Criteria abstracted from The Users' Guide to Medical Literature, from the Health Information Research Unit and Clinical Epidemiology and Biostatistics, McMaster University

Highlighted lines and questions below provide links to the pertinent description of criteria in The EBM User's Guide, now available at the Canadian Centres for Health Evidence

Article Reviewed:

Please visit the new Evidence Based Journal Club Reviews

Clinical Indicators of Intracranial Injury in Head-injured Infants

Greenes DS, Schutzman SA.

Pediatrics 1999; 104: 861-867. [Abstract] [full-text for subscribers, AAP members]

Reviewed by Pooja Tandon, MD, and Scot Bateman, MD, Boston Medical Center, Boston MA

Review posted August 25, 2001

I. What is being studied?

Study objective:

1. To determine whether clinical signs of brain injury are sensitive indicators of intracranial injury (ICI).
2. To determine whether radiographic imaging in otherwise asymptomatic infants with scalp hematoma is a useful means of detecting cases with ICI.
3. To determine whether head-injured infants without signs of brain injury or scalp hematoma may be managed safely without radiographic imaging.

Study design

One year prospective observational study

The patients included:

All infants < 2 years of age who presented to the Emergency Department of a tertiary care pediatric hospital with a complaint or diagnosis of head trauma. 608 patients were enrolled: Ages: 0-2 months: 92 (15%), 3-11 months: 224 (37%), > 12 months: 292 (48%).

Outcome measures:

1. Results of head CT or skull radiography as dictated by an attending radiologist
2. Need for intervention in the ED or subsequent hospital admission
3. Clinical deterioration, complications, or therapy as determined by telephone follow-up at 2 weeks.

II. Are the results of the study valid?

Note: These questions follow from Randolph AG et al. Understanding articles describing clinical prediction tools. Crit Care Med 1998;26:1603-1612. [abstract]

1. Was a representative group of patients completely followed up?

Yes. For all study patients, any interventions in the ED or during subsequent hospitalization was determined from review of medical record. For all patients who were admitted, clinical deterioration and discharge condition were noted. The authors report a 98% follow up with study subjects by phone call 2 weeks after the initial ED visit.

2. Were all potential predictors included?

Mostly. The study included the following potential predictors: mechanism of injury, loss of consciousness, lethargy, irritability, vomiting, seizures, depressed mental status, focal neurological deficits, bulging fontanelle, altered vital signs, age, and presence of scalp hematoma. Of note, mental status was rated by treating physicians by choosing one of four categories created by the authors ranging from "normal" to "severely depressed," rather than by using an established measure of mental status. Other potentially relevant physical exam findings such as scalp lacerations, signs of basilar skull fracture, and papilledema were not included among the potential predictors.

3. Did the investigators test the independent contribution of each predictor variable?

Yes. Chi-square testing was performed and odds ratios were calculated for univariate analysis of the relationship between the various clinical predictor variables and the outcome variables. This study does not consider the impact that many of these factors may have together or how the factors may interact with each other. A multivariate model for prediction of ICI was not used. However, they do analyze the data by controlling for those subjects who were symptomatic vs. asymptomatic and those who underwent head CT and those who did not.

4. Were outcome variables clearly and objectively defined?

Yes. The outcome variables were defined most clearly by the presence of intracranial injury on radiographic imaging, as defined by intracranial hematoma, cerebral contusion or brain swelling on CT. The investigators also followed up on need for particular treatment (such as anticonvulsants or surgery), clinical deterioration or any subsequent complications.

Of the 608 patients, 188 (31%) underwent CT, 122 (20%) underwent skull films but no CT and 298 (49%) underwent no imaging studies. Since the imaging studies were done at the discretion of the managing physicians, some cases of ICI could have been missed. Among symptomatic patients, 26% did not undergo imaging and 36% of patients with a significant scalp hematoma did not undergo imaging.

Of all subjects, 30 (5%) were diagnosed with ICI and 63 (10%) were diagnosed with isolated skull fracture with no ICI. 515 patients (85%) had neither ICI or skull fracture.

III. What are the results?

1. What is(are) the prediction tool(s)?

Lack of impact of history or clinical signs/symptoms

When only the patients who underwent CT scanning are considered, most historical aspects, clinical signs and symptoms were shown to be insensitive predictors of ICI in infants. This lack of effect clearly indicates that symptomatic patients presented with a range of symptoms making it difficult to find significant association between particular variables and the presence of ICI. (i.e., MVA was not associated with ICI!). The purpose of this study was not to suggest foregoing a head CT on symptomatic patients as 16/108 (15%) symptomatic patients who underwent a head CT had ICI.

