ophthalmologic exam. Three subjects with HSV had parenchymal or ventricular hemorrhage without SDH, a known complication of HSV. Of the remaining 14 viral meningitis subjects who received head imaging, 6 had nonhemorrhagic findings that were consistent with infection and 8 had normal head imaging.

The evaluation of an infant with SDH requires careful consideration of potential causes, including trauma and infections. This study identifies clear differentiating factors that may aid in initial clinical decision-making. Additionally, in legal settings involving children with AHT, this diagnosis may be questioned. Extracranial injuries (CSIs) are often attributed to a different theorized cause, whereas the intracranial findings are then discussed separately and may be attributed to a proposed mimic of AHT, such as viral meningitis. Although viral meningitis may rarely cause SDH, to assert that viral meningitis is a mimic of AHT, it would need to be shown that presenting features of children with viral meningitis have significant and consistent overlap with those seen in AHT. Our study identified significant differences in HPI and presenting clinical features between AHT and viral meningitis, the opposite of what would be expected if AHT could be frequently and easily confused with viral meningitis. Awareness of these differences may aid clinicians in their evaluation of the young infant with SDH and contribute to increased confidence in the diagnosis both clinically and in legal settings.

Two HPI features were >150 times more likely to occur in the AHT group than in viral meningitis (mental status change and lack of fever). Four other historical features (seizure, decreased tone, apnea, difficulty breathing) were at least 15 times more likely to be present in AHT. Conversely, infectious symptoms predominated in the viral meningitis group, particularly fever. These patterns again held true in the clinical domain, witnessed at the time of medical care. Simply put, the children with viral meningitis presented with common infectious symptoms such as fever, cough, or congestion by history and exam. The children with AHT presented with more severe illness (mental status changes, respiratory failure requiring intubation, seizures). These presenting features, taken as a group, provide clear differentiation between viral meningitis and AHT. Although leukocytosis is often associated with infection, acute stress responses related to head trauma can result in leukocytosis, as well.³⁰  Our results show leukocytosis was more likely in AHT and SDH-only subjects than viral meningitis subjects, thus the presence of leukocytosis should not be used as an indicator to rule out trauma. Demographic differences in race/ethnicity and insurance status may be explained by poverty and the historical effects of racism on poverty, given the known correlations between child maltreatment and poverty.^31–33 

Our data suggest that an overlap in presenting features of AHT with viral meningitis would be a very rare occurrence, typically with “mild” AHT accompanied by infectious signs/symptoms, or infrequently viral meningitis with more severe features. A careful evaluation may identify discriminating features such as lack of fever or concerning skin injury in cases with mild/nonspecific symptoms. In rare circumstances, evaluations for both AHT and viral meningitis may be required. However, given the frequent and significant differences in presentation between viral meningitis and AHT, differentiation between the two conditions should be readily apparent in nearly all circumstances.

Only 19 viral meningitis subjects had head imaging; however, most findings were consistent with infection and differed from those seen in AHT. This is similar to existing literature, which primarily describes intracranial hemorrhage in children with HSV with diffuse hemorrhage,^10,21,22  and rare reports of subdural effusions found in children with possible aseptic meningitis.¹⁹  It has been noted that retinal hemorrhage may be seen associated with coagulopathy in the setting of HSV, but to our knowledge has not otherwise been reported in viral meningitis.⁸  No retinal hemorrhages were identified in the 4 HSV subjects that received ophthalmologic exams. In our study, SDH confined to the posterior cranium was present in 2 viral meningitis subjects aged <10 days. Posteriorly located SDH, most resolving by 4 weeks of age, is well described in asymptomatic neonates from birth and differs from the typical locations of SDH seen in AHT.^34–36  It is likely that the SDH identified in these subjects was birth-related and not a result of viral meningitis.

SDH-only subjects were more similar to AHT than viral meningitis subjects, particularly regarding HPI features. However, SDH-only subjects were generally less acutely ill than AHT subjects. Across the 25 measures comprising the 3 domains studied, the SDH-only group had statistically significant differences from the viral meningitis group in 15 of 25 measures. Mental status change and lack of fever provide clear differentiation with viral meningitis in this group, as well. Statistically significant differences between the SDH-only and AHT groups were found in only 7 of 25 measures. Findings from this study suggest that the SDH-only group is unlikely to represent missed viral meningitis, given the stark differences between the two, but do not support a specific cause of SDH in this population. Given the similarities to our AHT group, it is possible the SDH-only subjects represent children presenting remotely from abusive injury or with less severely symptomatic AHT and lacking CSIs. To identify subjects with isolated SDH, we excluded subjects with CSIs from classification into the SDH-only group. This may select for those children with more “minor” abusive injuries because they lack multisystem trauma. Other possibilities include children who have SDH from a nonabusive mechanism, such as accidental injury or birth. Further study of this specific subpopulation is needed to offer more clarity around the diagnosis.

This study has several limitations. It is retrospective in nature. Evaluation for occult trauma was rare in the viral meningitis group and few SDH-only or AHT patients received lumbar puncture, introducing the possibility that viral meningitis subjects could have unidentified CSIs and SDH-only or AHT subjects could have unidentified viral meningitis. However, given the relatively low prevalence of both conditions, it would be very rare for a subject to have both AHT and viral meningitis at the same time, making misclassification unlikely. Skin findings that could be concerning for or confused with abuse were identified in 1 viral meningitis patient; however, it is possible that there could have been skin findings not documented on the initial exam or that evolved over the patient’s hospitalization. However, skin findings evolving while hospitalized would be less concerning for abuse and more consistent with a viral exanthem. Because viral meningitis subjects rarely had complete evaluation for occult trauma, it is possible that there could be missed occult injury in this group. However, because viral meningitis does not cause extracranial CSIs, an occult injury would require the consideration of abuse in addition to viral meningitis. AHT and SDH-only subjects were derived from hospital CAP team consultations; however, viral meningitis patients did not receive a CAP consultation. To allow for a more equal comparison among groups, features documented in the CAP note were not used for classification of subjects or for comparison across domains. This could have resulted in erroneous classification of AHT subjects as SDH only if an exam finding consistent with a CSI was noted by a CAP but not by the emergency department or inpatient provider, which would result in our AHT and SDH-only groups appearing more similar. Use of CAP team consultations to identify SDH-only subjects allowed for inclusion of those that had evaluation for occult trauma. However, this may have led to selectively capturing SDH-only subjects with more severe symptoms because it is possible that children with SDH only and minor or no symptoms may not have received a CAP consult.

In summary, the hypothesis that viral meningitis mimics AHT is not supported by this study. Subjects with AHT differed significantly from those with viral meningitis with respect to HPI and clinical presenting features, supporting a difference between how these medical conditions manifest. This is the opposite of what would be expected if viral meningitis is a mimic of AHT.

We thank Sommer Rose and Ashley Sherman for data analysis.

AHT

abusive head trauma

CAP

child abuse pediatrics

CSI

concomitant suspicious injury

HPI

history of present illness

HSV

herpes simplex virus

SDH

subdural hemorrhage

WBC

white blood count