Вы находитесь на странице: 1из 9

Advanced Emergency Nursing Journal

Vol. 40, No. 2, pp. 78–86


Copyright 
C 2018 Wolters Kluwer Health, Inc. All rights reserved.

R E S E A R C H T O

Practice
Column Editor: Dian Dowling Evans, PhD, FNP-BC, ENP-BC, FAANP
Downloaded from https://journals.lww.com/aenjournal by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCywCX1AWnYQp/IlQrHD38k+bDukTnmYBRphaf6yoiyNY068aJMc97SyFKWqve7/fjBekaEFpsQ== on 06/02/2018

Acute Headache in the Emergency


Department
Is Lumbar Puncture Still Necessary to Rule
Out Subarachnoid Hemorrhage?
Sarah Steffens, PA-C
Paula Tucker, DNP, FNP-C, ENP-C
Dian Dowling Evans, PhD, FNP-BC, ENP-BC, FAANP

Abstract
The purpose of the Research to Practice column is to review current primary journal articles that
directly affect the practice of the advanced practice nurse (APN) in the emergency department.
This review examines the findings of Carpenter et al. (2016) from their article, “Spontaneous
Subarachnoid Hemorrhage: A Systematic Review and Meta-Analysis Describing the Diagnostic
Accuracy of History, Physical Exam, Imaging, and Lumbar Puncture With an Exploration of Test
Thresholds.” The authors concluded that although no history or physical examination finding can
be used to rule in or rule out spontaneous subarachnoid hemorrhage (SAH), the complaint of neck
stiffness can increase the likelihood of SAH. In addition, the authors concluded that noncontrast
head computed tomography (CT) is accurate in ruling out/in SAH when performed within 6 hr of
symptom onset in adults with symptoms consistent with SAH and that the traditional gold standard
of confirmatory lumbar puncture after a negative head CT scan is only helpful in patients with
a very high pretest probability of SAH. By applying the evidence-based criteria presented in this
study, the emergency department APN can confidently rule out SAH and reduce patient risks
from unnecessary invasive and costly testing. Key words: computed tomography head, headache,
lumbar puncture, subarachnoid hemorrhage

Author Affiliations: Emory University Hospital, At-


lanta, Georgia (Ms Steffens); and Nell Hodgson
Woodruff School of Nursing, Emory University, Atlanta,
Georgia (Drs Tucker and Evans).
Disclosure: The authors report no conflicts of interest.
M RS. K., a 40-year-old African Ameri-
can woman presented to the emer-
gency department (ED) accompa-
nied by her husband with the worst headache
of her life. She awoke this morning feeling
Corresponding Author: Sarah Steffens, PA-C, Emory
University Hospital, 1364 Clifton Rd NE, Atlanta, well and drove herself to work. While sit-
GA 30322 (sarah.steffens@emoryhealthcare.org ting at her desk working on her computer
DOI: 10.1097/TME.0000000000000191 at 9:00 a.m., she developed a severe and

78

Copyright © 2018 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
April–June 2018 r Vol. 40, No. 2 Is Lumbar Puncture Still Necessary to Rule Out SAH? 79