When all patients were analyzed, there was a significantly higher risk of ICI in patients without a history of trauma (suspected abuse), falls > 3 feet, and falling down stairs.

Scalp hematoma

Significant scalp hematomas (defined as any hematoma for infants less than 1 year old. and a large, boggy hematoma for infants over 1 years of age ) were found to be more sensitive indicators of ICI than the signs and symptoms of brain injury as they were noted in 23 of 30 (77%) of subjects with ICI. Of these 23 patients with significant hematoma and ICI, 22 (96%) also had a skull fracture. Of the 30 subjects with ICI, 23 (77%) had skull fractures, corroborating earlier studies that scalp hematomas are highly associated with skull fractures in infants, which in turn are highly associated with ICI.

Among all subjects, significant scalp hematoma had an OR of 4.65 (95% CI: 2.00, 10.79) for association with ICI (p < 0.001). Among those subjects who underwent head CT, the OR was 2.78 (95% CI: 1.15, 6.70; p = 0.02). Among asymptomatic infants, significant scalp hematoma had an even stronger association with an odds ratio of 22.41 (95% CI: 2.90, 172.9) for association with ICI (p < 0.001).

Age under 3 months

Young age was associated with ICI as 13% of patients < 3 mo who presenting with head trauma had ICI. They did not calculate the odds ratio of young age. The authors do not state how many of the < 3 mo olds were symptomatic vs. asymptomatic.

Asymptomatic

It is unclear why some asymptomatic patients received a head CT and others did not. The authors state that the practice of each physician impacted the decision to get a head CT. Therefore, it is unclear how to interpret the data on ICI rate in asymptomatic patients. Nevertheless, 48% of the patients with ICI had no symptoms! In addition, 41% of symptomatic patients did not undergo CT scanning, thus potentially missing additional cases of ICI.

However, 265 subjects (43%) were in the low risk group (asymptomatic and no scalp hematoma) and only one patient (0.4%; 95% CI: 0.1, 1.9%) had ICI but required no intervention (although it may have been higher if they had imaged every patient). The important thing is that no one (95% CI; 0, 1.2%) in this group (head CT or not) had any subsequent clinical deterioration, suggesting that infants with these characteristics could be safely managed without radiographic imaging. The upper limit of the 95% confidence interval suggests that one in one hundred patients could be missed with this strategy.

2. How well does the model categorize patients into different levels of risk?

The paper suggests that the following groups of children under 2 with a history of head trauma should have radiologic screening:

1. Infants who are symptomatic
2. Infants with a significant scalp hematoma
3. Infants < 3 months of age regardless of symptoms or scalp hematoma

If the patient is over 3 months and asymptomatic there a very low likelihood (between zero and 1.2%) of ICI.

3. How confident are you in the estimates of the risk?

The low risk patient (asymptomatic with no scalp hematoma) has between zero and 1.2% chance of clinically significant injury, with 95% confidence. For some providers, a 1% chance of missing an intracranial injury is too high. If that same patient has a significant scalp hematoma, then the chance of an intracranial injury is much higher (OR of ICI is 22.41; 95% CI 2.90, 172.9).

IV. Will the results help me in caring for my patients?

1. Does the tool maintain its prediction power in a new sample of patients?

This prediction tool was only studied in the ED of one tertiary care pediatric hospital. It would be important to validate the prediction power of this tool by studying it on other samples of patients and perhaps in other settings.

2. Are your patients similar to those patients used in deriving and validating the tool(s)?

Yes. As a hospital-based pediatrician, most of the patients I see with head injury present to a tertiary care hospital Emergency Department.

3. Does the tool improve your clinical decisions?

The findings of this study do not contribute to clinical decision-making with regards to symptomatic infants with head injury. This study confirms that significant scalp hematomas predict skull fractures which in turn are associated with ICI. One is still left with the concern that not all ICI's were detected by this study design; relying solely on clinical course may also be misleading, since one patient who was asymptomatic had a large epidural hematoma that required evacuation. It does however provide moderately reassuring data on when to avoid unnecessary imaging in certain low risk infants.

 

Comments

Back to the EB Journal Club Index