Original Article

Emergency room decision-making

for urgent cranial computed tomography:
selection criteria for subsets
of non-trauma patients

Acta Radiologica
2014, Vol. 55(7) 847854
! The Foundation Acta Radiologica
2013
Reprints and permissions:
sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/0284185113506137
acr.sagepub.com

Christer Tung1,2, Arne Lindgren3,4, Roger Siemund1,2 and

Danielle van Westen1,2

Abstract
Background: Computed tomography (CT) of the brain is used extensively in the urgent work-up of patients with a
suspicion of intracranial pathology, but is often normal. Previously proposed selection criteria aim at limiting the ordering
of urgent cranial CT in the non-trauma population, while maintaining high sensitivity for diagnoses demanding immediate
attention.
Purpose: To retrospectively evaluate these selection criteria in a general non-trauma population from a Swedish tertiary
hospital, as well as in a nested subgroup that lacks guidelines at present, namely where the chief complaint was not
headache, symptoms clearly indicating stroke, seizures, or vertigo.
Material and Methods: Medical records of 346 patients (114 in the nested group) who had undergone urgent cranial
CT were reviewed. Selection criteria as proposed by Rothrock (patient age 60 years, presence of new onset focal
neurologic deficit, headache with vomiting, or altered mental status) were used. Acute cerebral infarction, intracranial
hemorrhage, malignancy, infection, cerebral edema, or hydrocephalus were considered significant findings.
Results: The prevalence of significant findings was 10.1%. The Rothrock criteria had a sensitivity of 97.1% (identifying 34
of 35 significant findings) among all 346 patients and 100% (10/10) among the 114 patients in the subgroup and resulted in
a potential scan reduction rate of 22.8% and 11.4%, respectively. In the patient with significant pathology, that was not
selected for CT, focal neurological symptoms were not described as newly onset.
Conclusion: Although 100% sensitivity was not achieved, our results may contribute to the evidence that in the absence
of focal neurologic deficit, headache with vomiting or altered mental status in patients aged <60 years cranial tomography can be refrained from, in the general population as well as in the subgroup defined above. Further research might
validate patient history as a parameter.

Keywords
Emergency, cranial computed tomography, diagnostic study, cerebral, observational
Date received: 12 June 2013; accepted: 1 September 2013

Introduction

Department of Diagnostic Radiology, Lund University, Lund, Sweden

Center for Medical Imaging and Physiology, Skane University Hospital,
Lund, Sweden
3
Department of Clinical Sciences Lund, Neurology, Lund University,
Lund, Sweden
4
Department of Neurology, Skane University Hospital, Lund, Sweden
2

Cranial computed tomography (CT) is a readily obtainable service and is commonly employed in the work-up
of patients with suspected intracranial pathology. With
the geographically widespread and round-the-clock
availability of CT scanners (1), American emergency
departments (EDs) have seen a remarkable increase in
the use of CT. The proportion of patients visiting an
American ED that underwent a CT or magnetic

Corresponding author:
Danielle van Westen, Department of Radiology, Skane University
Hospital, SE 221 85 Lund, Sweden.
Email: danielle.van_westen@med.lu.se