sudden-onset headache. She did not have a II–XII, normal cerebellar function, normal
history of headaches. She admitted that she sensation, and normal gait. A National Insti-
had been under increased stress at work be- tutes of Health Stroke Scale was administered
cause she was just promoted to a management and the score was zero. All other physical ex-
position. Because her headache pain did not amination findings were within normal limits.
reduce with ibuprofen, she became anxious A comprehensive metabolic panel, com-
and called her husband to drive her to the plete blood cell count, and urine pregnancy
ED after her coworker told her she might were ordered by the nurse in triage and
be having a “bleeding aneurysm.” She de- showed no acute electrolyte abnormalities,
scribed the headache as sharp and diffuse, but no anemia or leukocytosis, and no pregnancy.
slightly worse on the right side. The headache The patient anxiously asked if she could be
went from mild to severe over the course of having a bleeding aneurysm or a stroke and
20 min. She denied a preheadache aura, vi- to order a scan of her brain. It was now
sion changes, tingling, numbness, weakness, 11:00 a.m., 2 hr after symptom onset, and
dizziness, neck pain, or stiffness. She denied she was still in severe pain. You decided to
any recent upper respiratory tract symptoms initially treat her pain with acetaminophen
or fever. Her past medical history included and 1-L bolus of intravenous normal saline as
hypertension and hypothyroidism for which you considered the next steps. Your differ-
she took hydrochlorothiazide 12.5 mg once a ential diagnoses were broad, but given the
day and levothyroxine 25 mcg once a day. Her sudden onset of the patient’s symptoms, and
past surgical history included a cesarean de- her description of the “worst headache of
livery 12 years ago and appendectomy about her life,” you were concerned that she could
30 years ago. Her social history included be having a subarachnoid hemorrhage (SAH).
drinking one glass of wine with dinner once Should you expose her to radiation with com-
per week; she denied tobacco or illicit drug puted tomography (CT) of the head? Should
use. She lived with her husband and two chil- you jump straight to CT angiography to as-
dren and had worked in the local university sess for an aneurysm? Should you perform a
admissions office for the past 6 years. lumbar puncture (LP) to look for xanthochro-
On examination, the patient appeared in mia? She had no focal neurological deficits,
moderate distress holding her sweater over and you wanted to avoid unnecessary tests,
her face to block the light in the examina- but you also did not want to miss a poten-
tion room. Vitals signs included temperature tially life-threatening emergency.
37.1 ◦ C, heart rate 99 beats/min, blood pres-
sure 160/90 mmHg, and respiration rate 19
RESEARCH ARTICLE
breaths/min. Her body mass index was 26,
and she appeared well nourished. She was Carpenter, C. R., Hussain, A. M., Ward, M.
alert and oriented to person, place, time, and J., Zipfel, G. J., Fowler, S., Pines, J. M., &
situation. Her skin was warm, dry, with no Sivilotti, M. L. (2016). Spontaneous sub-
rash. Pupils were equal, round, reactive to arachnoid hemorrhage: A systematic review
light, with normal conjunctiva. Her mucous and meta-analysis describing the diagnostic
membranes were dry, she had no sinus tender- accuracy of history, physical, examination,
ness, tonsillar swelling, or erythema, and her imaging, and lumbar puncture with an
tympanic membranes were pearly gray with- exploration of test thresholds. Academic
out cerumen impaction. Musculoskeletal ex- Emergency Medicine, 23(9), 963–1003.
amination revealed no neck tenderness, with
full range of motion without meningeal signs.
PURPOSE/METHODS
Her muscle strength was 5/5 in upper and
lower extremities bilaterally. Her neurologi- The primary purpose of Carpenter et al.’s
cal examination revealed intact cranial nerves (2016) study was to review and summarize

Copyright © 2018 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
80 Advanced Emergency Nursing Journal

the current literature on the diagnostic findings (Whiting et al., 2011), which was
accuracy of clinical history and physical the purpose of Carpenter et al.’s systematic
examination characteristics, CT, cere- review and meta-analysis. QUADAS-2 analysis
brospinal fluid (CSF) analysis, and prediction is performed by two independent researchers
rules to determine the most sensitive and who apply specific, agreed-upon questions to
accurate characteristics/methods associated systematically analyze each study to rate the
with a positive/negative diagnosis of sponta- likelihood of bias in patient selection, index
neous SAH. The secondary purpose was to testing, and interpretation (i.e., CSF analysis
determine disease probability thresholds and or CT results), application of a reference stan-
likelihood ratios of those clinical characteris- dard (criteria established to have 100% diag-
tics associated with SAH to establish optimal nostic sensitivity), and flow and timing (eval-
diagnostic strategies for initiating further uates whether the index test and reference
imaging or LP to improve the accuracy standard were determined for all patients in
of ruling in/out SAH with consideration the same manner; Whiting et al., 2011).
of risk–benefit factors, such as radiation A flow diagram is then constructed to
exposure or procedural complications. Study report studies with low, high, or unclear risk
methods included a systematic review and of bias. Carpenter et al. (2016) performed
meta-analysis of all published literature in prevalence-adjusted, bias-adjusted κ values
PubMed, EMBASE, and Scopus databases to determine interrater reliability in the
using the following search terms: headache, QUADAS-2 analysis. The authors then used
SAH, physical examination, and diagnostic the MOOSE criteria for reporting their meta-
accuracy. Inclusion criteria were studies de- analysis results. MOOSE criteria are applied
scribing ED patients, 14 years or older, seen in the analysis of observational research
for acute headache or other symptoms such studies to further reduce misinterpretation
as syncope, mental status change, or unex- of data due to bias errors and study het-
plained nausea concerning for SAH. Studies erogeneity (Stroup et al., 2000). Carpenter
were also only included if they reported et al. (2016) performed meta-analysis esti-
sufficient detail for comparative analysis. mates if any history, physical examination, or
The authors used the Preferred Report- index testing variables were reported in more
ing Items for Systematic Reviews and Meta- than one study. These estimates were then
Analyses (PRISMA) and the Meta-analysis of used to determine the diagnostic accuracy of
Observational Studies (MOOSE) criteria to de- clinical characteristics or diagnostic results
velop their research design (Carpenter et al., and likelihood ratios of SAH probability.
2016). PRISMA is a recommended and spe- Finally, Pauker–Kassirer statistical analysis
cific technique that uses a 27-item check- was conducted to estimate test–treatment
list for selection, comparison, and synthesis thresholds for cranial CT stratified by time
of study findings to standardize and improve from headache onset and LP based on CSF red
analysis of data in systematic reviews (Moher, blood cells (RBCs) or visible xanthochromia.
Liberati, Tetzlaff, Altman, & PRISMA Group, The Pauker–Kassirer test provides a value
2009). PRISMA can be applied to randomized that factors in the probability of having a
controlled trials and other types of research disease (SAH), the benefit of treatment if
studies. Because PRISMA is not a quality as- having the disease, risk of treatment if no
sessment tool (Moher et al., 2009), Carpenter disease, risk associated with a diagnostic
et al. (2016) used the Quality Assessment test, and probabilities of a negative/positive
of Diagnostic Accuracy Studies-2 (QUADAS- result in a patient with/without disease to
2) tool to assess risk of bias and study qual- establish a test–treatment threshold (Pauker
ity within the systematic review. QUADAS-2 & Kassirer, 1980). In the present analysis,
is recommended when conducting a system- the test was CT or LP and the treatment was
atic review to evaluate diagnostic accuracy angiography.