848
resonance imaging (MRI) scan of any kind increased
from 2.4% in 1992 (2) to 11.2% in 2005 (3). Similarly,
the number of urgent cranial CT scans performed in
both in- and outpatients at our hospital, the Skane
University Hospital Lund, has increased by approximately 60% from 2003 (5100 studies) to 2009 (8200
studies), while the number of visits to the ED in this
period increased by approximately 40% (45,000 in 2003
and 65,400 in 2009). In this context, the word urgent
denotes those examinations that the referring physician
thought necessary to be performed within 24 h. The
assumption that the disease spectrum in the population
has been relatively unchanged during this period seems
reasonable, suggesting that more negative examinations
are performed. A recent radiation dose survey from the
German agency for radiation protection stated that CT
examinations account for 7% of all radiological examinations but contribute with 56% of the total population dose of all radiological examinations (4). Thus, an
attempt to prevent an increase of CT examinations with
negative ndings is warranted in order to reduce costs
and limit extraneous radiation exposure. For the wellcharacterized group of patients with minor head
trauma, clinical selection guidelines for utilization of
cranial computed tomography are available (58).
However, urgent cranial CT is commonly ordered in
a large number of non-trauma patients with a heterogeneous collection of subgroups, many of which lack
guidelines pertaining to the use of CT. As a result, clinicians are often uncertain whether urgent cranial CT is
warranted.
Previous selection clinical criteria by which physicians might limit the number of cranial CT scans
performed in the adult non-trauma general ED population, have been proposed by Rothrock et al. (9). In
their material of 806 patients, all 61 (8%) patients with
the a priori dened signicant ndings of acute cerebral
infarct, intracranial hemorrhage, malignancy, infection,
cerebral edema, hydrocephalus, or a combination
thereof, were identied by any of the four clinical:
patient age 60 years, presence of new onset focal
neurologic decit, headache with vomiting or altered
mental status; henceforth these are denoted the
Rothrock criteria (Table 1). Presence of any one or
more items of these Rothrock criteria was found in
72% of the study population so that application of
these criteria would have reduced the number of cranial
CT scans by 28%. The Rothrock criteria were further
validated in a prospective study on 1911 patients where
the signicant nding rate was 21.7%, while the sensitivity of the Rothrock criteria was 98.8% (95% CI
97.299.6%) (10).
Guidelines for subgroups on the general non-trauma
are available for some common conditions, notably
headache, stroke, and seizure (1117). However, the

Acta Radiologica 55(7)

Table 1. Rothrock criteria, original and as modified in the present work. Altered mental status and GCS <14 were substituted
for by the RLS, level 2 (19).
Rotdrock criteria, original
Age 60 years
Focal neurological deficit
Headache with vomiting
Altered mental status
Rothrock criteria as modified in the present study
Age 60 years
New onset focal neurological deficit
Headache with vomiting
RLS 2

utilization of urgent cranial CT for patients presenting

with vertigo has not been dened, in spite of cranial CT
being of limited value in these patients (18). As urgent
cranial CT is ordered in patients with a variety of complaints and symptoms, many of which are small in
number and consequently do not attract extensive
research, selection criteria are of interest to avoid
unnecessary investigations.
In light of the above, the present, retrospective study
was initiated with the following objectives: (i) to examine the general non-trauma patient population
undergoing urgent cranial CT and investigate the performance and potential scan reduction rate of the
Rothrock selection criteria, respectively, for urgent cranial CT in a Swedish tertiary hospital; and (ii) to examine performance and potential scan reduction rate of
the Rothrock criteria in a nested group of patients
whose chief complaint was not headache, seizures, or
vertigo, nor primarily raised suspicion of acute cerebral
infarct or hemorrhage.

Material and Methods

According to the Ethics Committee at Lund University,
the study did not need approval from the Committee.
It was conducted at Skane University Hospital Lund, a
tertiary care center with an annual ED census of
approximately 65,000 visits.
Medical records were retrospectively reviewed for
patients identied through the hospitals radiology
information system (RIS) as having undergone urgent
cranial CT, with or without administration of intravenous contrast agents, between 1 January 2009 and
30 April 2009, in total 805 patients. Thus, the inclusion
criterion was to have undergone urgent cranial CT.
Exclusion criteria were: (i) head trauma during the
last 72 h; (ii) hospitalization, ED admission, visit to a
hospital outpatient clinic, or neuroimaging during the
previous 7 days; (iii) referral for cranial CT occurring

Tung et al.

849

Fig. 1. Age distribution of the general study population (n 360), counts, and years.