Copyright © 2018 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
April–June 2018 r Vol. 40, No. 2 Is Lumbar Puncture Still Necessary to Rule Out SAH? 81

The authors defined “disease positive” and physical examination findings for SAH.
as acute SAH based on one of the three These studies compared descriptions such as
following criteria: subarachnoid blood on “worst headache of life” as well as associated
noncontrast head CT scan, CSF xanthochro- symptoms and examination findings such as
mia, or CSF erythrocyte count above 5 photophobia, vomiting, altered mental status,
× 106 /L in the final tube of CSF (Carpenter neck stiffness, and time of onset to deter-
et al., 2016). Patients with a negative head mine which findings would increase or de-
CT scan were only deemed “disease nega- crease the probability of SAH (Carpenter et al.,
tive” if a subsequent LP was also negative or if 2016). Discrepancies in the definition of these
no patients were identified as having SAH at specific findings among the studies led to sta-
the 6-month telephone follow-up (Carpenter tistical heterogeneity, and no single history
et al., 2016). Next, the authors collected data or physical examination finding was found to
by generating 2 × 2 diagnostic accuracy ta- have a high pooled positive or negative like-
bles. Estimates for diagnostic accuracy were lihood ratio. Although no specific historical
calculated using random-effects models. The finding could rule in or rule out SAH, the com-
authors chose to report outcomes as posi- plaint of neck stiffness was the most closely as-
tive likelihood ratios (LR+) and negative like- sociated with an increased probability of SAH
lihood ratios (LR−) because these values are (LR+ 4.12; 95% CI [2.24, 7.59]) (Carpenter
independent of disease prevalence and can et al., 2016). Similarly, headaches of a grad-
be used to estimate posttest probabilities for ual onset (evolved over 1 hr from onset) (LR–
individual patients (Gallagher, 1998). To mea- 0.06; 95% CI [0, 0.95]) and the absence of
sure the consistency of results from each “worst headache of life” (LR– 0.36; 95% CI
study included in the meta-analysis, the au- [0.01, 14.22]) were the complaints found to
thors reported the index of inconsistency (I2 ) most reduce the probability of SAH. The phys-
(Carpenter et al., 2016). High I2 values indi- ical examination finding most associated with
cate that the variability among studies in a SAH was objective neck stiffness (LR+ 6.59;
meta-analysis is due to heterogeneity rather 95% CI [3.95, 11]) (Carpenter et al., 2016).
than chance (Higgins, Thompson, Deeks, & There were an insufficient number of stud-
Altman, 2003). The authors then used pooled ies examining SAH clinical decision rules for
estimates from the meta-analysis to determine the authors to perform a meta-analysis on the
test–treatment thresholds based on the diag- diagnostic accuracy of such tools. The au-
nostic accuracy of the test, as well as the risks thors noted, however, that of the four tools
and benefits of testing and treatment. reviewed, the Ottawa SAH Rule (Perry et al.,
2013) was the most accurate in ruling out
SAH (LR– 0.02; 95% CI [0, 0.39]). The au-
RESULTS
thors also reviewed the diagnostic accuracy
Of the 5,022 studies identified by the authors of noncontrast CT head. Using pooled data,
in their initial literature search, only 22 stud- Backes, Rinkel, Kemperman, Linn, and Ver-
ies met inclusion criteria and were selected gouwen (2012) and Perry et al. (2011) found
for the systematic review and meta-analysis. CT head to be 100% sensitive (95% CI [98,
The overall prevalence of SAH within the sam- 100]; I 2 = 0%) and 100% specific (95% CI
ple reviewed ranged from 0.91% to 68%, with [99, 100]; I 2 = 80.5%) for diagnosis of SAH
a weighted average of 7.5%. QUADAS-2 anal- if performed within the first 6 hr of symptom
ysis found that although study quality, index onset (Carpenter et al., 2016). The sensitivity
testing, and reference standards varied, the of CT accuracy decreased to 89% (95% CI [83,
overall risk of biases was low and interrater re- 93]; I 2 = 89%) and LR– 0.07 (95% CI [0.01,
liability (κ values) was acceptable (Carpenter 0.61]; I 2 = 63%) when it was obtained after
et al., 2016). Eight of the studies investigated 6 hr of symptom onset (Carpenter et al.,
the diagnostic accuracy of specific history 2016). Six of the studies also evaluated the