>24 h after hospital admission; (iv) previous neuroimaging-conrmed diagnosis of intracranial pathology or
CT for preoperative or post-therapeutic use; (v) incomplete medical record. Subsequently data from 346
patients (114 in the nested group) were available for
analysis; the age distribution is shown in Fig. 1, specifically, 221 patients were aged >60 years. Medical records including CT referrals and reports were reviewed
for clinical information and for information pertaining
to the Rothrock criteria as well as to the CT ndings.
No age limit was imposed in the patient selecting
process.
Patients chief complaints were construed from the
reason for visit, urgent problem, and anamnesis
entries in the medical records and sorted as headache,
suspicion of stroke/TIA, seizures, vertigo, altered
mental status, impaired general condition, syncope,
confusion, or gait disturbance. Sudden onset aphasia,
unilateral weakness or paresthesia, and symptom constellations primarily raising concerns regarding cerebrovascular insult were categorized as suspected
stroke/TIA. Instances in which the provided journal
information did not allow for condent categorization
into any of the above were categorized as miscellaneous. In addition, a group of patients within the total
group, henceforth denoted the nested group was
dened to comprise those patients whose chief complaint was not headache, seizures, vertigo, nor primarily raised suspicion of acute cerebral infarct or
hemorrhage, i.e. those complaints where selection criteria for urgent CT are not available at present.
The Rothrock criteria were adapted to Swedish
standards, specically altered mental status was
expressed using the in Scandinavia commonly used
Reaction Level Scale (RLS) instead of the Glasgow
Coma Scale (GCS) (19). While the GCS score is the
composite of three ordinal scales (eye, verbal, and
motor, with a high score indicating level of consciousness closer to normal), the RLS is a hierarchically
ordered scale with eight categories (reaction levels)

with a lower score indicating level of consciousness

closer to normal with lower inter-observer variability
than the GCS (20,21). The criterium GCS <14 in the
Rothrock criteria thus was replaced by RLS level 2.
In case the RLS was not documented, this was assessed
from joint evaluation of patient history and description
of the patients general condition. Our modied
Rothrock criteria demanded the presence of one or
more of: (i) age 60 years; (ii) new onset focal neurologic decit; (iii) headache with vomiting; (iv) RLS 2.
In this study, acute cerebral infarct, intracranial
hemorrhage, malignancy, infection, cerebral edema,
hydrocephalus, or any combination of the above,
were considered signicant CT ndings. All other ndings, including chronic small vessel disease, reportedly
non-acute cerebral infarction or hemorrhage, atherosclerosis, cerebral atrophy as well as a residual group
of clinically non-signicant intracranial ndings such as
benign tumors and extracranial ndings (e.g. hypertrophic sinonasal mucosa) were dened as non-signicant. Non-signicant ndings were only documented if
they were newly discovered or sucient in progress in
relation to past neuroimaging studies to warrant mention by the interpreting radiologist.
Pathology or conditions that the ED physicians suspected or wanted to rule out, were identied from the
CT referral forms. Entries were categorized as enquiries
for intracerebral hemorrhage, ischemia, mass lesion,
increased intracranial pressure, venous sinus thrombosis, hydrocephalus, fracture, or sinusitis. In cases
where the enquiry could not be sorted to any of the
above, the condition was labeled as specic other if
the query was specic in nature. Unspecic queries
asking for presence of pathology in general were classied as undened. Stroke queries counted
towards both intracerebral hemorrhage and ischemia.
Statistical analysis was performed using SPSS 19.0
for Windows (SPSS Inc., Chicago, IL, USA).
Demographic data (Table 2) are represented as counts
for the total number and as the percentage of the

850

Acta Radiologica 55(7)

Table 2. Demographics and clinical characteristics of the general and nested population.

Age 60 years

General population
(n 346)
n (%)*

Residual populationy
(n 232)
n (%), [95% CI]*

Nested population
(n 114)
n (%), [95% CI]

221 (63.9%)

126 (54.3%)
[47.760.9]z
40 (17.3%)
[12.622.7]
15 (6.5%)
[3.710.4]
76 (32.8%)
[26.839.2]
11 (4.7%)
[2.48.3]
126 (54.3%)
57.9
[55.260.7]
63
25 (10.8%)
[7.115.5]

95 (83.3%)
[75.289.7]
1 (0.9%)
[0.024.8]
1 (0.9%)
[0.024.8]
35 (30.7%)
[22.440.0]
31 (27.2%)
[19.336.3]
59 (51.8%)
71.8
[67.273.5]
76
10 (8.8%)
[4.315.5]

Headache with nausea or vomiting

41 (11.8%)

Headache with vomiting

16 (4.6%)

New onset focal neurologic deficit

111 (32.1%)

RLS 2

42 (12.1%)

Female
Mean age

185 (53.5%)
62.5

Median age
Significant findings

67.5
35 (10.1%)

*The total number is shown as the count, and as the percentage of respective population in brackets, while 95% CIs are shown in square brackets. The
Students t-test was used for calculating CIs for mean age and for comparison between groups.
y
The residual population contains the patients whose chief complaint was headache, vertigo, TIA/stroke, or seizures.
z
Significant differences between the nested and residual populations are in bold format.