Copyright © 2018 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
82 Advanced Emergency Nursing Journal

diagnostic accuracy of LP for SAH. Visible limitation of this systematic review is that the
xanthochromia was found to be only 85% results may not be generalizable to commu-
sensitive (95% CI [66, 96]; I 2 = 0%) and 97% nity hospitals with limited resources (Backes
specific (95% CI [96, 98]; I 2 = 13%) for di- et al., 2012; Perry et al., 2011). Both Perry
agnosis of SAH. The authors report that CSF et al. (2011) and Backes et al. (2012) used
analysis was less accurate for diagnosis than a third-generation CT scanners and neuroradiol-
CT scan obtained less than 6 hr from symptom ogists to interpret most scans. The increased
onset (Carpenter et al., 2016). accuracy of imaging from third-generation CT
Finally, the authors examined the test– scanners may limit the conclusions from this
treatment threshold for obtaining diagnostic study from being generalized to EDs where
angiography for confirming SAH by estimat- there may be less advanced scanners. This is
ing the pretest probability ranges for which unlikely to pose a limitation in North America,
the benefit of first testing using either CT or as most hospitals have used third-generation
LP outweighed the risks (Carpenter et al., CT scanners since the 1980s (Edlow & Wyer,
2016). Given the relatively low risks of CT, 2000); however, this may pose a limitation
the authors suggest obtaining a CT scan even to hospitals outside of North America. A re-
if the pre-CT probability of SAH is as low cent study found that general radiologists in
as 0.7% (Carpenter et al., 2016). The data nonacademic hospitals were just as accurate
suggest obtaining an LP after a negative CT at diagnosing SAH as neuroradiologists (Blok
scan if the pre-CT probability of SAH was et al., 2015). These findings argue that al-
greater than 20% (Carpenter et al., 2016). though the use of third-generation scanners
The authors suggest proceeding directly to may pose a limitation, the availability of neu-
angiography if the pre-CT probability of SAH roradiologists may not necessarily limit the re-
was greater than 10% (Carpenter et al., 2016). sults from Carpenter et al.’s (2016) systematic
However, if a negative CT scan is obtained review from being generalized to community
within 6 hr of symptom onset, then the data hospitals.
suggest only proceeding to angiography if The researchers concluded their findings
the pre-CT probability of SAH was nearly 70% support that noncontrast head CT is accurate
(Carpenter et al., 2016). Therefore, the data in ruling out/in SAH when performed within
suggest that the utility of LP after a negative 6 hr of symptom onset in adults with symp-
CT scan is limited and most high-risk patients toms consistent with SAH (Carpenter et al.,
should proceed directly to angiography. 2016). Additional confirmatory LP after CT
or proceeding directly to angiography is only
helpful beyond the 6-hr window or in patients
STUDY LIMITATIONS, STRENGTHS,
with a very high pretest probability of SAH.
AND CONCLUSIONS
Although some signs and symptoms increase
A major strength of this study is that it is the the likelihood of SAH, such as neck stiffness,
first systematic review to assess diagnostic ac- there is no single history or physical exami-
curacy of patient history, physical examina- nation finding that can be used to rule in or
tion findings, and imaging for evaluation of rule out SAH (Carpenter et al., 2016). Clini-
SAH in patients presenting to the ED with cal decision tools for SAH have the potential
acute-onset headache. The study was care- to help ED clinicians identify which patients
fully designed according to recommended would benefit most from further testing. The
methodologies for reducing bias and error authors recommended further research to val-
of interpretation. Despite the limited number idate the Ottawa SAH Rule, as their analy-
and heterogeneity among study variables, er- sis did not find adequate sample size to per-
ror rates due to bias were generally low and form a meta-analysis to determine predictive
study conclusions were well supported. One accuracy.

Copyright © 2018 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
April–June 2018 r Vol. 40, No. 2 Is Lumbar Puncture Still Necessary to Rule Out SAH? 83

SARAH’S COMMENTS 2014). There is no consensus in the literature


on the number of RBCs in the CSF required
Headache accounts for more than 2% of all ED to diagnose SAH (Gorchynski et al., 2007).
visits (Edlow et al., 2008); yet, this systematic There is also variability among hospitals in
review found that only 7.5% of patients with the method to assess xanthochromia. Nearly
a concerning acute headache were eventually all North American hospitals use visual in-
diagnosed with SAH (Carpenter et al., 2016). spection with the naked eye (Carpenter et al.,
Although rare, SAH is deadly, with a mortality 2016). Guidelines in the United Kingdom rec-
rate of about 25% in the first 24 hr and a ommend centrifuging the CSF and perform-
50% overall mortality rate (Long, Koyfman, ing a spectrophotometric scan on the super-
& Runyon, 2017). Thoughtful consideration natant between 350 and 600 nm (Cruickshank
of appropriate tests and procedures to rule et al., 2008). Subarachnoid hemorrhage is di-
out this life-threatening emergency is of agnosed if the net bilirubin absorbance is
utmost importance. In an age of modern 0.007 or less absorbance units and the net
CT scanners with improved diagnostic oxyhemoglobin absorbance is 0.1 or more
sensitivity, an increased emphasis on patient absorbance units (Cruickshank et al., 2008).
comfort and satisfaction, overcrowded EDs Another limitation of xanthochromia, how-
in which clinicians are pressed for time, and ever, is that it can potentially delay diagno-
consideration of Choosing Wisely principles, sis as it can take up to 12 hours postictus to
which encourages use of tests, treatments, develop (Carpenter et al., 2016; Gorchynski
and procedures that are evidenced-based et al., 2007). Diagnosing SAH based on LP
and truly necessary (Choosing Wisely, 2018), analysis is also prone to false-positives due to
can reduce patient risks from unnecessary hyperbilirubinemia or due to traumatic tap,
invasive and costly testing. which occurs in 10%–30% of LPs (Carpenter
The traditional gold standard to assess for et al., 2016; Gorchynski et al., 2007). A re-
SAH has been LP to determine the pres- cent prospective cohort study, however, does
ence or absence of xanthochromia, the yel- suggest that it is possible to distinguish be-
low discoloration caused by the breakdown of tween SAH and traumatic tap if there is no
hemoglobin in CSF, or to measure the amount xanthochromia and RBC count is less than
of RBCs in the final tube of CSF (Gorchynski, 2000 × 106 /L (Perry et al., 2015). In addi-
Oman, & Newton, 2007; Tintinalli, 2015). tion to the potential inaccuracy of LP, there
This gold standard was developed because are patient risks associated with LP. Risk of
early CT scanners were not sensitive enough post-LP headache can be as high as 4%–11%,
to assess for small subarachnoid bleeds; depending on the type of needle used (Nath
therefore, LP was required for confirmation et al., 2017).
(Edlow, 2003). Recent studies, including this This systematic review found that no sin-
meta-analysis, have sought to question and po- gle history or physical examination finding
tentially change this gold standard for a mul- effectively rules in or rules out SAH. Patient
titude of reasons. First and foremost, modern complaints such as “thunderclap headache”
CT scanners have been shown to be far supe- or “worst headache of life,” which were tradi-
rior to older CT scanners due to better reso- tionally thought of as red flags for SAH (Bassi,
lution and thinner slice protocols (Dubosh, Bandera, Loiero, Tognoni, & Mangoni, 1991),
Bellolio, Rabinstein, & Edlow, 2016; Perry were not found to be reliable indicators of
et al., 2011). Although LP may be helpful in SAH (Carpenter et al., 2016). Likewise, the
ruling out other etiologies of headache such absence of those complaints does not pro-
as meningitis, studies suggest that CSF xan- vide assurance that a headache is not an SAH
thochromia is not as accurate in diagnosing (Carpenter et al., 2016). These findings high-
SAH as once thought (Carpenter et al., 2016; light the importance of maintaining a broad
Chu, Hann, Greenslade, Williams, & Brown, differential diagnosis for acute headaches

Copyright © 2018 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
84 Advanced Emergency Nursing Journal