Altered mental status and GCS <14 were substituted for by the RLS, level 2 (20).

respective population; the Students t-test was used to

compare groups. The Clopper-Pearson method was
used to compute the 95% CIs for the statistical measures of performance (e.g. sensitivity, specicity, positive
predictive value, and negative predictive value).
Logistic regression was used to determine odds ratios
(ORs) for the Rothrock criteria separately in relation to
a dichotomous variable indicating signicant or nonsignicant CT ndings.

Results
Demographics
As mentioned above, in total 805 patients were identied as having undergone urgent cranial CT after referral from the ED during the period from 1 January to
30 April 2009, using the Department of Radiologys
database. Exclusion criteria eliminated the following
numbers of patients from each category: (i) head
trauma during the last 72 h, n 201; (ii) hospitalization,
ED admission, visit to an outpatient clinic, or neuroimaging during the last 7 days, n 39; (iii) referral for
cranial CT >24 h after admission to hospital, n 105;
(iv) known neuroimaging-conrmed diagnosis, or CT
performed for preoperative or post-therapeutic use,
n 85; (v) incomplete medical record, n 29. Thus,

346 (43%) of 805 were eligible for the study.

Distribution in terms of gender was fairly even with
women comprising 53.5% (185/346) of the patient
population. The mean age was 62.5 years, of which
2.6% (9/346) were aged 15 years and 63.9% (221/
346) were aged 60 years. The nested population,
that is the subgroup of patients whose chief complaint
was not headache, seizures, vertigo, and whose symptoms did not suggest acute cerebral infarction or hemorrhage, comprised 114 of the general population of
346 patients. Demographic characteristics diered
between this nested group and the residual population
with regard to age and symptoms (Table 2).
The most frequent primary complaints were headache (25.1%, 87/346), symptoms suggesting stroke/
TIA as the working diagnosis (22.5%, 78/346), vertigo
(13.9%, 48/346), and seizures 5.5% (19/346). The
remaining subgroups of confusion (10.7%, 37/346),
impaired general condition (4.3%, 15/346), syncope
(2.9%, 10/346), gait disturbance (2.3%, 8/346), and
miscellaneous (12.7%, 44/346) collectively comprised
33.0% (114/346) of the population and together form
the nested population. The unspecic category of
miscellaneous encompassed a heterogeneous group
of patients including those whose symptoms were not
evidently related to the central nervous system, acute
and chronic pain conditions, a history of head trauma

Tung et al.

851

Table 3. Primary complaints and distribution of significant

findings.

Symptoms of stroke/TIA
Headache
Vertigo
Seizures
Impaired general condition
Altered mental status
Syncope
Confusion
Gait disturbance
Miscellaneous

Primary
complaint
(n 346)
n (%)

Significant
CT finding
(n 35)
n (%)

78
87
48
19
15
20
10
17
8
44

14
6
4
1
2
3
2
1
0
2

(22.5%)
(25.1%)
(13.9%)
(5.5%)
(4.3%)
(5.8%)
(2.9%)
(4.9%)
(2.3%)
(12.7%)

(17.9%)
(14.5%)
(8.3%)
(5.3%)
(13.3%)
(15.0%)
(20.0%)
(5.9%)
(0%)
(4.5%)

older than 72 h, adverse eects of pharmacological

therapy, unspecic paresthesia, among others (Table 3).

Cranial CT findings
Signicant CT ndings were found in 35 patients,
resulting in an overall prevalence rate of 35/346 patients
(10.1%); non-signicant ndings, newly discovered or
noteworthy progressed were found in 102 (29.5%)
patients; normal or unaltered appearance was demonstrated in 209 (60.4%) patients. The signicant ndings
in the general population were: acute cerebral infarct
(n 15), cerebral parenchymal hemorrhage (n 5), subarachnoid hemorrhage (n 2), subdural hemorrhage
(n 4), intracerebral tumor with surrounding edema
or mass eect (n 4), and metastatic malignancy
(n 5). In the nested population signicant ndings
were found in 10 patients (10/114, 8.8 %), divided as
follows: acute cerebral infarct (n 4), cerebral parenchymal hemorrhage (n 1), subdural hemorrhage
(n 3), intracerebral tumor with edema (n 1), and
metastatic malignancy (n 1).
The most prevalent non-signicant ndings were
chronic small vessel disease (n 49/346, 14.2%), nonacute cerebral ischemic infarction (n 32/346, 9.2%),
and hypertrophic sinonasal mucosa (n 22/346,
6.4%). Also, the non-signicant intracranial ndings
included a 5-mm-wide hygroma without mass eect, a
meningioma without mass eect, an arachnoid cyst, a
dermoid cyst, minor (<5 mm) herniation of the cerebellar tonsils, and two cases with calcications of benevolent appearance. Diagnoses most commonly suspected
or asked by the referring physician to be ruled out were
intracerebral hemorrhage (260/346, 75.1%) and ischemia (193/346, 55.8%), followed by expansive lesion

(91/346, 26.3%) and unspecied enquiries for pathology (52/346,15.0%). The mean number of query
terms entered on the radiology referrals was 2.11.

Performance of modified Rothrock on the general

non-trauma population
Our modied Rothrock criteria (presence of any of: age
60 years, new onset focal neurologic decit, RLS 2
or headache with vomiting, or a combination thereof)
selected 267 of 346 (77.2%) patients in our study group.
The remaining 79 patients accounted for a 22.8%
potential scan reduction rate. The criteria identied
34 of 35 signicant ndings resulting in a 97.1%
(95% CI 85.199.9%) sensitivity and a specicity of
25.1% (78/311, 95% CI 20.430.3%). The positive predictive value was 12.7% (95% CI 9.017.3%) and the
negative predictive value 98.7% (95% CI 93.2
99.97%). The patient not identied using these criteria
was a 37-year-old man with malignant melanoma and a
known subcutaneous metastasis who sought medical
attention for 1 weeks duration of dysphasia and new
onset tongue cramps. However, the neurologic exam
was reportedly without any new onset neurological deficits and thus the patient was not selected for CT
according to our criteria. Four metastases with surrounding edema were later detected on cranial CT.
ORs using binary logistic regression for the modied
Rothrock criteria were signicant for new onset focal
neurologic decit (OR 2.83, 95% CI 1.395.74,
P 0.004) and RLS 2 (OR 3.49, 95% CI 1.547.92,
P 0.003) (Table 4).

Performance of modified Rothrock criteria

in the subgroup where the chief complaint
was not headache, stroke/TIA, seizures,
or vertigo
This subgroup that included patients whose primary
complaint was altered mental status, impaired general
condition, syncope, confusion, and gait disturbance,
comprised 114 patients, of which 10 (8.8%) demonstrated signicant ndings (Table 2). Our modied
Rothrock criteria selected 88.6% (101/114) of patients
in the subgroup, resulting in a potential scan reduction
rate of 11.4%. All 10 cases with signicant ndings
were identied, resulting in 100% sensitivity (95% CI
74.1100%). The specicity was 12.5% (13/104, 95%
CI 6.820.4%), the positive predictive value 9.9% (95%
CI (4.917.5%) and the negative predictive value 100%
(95% CI 79.4100%) (Table 4).
Logistic regression showed signicant results for
patients with new onset focal neurologic decit
(P 0.047) and RLS 2 (P 0.023). Cranial CT

852

Acta Radiologica 55(7)

Table 4. Performance of the modified criteria in the general

and nested populations.
Modified Rothrock
criteria % (CI)
General population (n 346)
Sensitivity
Specificity
Positive predictive value
Negative predictive value
False-positive rate
False-negative rate

97.1
25.1
12.7
98.7
74.9
2.9

(85.199.9)
(20.430.3)
(9.017.3)
(93.299.97)
(69.779.6)
(0.114.9)

Nested population (n 114)

Sensitivity
Specificity
Positive predictive value
Negative predictive value
False-positive rate
False-negative rate

100
12.5
9.9
100
87.5
0

(74.1100)
(6.820.4)
(4.917.5)
(79.4100)
(79.693.2)
(025.9)

Numbers are shown as percentages, and 95% CIs (brackets).