despite a patient’s history and physical exam- further external validation studies of the Ot-
ination findings. tawa SAH rule would be most helpful.
Because the diagnosis of SAH cannot be One of the most striking findings of this
made by history and physical examination systematic review is that CT performed with
alone, the ED provider must decide which, a third-generation scanner within 6 hr of
if any, diagnostic tests to order. Testing can headache onset is nearly 100% sensitive and
be time-consuming and can involve risks to 100% specific for diagnosis of SAH (Carpenter
the patient such as radiation exposure with et al., 2016). Such bold claims are rare in
CT scan or the possibility of discomfort with medical literature; therefore, some may ques-
LP. Emergency providers therefore frequently tion the validity of these findings. The authors
utilize clinical decision tools to identify high- of this systematic review pooled data from
risk patients and limit unnecessary tests. Clin- Backes et al.’s (2012) and Perry et al.’s (2013)
ical decision tools such as the Canadian CT studies to determine the diagnostic accuracy
Head Rule for head injury (Stiell et al., 2001) of CT. These researchers (Backes et al., 2012;
and the PERC Rule for pulmonary embolism Perry et al., 2013) were extremely diligent
(Kline, Mitchell, Kabrhel, Richman, & Court- about follow-up, and it does not appear that
ney, 2004) have been prospectively (Kline any of the patients with negative head CT
et al., 2004; Stiell et al., 2001) and exter- scan who were discharged home suffered
nally (Smits et al., 2005) validated and are from an aneurysmal SAH. Backes et al. (2012)
designed to help the clinician risk-stratify pa- were also meticulous to avoid missed diag-
tients who fit specific inclusion criteria. Al- noses and confirmed findings by performing
though there were an insufficient number of an LP on every patient with a negative head
SAH clinical decision tool studies to perform CT scan. Of the 69 patients with negative
a meta-analysis in Carpenter et al.’s study, the head CT scan within 6 hr of symptom onset,
authors conclude that the Ottawa SAH Rule only one patient had a positive LP finding
was the most accurate (Perry et al., 2013). Af- and was eventually diagnosed with a cervical
ter the publication of Carpenter et al.’ system- arteriovenous malformation. This patient,
atic review, a multicenter prospective cohort however, had endorsed only neck pain and
study was conducted to validate the Ottawa did not endorse headache (Backes et al.,
SAH Rule in patients with acute headache 2012). Therefore, the authors’ (Carpenter
in the ED (Perry et al., 2017). In this recent et al., 2016) recommendation that a negative
study, ED physicians at six tertiary care hos- head CT scan within 6 hr of headache onset is
pitals in Canada enrolled 1,153 neurologically sufficient to rule out SAH without subsequent
intact patients with acute headache concern- confirmatory LP. However, patients with
ing for SAH, 67 of whom were found to have atypical presentations (without headache)
SAH (5.8%) (Perry et al., 2017). The authors may require further workup with LP. Con-
report that the Ottawa SAH Rule had 100% firmation of results with the traditional gold
sensitivity (95% CI [94.6, 100%]) and a speci- standard of LP and diligent follow-up supports
ficity of 13.6% (95% CI [13.1, 15.8]) (Perry Carpenter et al.’s conclusions that a negative
et al., 2017). Perry et al. (2017) concluded CT scan within 6 hr of headache onset effec-
that the Ottawa SAH can be used to rule out tively rules out SAH (Carpenter et al., 2016).
SAH in neurologically intact patients with a A key finding from Carpenter et al.’s study
rapidly peaking headache. This study further is that the diagnostic accuracy of head CT is
supports the use of the Ottawa SAH Rule; most sensitive within the 6-hr window follow-
however, the same hospitals used in the study ing symptom onset. The 6-hr mark is impor-
to derive the rule were also used to validate tant because the sensitivity of CT to detect
the rule (Perry et al., 2017). Of the many po- subarachnoid blood decreases with time, as
tential future research opportunities based on the bleed becomes progressively diluted by
Carpenter et al.’s (2016) systematic review, CSF flow (Dubosh et al., 2016). For patients

Copyright © 2018 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
April–June 2018 r Vol. 40, No. 2 Is Lumbar Puncture Still Necessary to Rule Out SAH? 85