Table 5. Odds ratios (ORs) of criteria items in the general and

nested populations.
P value

ORs (95% CI)

Age 60 years

Headache with nausea and vomiting

0.076
0.662

2.11 (0.934.78)
1.25 (0.463.43)

Headache with vomiting

0.29

2.01 (0.557.39)

New onset focal neurologic deficit

0.003

2.88 (1.425.84)

RLS 2

0.004

3.31 (1.467.47)

Age 60 years

Headache with nausea and vomiting

0.998
0.999

Headache with vomiting

0.999

New onset focal neurologic deficit

0.039

4.09 (1.0815.53)

RLS 2

0.028

4.50 (1.1817.15)

General population (n 350)

Nested population (n 118)

scans with signicant ndings were more likely to be

found in patients with new onset focal neurologic deficit (OR 3.88, 95% CI 1.0214.75) or RLS 2 (OR
4.74, 95% CI 1.2418.15). ORs are detailed in Table 5.

Discussion
For nancial and healthcare purposes, i.e. radiation
exposure, a highly sensitive decision rule in addition
to clinical practice guidelines, that selects patients of a
wide range of complaints for radiological investigation,
would be of high value. The aim of our study was to

assess the performance and potential scan reduction

rate of the modied Rothrock selection criteria for
urgent cranial CT in the non-trauma general population in a Swedish tertiary hospital as well as in a novel,
nested patient group for whom selection criteria are not
available at present, namely those patients whose chief
complaint was not headache, seizures or vertigo, nor
primarily raised suspicion of acute stroke. The prevalence of signicant ndings in our material was 10.1%,
to be compared to 8% in the work of Rothrock et al.
(10). The sensitivity of the Rothrock criteria, modied
regarding level of consciousness to be in accordance
with local practice, was 97.1% identifying 34 of 35 signicant ndings in 346 patients and 100% (10/10)
among 114 patients in the nested group, resulting in a
potential scan reduction rate of 22.8% and 11.4%,
respectively. It should be noted though that our retrospective study design where only patients that had
undergone a cranial CT were included, may induce a
selection bias and may inuence the reported statistics.
In this study we tested slightly modied Rothrock
criteria against the general non-trauma population, and
more importantly a nested group. The Rothrock criteria identied all signicant ndings in the nested
group and demonstrated, in accordance with previous
studies, high sensitivity and a moderate potential scan
reduction rate, in our general population. Even so, the
potential scan reduction rate in the nested group was
lower suggesting the possibility of further adaptation of
the Rothrock criteria for this group.
In their pioneering work comprising 806 patients,
Rothrock et al. derived selection criteria from a prospective observational study. Four criteria: age >60
years, altered mental status, focal weakness, and headache with vomiting combined in a decision algorithm
requiring one of these, had a sensitivity of 100% and a
negative predictive value of 100% for detecting signicant CT scan abnormalities. These results were supported by the study by Tan et al. who evaluated the
Rothrock criteria prospectively in 1911 patients attending the ED of Westmead Hospital in Sydney between
2002 and 2005 (11). In their study, the signicant nding rate was 21.7%, while the sensitivity of the
Rothrock criteria was 98.8% (95% CI 97.299.6%).
Our modied Rothrock criteria have a comparable
high sensitivity of 97.1% (95% CI 85.199.9%) in the
general population. In addition, all patients with signicant ndings in our nested group were detected
using the modied Rothrock criteria. Although promising, this result must be viewed in the context of the
small material size with signicant ndings being present in 10 out of 114 patients and of the resulting condence interval (95% CI 74.1100%). A third work in
this eld has been performed by Harris et al., who
modied the Rothrock criteria so that altered mental

Tung et al.
status was replaced by GCS <14, and the criterion
age 60 years was removed (22). In their own material, these modied Harris criteria were retrospectively sucient to detect all 22 (35%) of the 62
patients with signicant CT ndings. Had the original
Rothrock criteria been applied in the Harris study, the
amount of scans would be reduced by 11%. Tan et al.
found reduced sensitivity for the Harris criteria in comparison to the Rothrock criteria suggesting the age criterion is of importance; in our material, this was also
the case, thus we have chosen not to report results from
the Harris criteria.
Urgent cranial CT is an important tool in the
workup of patients with suspected intracranial pathology that attend the ED. However, development of
highly sensitive clinical selection criteria for common
patient groups such as those in the non-trauma general
population with symptoms that may hint of intracranial pathology remains alluring. While it is highly
improbable to attain 100% sensitivity for the aforementioned signicant ndings using reasonable clinical
selection criteria, a very high sensitivity in conjunction
with a substantial scan reduction rate may be of value
for clinicians; the latter being dependent on the local
prevalence rate in patients undergoing cranial CT. An
additional parameter to validate for possible inclusion
in the decision rule would be previous history (as was
the case in the patient not selected by the Rothrock
criteria in our material). Apart from the sensitivity
not being 100%, in a clinical setting where practice
guidelines play an augmented role, simplicity and pragmatism is often depreciated. In the ED, the surging seek
for medical attention prompting fast sorting of patients
where ruling out disease has become increasingly
important.
Some limitations of this study include that 29
patients were excluded solely on the grounds of incomplete medical records. Quality of the medical documentation of neurologic examinations varied and the
conclusion that neurologic signs had been ruled out
may have based on incomplete description of patients
status. The patients chief complaint had to be constructed from three dierent headings in the medical
records. Due to clinicians varying descriptions on
what type of information was recorded under the
reason for visit and urgent problem headings, it
was often necessary to seek additional information provided under the patient history heading; hence interobserver reliability is limited.
As pointed out previously, there may be a selection
bias due to the fact that only patients that did undergo
cranial CT were included, which may inuence the
results. There was no control group in this study,
hence it is unknown how the criteria would have performed on patients suspected of having intracranial

853
pathology but did not undergo an urgent cranial CT.
The criteria, we therefore believe, should not be used to
select patients for CT who normally would not be
referred. Examples of such situations are that gastroenteritis patients with headache and vomiting or
patients aged 60 years whose complaint is of apparent
extracranial cause, are normally not referred for urgent
cranial CT. Rather the criteria may provide justication
for abstaining CT in patients not exhibiting any of the
criteria items nor whom have specic risk factors.
In addition, our analysis will miss false-negatives
caused by patients whose ED visit does not result in a
cranial CT, but who return within 7 days with a signicant nding, by means of the exclusion criteria.
In conclusion, although 100% sensitivity was not
achieved, our results may contribute to the evidence
that in the absence of focal neurologic decit, headache
with vomiting, or altered mental status in patients aged
<60 years cranial CT can be refrained from. Further
research might validate patient history as a parameter.
Funding
This research received no specic grant from any funding
agency in the public, commercial, or not-for-prot sectors.

References
1. Ginde AA, Foianini A, Renner DM, et al. Availability
and quality of computed tomography and magnetic resonance imaging equipment in U.S. emergency departments. Acad Emerg Med 2008;15:780783.
2. Schappert SM. National Hospital Ambulatory Medical
Care Survey: 1992 Emergency Department Summary.
Vital Health Stat 1997;125:1108.
3. Nawar EW, Niska RW, Xu J. National Hospital
Ambulatory Medical Care Survey: 2005 emergency
department summary. Adv Data 2007;386:132.
4. Radiation exposures from medical applications. See
http://www.bfs.de/de/ion/medizin/Medizin.pdf
5. Ingebrigtsen T, Romner B, Kock-Jensen C. Scandinavian
guidelines for initial management of minimal, mild, and
moderate head injuries. J Trauma 2000;48:760766.
6. Jagoda AS, Bazarian JJ, Bruns JJ Jr, et al. Clinical policy:
neuroimaging and decisionmaking in adult mild traumatic brain injury in the acute setting. J Emerg Nurs
2009;35:e5e40.
7. Stiell IG, Wells GA, Vandemheen K, et al. The Canadian
CT Head Rule for patients with minor head injury.
Lancet 2001;357:13911396.
8. Vos PE, Battistin L, Birbamer G, et al. EFNS guideline
on mild traumatic brain injury: report of an EFNS task
force. Eur J Neurol 2002;9:207219.
9. Rothrock SG, Buchanan C, Green SM, et al. Cranial
computed tomography in the emergency evaluation of
adult patients without recent history of head trauma: a
prospective analysis. Acad Emerg Med 1999;4:654661.
10. Tan I, Young N, Sindhusake DP, et al. Prospective evaluation of selected clinical criteria for cranial computed