who cannot remember the time of symptom confusion, or fever. By applying evidence-
onset, it is presumably safest to proceed with based criteria, you can confidently rule out
LP as confirmation (Carpenter et al., 2016). SAH and avoid subjecting Mrs. K. to unnec-
Emergency providers must remain vigilant essary and potentially painful tests, while
for diagnosis of possible SAH in any patient reducing her costs of care and arriving at a
presenting with an acute-onset headache that diagnosis in a timely manner.
is unusual and sudden, although this type of
REFERENCES
headache may be of a migrainous etiology.
Understanding the test thresholds for diagnos- Backes, D., Rinkel, G. J., Kemperman, H., Linn, F.
tic testing for SAH and staying up to date on H., & Vergouwen, M. D. (2012). Time-dependent
current evidence are essential for the ED clin- test characteristics of head computed tomogra-
phy in patients suspected of nontraumatic sub-
ician to provide the best care in a high-stakes arachnoid hemorrhage. Stroke, 43(8), 2115–2119.
environment. The results from this system- doi:10.1161/STROKEAHA.112.658880
atic review and meta-analysis are justifications Bassi, P., Bandera, R., Loiero, M., Tognoni, G., & Mangoni,
for practice change in the evaluation of acute A. (1991). Warning signs in subarachnoid hemor-
headache in the ED. rhage: A cooperative study. Acta Neurologica Scan-
dinavica, 84(4), 277–281.
Blok, K. M., Rinkel, G. J., Majoie, C. B., Hendrikse,
J., Braaksma, M., Tijssen, C. C., . . . Vergouwen, M.
CONCLUSION
D. (2015). CT within 6 hours of headache onset
You return to Mrs. K. and explain that you to rule out subarachnoid hemorrhage in nonaca-
demic hospitals. Neurology, 84(19), 1927–1932.
would like to order a noncontrast head CT
doi:10.1212/WNL.0000000000001562
to assess for potential bleeding in the brain. Carpenter, C. R., Hussain, A. M., Ward, M. J., Zipfel, G. J.,
Through shared decision-making, you discuss Fowler, S., Pines, J. M., & Sivilotti, M. L. (2016). Spon-
the risks and benefits of a CT scan and cite taneous subarachnoid hemorrhage: A systematic re-
a recent meta-analysis that demonstrated that view and meta-analysis describing the diagnostic ac-
curacy of history, physical examination, imaging, and
CT scans performed within 6 hr of symptoms
lumbar puncture with an exploration of test thresh-
in patients with acute-onset severe headaches olds. Academic Emergency Medicine, 23(9), 963–
were accurate in ruling out/in subarachnoid 1003. doi:10.1111/acem.12984
bleeds. You explain that if the CT scan is neg- Choosing Wisely. (2018). American College of Emer-
ative, then likely no further testing will be gency Physicians ten things physicians and pa-
tients should question. Retrieved from http://www.
necessary. The patient voices understanding
choosingwisely.org/societies/american-college-of-
of the risks and benefits and agrees with the emergency-physicians/
plan for CT. A noncontrast cranial CT scan Chu, K., Hann, A., Greenslade, J., Williams, J., & Brown,
is obtained within 3 hours of Mrs. K.’s symp- A. (2014). Spectrophotometry or visual inspection
tom onset and is read by the neuroradiologist to most reliably detect xanthochromia in subarach-
noid hemorrhage: Systematic review. Annals of
as “negative” for acute intracranial abnormal-
Emergency Medicine, 64(3), 256.e255–264.e255.
ity. Mrs. K. is relieved when you share the doi:10.1016/j.annemergmed.2014.01.023
news and she explains that her headache is Cruickshank, A., Auld, P., Beetham, R., Burrows, G.,
starting to reduce. Over the next hour, you Egner, W., Holbrook, I., . . . UK NEQAS Specialist
continue to manage her pain with a migraine Advisory Group for External Quality Assurance of
CSF Proteins and Biochemistry. (2008). Revised na-
cocktail of intravenous prochlorperazine 10
tional guidelines for analysis of cerebrospinal fluid for
mg, diphenhydramine 25 mg, and dexam- bilirubin in suspected subarachnoid haemorrhage.
ethasone 10 mg and her symptoms resolve Annals of Clinical Biochemistry, 45(Pt. 3), 238–244.
(Orr et al., 2016). You provide her with in- doi:10.1258/acb.2008.007257
structions to follow-up with her primary care Dubosh, N. M., Bellolio, M. F., Rabinstein, A. A.,
& Edlow, J. A. (2016). Sensitivity of early
provider and to return to the ED if symptoms
brain computed tomography to exclude aneurys-
worsen or if she develops any neurological mal subarachnoid hemorrhage: A systematic re-
deficits such as dizziness, vision changes, tin- view and meta-analysis. Stroke, 47(3), 750–755.
gling/numbness, weakness, slurred speech, doi:10.1161/STROKEAHA.115.011386

Copyright © 2018 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
86 Advanced Emergency Nursing Journal