854

11.

12.

13.

14.

15.

Acta Radiologica 55(7)

tomography in non-trauma adult patients. Emerg Med
Australas 2009;21:4351.
Adams HP Jr, del Zoppo G, Alberts MJ, et al. Guidelines
for the early management of adults with ischemic stroke:
a guideline from the American Heart Association/
American Stroke Association Stroke Council, Clinical
Cardiology Council, Cardiovascular Radiology and
Intervention Council, and the Atherosclerotic Peripheral
Vascular Disease and Quality of Care Outcomes in
Research Interdisciplinary Working Groups: The
American Academy of Neurology affirms the value of
this guideline as an educational tool for neurologists.
Stroke 2007;38:16551711.
Masdeu JC, Irimia P, Asenbaum S, et al. EFNS guideline
on neuroimaging in acute stroke. Report of an EFNS
task force. Eur J Neurol 2006;13:12711283.
Edlow JA, Panagos PD, Godwin SA, et al. Clinical
policy: Critical issues in the evaluation and management
of adult patients presenting to the emergency department
with acute headache. J Emerg Nurs 2009;35:e43e71.
Sandrini G, Friberg L, Coppola G, et al.
Neurophysiological tests and neuroimaging procedures
in non-acute headache. 2nd edn. Eur J Neurol
2011;18:373381.
Harden CL, Huff JS, Schwartz TH, et al. Reassessment:
neuroimaging in the emergency patient presenting with
seizure (an evidence-based review): report of the
Therapeutics and Technology Assessment Subcommittee
of the American Academy of Neurology. Neurology
2007;69:17721780.

16. Krumholz A, Wiebe S, Gronseth G, et al. Practice

Parameter: evaluating an apparent unprovoked first seizure in adults (an evidence-based review): report of the
Quality Standards Subcommittee of the American
Academy of Neurology and the American Epilepsy
Society. Neurology 2007;69:19962007.
17. ACEP Clinical Policies Committee; Clinical Policies
Subcommittee on Seizures. Clinical policy: Critical
issues in the evaluation and management of adult patients
presenting to the emergency department with seizures.
Ann Emerg Med 2004;43:605625.
18. Labuguen RH. Initial evaluation of vertigo. Am Fam
Physician 2006;15:244251.
19. Starmark JE, Stalhammar D, Holmgren E. The Reaction
Level Scale (RLS85). Manual and guidelines. Acta
Neurochir (Wien) 1988;91:1220.
20. Tesseris J, Pantazidis N, Routsi C, et al. A comparative
study of the Reaction Level Scale (RLS85) with
Glasgow Coma Scale (GCS) and Edinburgh-2 Coma
Scale (modified) (E2CS (M)). Acta Neurochir (Wien)
1991;110:6576.
21. Stalhammar D, Starmark JE, Holmgren E, et al.
Assessment of responsiveness in acute cerebral disorders.
A multicentre study on the reaction level scale (RLS 85).
Acta Neurochir (Wien) 1988;90:7380.
22. Harris JE, Draper HL, Rhodes AI, et al. High yield criteria for emergency cranial computed tomography in
adult patients with no history of head injury. J Accid
Emerg Med 2000;17:1517.

Copyright of Acta Radiologica is the property of Sage Publications, Ltd. and its content may
not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's
express written permission. However, users may print, download, or email articles for
individual use.