Edlow, J. A. (2003). Diagnosis of subarachnoid hem- Pauker, S. G., & Kassirer, J. P. (1980). The thresh-
orrhage in the emergency department. Emer- old approach to clinical decision making. The New
gency Medicine Clinics of North America, 21(1), England Journal of Medicine, 302(20), 1109–1117.
73–87. doi:10.1056/NEJM198005153022003
Edlow, J. A., Panagos, P. D., Godwin, S. A., Thomas, T. L., Perry, J. J., Alyahya, B., Sivilotti, M. L., Bullard, M. J.,
& Decker, W. W., & American College of Emergency Emond, M., Sutherland, J., . . . Stiell, I. G. (2015). Dif-
Physicians. (2008). Clinical policy: Critical issues in ferentiation between traumatic tap and aneurysmal
the evaluation and management of adult patients pre- subarachnoid hemorrhage: Prospective cohort study.
senting to the emergency department with acute BMJ, 350, h568. doi:10.1136/bmj.h568
headache. Annals of Emergency Medicine, 52(4), Perry, J. J., Sivilotti, M. L. A., Sutherland, J., Hohl, C.
407–436. doi:10.1016/j.annemergmed.2008.07.001 M., Emond, M., Calder, L. A., . . . Stiell, I. G. (2017).
Edlow, J. A., & Wyer, P. C. (2000). Evidence-based emer- Validation of the Ottawa Subarachnoid Hemorrhage
gency medicine/clinical question. How good is a neg- Rule in patients with acute headache. CMAJ: Cana-
ative cranial computed tomographic scan result in dian Medical Association Journal, 189(45), E1379–
excluding subarachnoid hemorrhage? Annals of E1385. doi:10.1503/cmaj.170072
Emergency Medicine, 36(5), 507–516. Perry, J. J., Stiell, I. G., Sivilotti, M. L., Bullard, M. J.,
Gallagher, E. J. (1998). Clinical utility of likelihood ratios. Emond, M., Symington, C., . . . Wells, G. A. (2011).
Annals of Emergency Medicine, 31(3), 391–397. Sensitivity of computed tomography performed
Gorchynski, J., Oman, J., & Newton, T. (2007). Interpre- within six hours of onset of headache for diagno-
tation of traumatic lumbar punctures in the setting sis of subarachnoid haemorrhage: Prospective cohort
of possible subarachnoid hemorrhage: Who can be study. BMJ, 343, d4277. doi:10.1136/bmj.d4277
safely discharged? The California Journal of Emer- Perry, J. J., Stiell, I. G., Sivilotti, M. L., Bullard, M. J.,
gency Medicine, 8(1), 3–7. Hohl, C. M., Sutherland, J., . . . Wells, G. A. (2013).
Higgins, J. P., Thompson, S. G., Deeks, J. J., & Altman, D. Clinical decision rules to rule out subarachnoid hem-
G. (2003). Measuring inconsistency in meta-analyses. orrhage for acute headache. JAMA, 310(12), 1248–
BMJ, 327(7414), 557–560. doi:10.1136/bmj.327 1255. doi:10.1001/jama.2013.278018
.7414.557 Smits, M., Dippel, D. W., de Haan, G. G., Dekker, H.
Kline, J. A., Mitchell, A. M., Kabrhel, C., Richman, P. B., M., Vos, P. E., Kool, D. R., . . . Hunink, M. G. (2005).
& Courtney, D. M. (2004). Clinical criteria to prevent External validation of the Canadian CT Head Rule and
unnecessary diagnostic testing in emergency depart- the New Orleans Criteria for CT scanning in patients
ment patients with suspected pulmonary embolism. with minor head injury. JAMA, 294(12), 1519–1525.
Journal of Thrombosis and Haemostasis, 2(8), doi:10.1001/jama.294.12.1519
1247–1255. doi:10.1111/j.1538-7836.2004.00790.x Stiell, I. G., Wells, G. A., Vandemheen, K., Clement, C.,
Long, B., Koyfman, A., & Runyon, M. S. (2017). Subarach- Lesiuk, H., Laupacis, A., . . . Worthington, J. (2001).
noid hemorrhage: Updates in diagnosis and manage- The Canadian CT Head Rule for patients with minor
ment. Emergency Medicine Clinics of North Amer- head injury. The Lancet, 357(9266), 1391–1396.
ica, 35(4), 803–824. doi:10.1016/j.emc.2017.07.001 Stroup, D. F., Berlin, J. A., Morton, S. C., Olkin, I.,
Moher, D., Liberati, A., Tetzlaff, J., & Altman, D. Williamson, G. D., Rennie, D., . . . Thacker, S. B.
G., & PRISMA Group (2009). Preferred reporting (2000). Meta-analysis of observational studies in epi-
items for systematic reviews and meta-analyses: The demiology: A proposal for reporting. Meta-analysis
PRISMA statement. PLoS Medicine, 6(7), e1000097. Of Observational Studies in Epidemiology (MOOSE)
doi:10.1371/journal.pmed.1000097 group. JAMA, 283(15), 2008–2012.
Nath, S., Koziarz, A., Badhiwala, J. H., Alhazzani, W., Tintinalli, J. E., Stapczynski, J. S., Ma, O. J., Yealy,
Jaeschke, R., Sharma, S., . . . Almenawer, S. A. (2017). D., Meckler, G. D., Cline, D. M., & Kelen, G.
Atraumatic versus conventional lumbar puncture D. (2015). Emergency medicine: A comprehensive
needles: A systematic review and meta-analysis. The study guide (8th ed.). New York, NY: McGraw-Hill
Lancet. doi:10.1016/S0140-6736(17)32451-0 Professional.
Orr, S. L., Friedman, B. W., Christie, S., Minen, M. T., Bam- Whiting, P. F., Rutjes, A. W., Westwood, M. E., Mal-
ford, C., Kelley, N. E., & Tepper, D. (2016). Manage- lett, S., Deeks, J. J., & Reitsma, J. B., QUADAS-
ment of adults with acute migraine in the emergency 2 Group (2011). QUADAS-2: A revised tool for
department: The American Headache Society evi- the quality assessment of diagnostic accuracy stud-
dence assessment of parenteral pharmacotherapies. ies. Annals of Internal Medicine, 155(8), 529–536.
Headache, 56(6), 911–940. doi:10.1111/head.12835 doi:10.7326/0003-4819-155-8-201110180-00009

Copyright © 2018 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.