Low risk for subsequent subarachnoid hemorrhage for emergency department patients with headache, bloody cerebrospinal fluid, and negative findings on cerebrovascular imaging

Clinical article

Full access

Object

When patients present to the emergency department (ED) with acute headache concerning for subarachnoid hemorrhage (SAH) and a lumbar puncture (LP) shows blood in the CSF, it is difficult to distinguish the results of a traumatic LP from those of SAH. CT angiography (CTA) is often performed, but the long-term outcome for patients with a positive LP and normal neurovascular imaging remains uncertain. The primary objective of this study was to determine whether patients who presented to the ED with acute headache and had red blood cells (RBCs) in their CSF on LP but negative cerebrovascular imaging suffered subsequent SAH.

Methods

A case series study of consecutive adult ED patients who incurred charges for LP between 2001 and 2009 was performed from 2010 through 2011. Inclusion criteria were: headache, > 5 RBCs/mm3 in CSF, noncontrast head CT with no evidence of hemorrhage, and cerebrovascular CTA or MRA without aneurysm or vascular lesion within 2 weeks of the ED visit. Patients with less than 6 months of available follow-up were excluded. The primary outcomes were 1) subsequent nontraumatic SAH and 2) new vascular lesion. Secondary outcomes were complications related to SAH, or LP or angiography.

Results

Of 4641 ED patients billed for an LP, 181 patients (mean age 42 years) were included in this study. Over a median follow-up of 53 months, 0 (0%) of 181 patients (95% CI 0%–2.0%) had a subsequent SAH or new vascular lesion identified. Although not the primary outcome, there was 1 patient who was ultimately diagnosed with vasculitis. Eighteen (9.9%) of 181 patients (95% CI 6.0%–15.3%) had an LP-related complication and 0 (0%) of 181 patients (95% CI 0%–2.0%) had an angiography-related complication.

Conclusions

Patients who present to the ED with acute headache concerning for SAH and have a finding of bloody CSF on LP but negative findings on cerebrovascular imaging are at low risk for subsequent SAH and thus are likely to be safe for discharge. Replacement of the CT/LP with a CT/CTA diagnostic algorithm merits further investigation.

Abbreviations used in this paper:CTA = CT angiography; DSA = digital subtraction angiography; ED = emergency department; ICH = intracranial hemorrhage; IQR = interquartile range; IRB = institutional review board; LMR = longitudinal medical record; LP = lumbar puncture; MRA = MR angiography; RBC = red blood cell; SAH = subarachnoid hemorrhage; WBC = white blood cell.

Object

When patients present to the emergency department (ED) with acute headache concerning for subarachnoid hemorrhage (SAH) and a lumbar puncture (LP) shows blood in the CSF, it is difficult to distinguish the results of a traumatic LP from those of SAH. CT angiography (CTA) is often performed, but the long-term outcome for patients with a positive LP and normal neurovascular imaging remains uncertain. The primary objective of this study was to determine whether patients who presented to the ED with acute headache and had red blood cells (RBCs) in their CSF on LP but negative cerebrovascular imaging suffered subsequent SAH.

Methods

A case series study of consecutive adult ED patients who incurred charges for LP between 2001 and 2009 was performed from 2010 through 2011. Inclusion criteria were: headache, > 5 RBCs/mm3 in CSF, noncontrast head CT with no evidence of hemorrhage, and cerebrovascular CTA or MRA without aneurysm or vascular lesion within 2 weeks of the ED visit. Patients with less than 6 months of available follow-up were excluded. The primary outcomes were 1) subsequent nontraumatic SAH and 2) new vascular lesion. Secondary outcomes were complications related to SAH, or LP or angiography.

Results

Of 4641 ED patients billed for an LP, 181 patients (mean age 42 years) were included in this study. Over a median follow-up of 53 months, 0 (0%) of 181 patients (95% CI 0%–2.0%) had a subsequent SAH or new vascular lesion identified. Although not the primary outcome, there was 1 patient who was ultimately diagnosed with vasculitis. Eighteen (9.9%) of 181 patients (95% CI 6.0%–15.3%) had an LP-related complication and 0 (0%) of 181 patients (95% CI 0%–2.0%) had an angiography-related complication.

Conclusions

Patients who present to the ED with acute headache concerning for SAH and have a finding of bloody CSF on LP but negative findings on cerebrovascular imaging are at low risk for subsequent SAH and thus are likely to be safe for discharge. Replacement of the CT/LP with a CT/CTA diagnostic algorithm merits further investigation.

Spontaneous subarachnoid hemorrhage (SAH) can be a devastating event, with an annual incidence of approximately 25,000 cases in the United States. The current standard of care for emergency department (ED) evaluation of SAH suggests that if a lumbar puncture (LP) is positive for red blood cells (RBCs) or xanthochromia after normal findings on noncontrast head CT, then cerebrovascular imaging such as CT angiography (CTA) is recommended to assess for a cerebral aneurysm.4,7,9,10,12,20,23,32 Magnetic resonance angiography (MRA) may also be performed, if available.8,30,31

If patients are found to have an aneurysm on angiography, then SAH is confirmed and definitive management follows. However, if CTA is nondiagnostic, there are a few possibilities. Approximately 15% of patients with SAH have a nonaneurysmal cause of hemorrhage.28 These cases can be divided into 1) those with positive findings on head CT with or without a perimesencephalic blood distribution and 2) those with negative findings on head CT but positive findings on LP.2 The results of CSF evaluation can be true positive due to SAH or falsely positive due to a traumatic LP.

The risk of subsequent SAH in patients with SAH on CT and negative CTA has been described in multiple series.3,5,15,18,19,28,29 Perimesencephalic nonaneurysmal SAH has a good short- and long-term prognosis. Nonperimesencephalic nonaneurysmal SAH has a varied etiology and prognosis, and the prognosis partly depends on the distribution and extent of hemorrhage.2,19,27 However, the risk of subsequent SAH with bloody CSF (presence of xanthochromia or > 5 RBCs/mm3 in CSF from LP) and negative results on modern CT or CTA is not well described.2,18,27

Multidetector CTA is considered to be excellent in the detection of aneurysms1,9,13,34 and has been estimated to have a sensitivity of 99.4% in detecting ruptured aneurysms compared with intraoperative findings.24 CTA can sometimes detect aneurysms missed by CT/LP6 and digital subtraction angiography (DSA).3

We hypothesized that acute headache patients with negative findings on noncontrast head CT and an LP positive for CSF RBCs who have no aneurysm or neurovascular lesion on cerebrovascular imaging are at low risk for subsequent catastrophic SAH. If this is true, then it will confirm the safety of the current practice of discharging these patients home. It also allows further consideration of an alternative diagnostic algorithm of CT/CTA in lieu of CT/LP to assess for acute SAH or risk of imminent SAH in appropriate patients presenting to the ED with headache. Our primary goal in this investigation is to determine whether the combination of CT and CTA of the head is sufficiently sensitive to assess for acute aneurysmal SAH or increased risk of future SAH in patients with bloody CSF.

Methods

Study Design

This was a retrospective case series of all consecutive adult patients who incurred charges for an LP performed in the ED of Massachusetts General Hospital between 2001 and 2009 and had headache, bloody CSF, and negative findings on neurovascular imaging. The study was approved by the hospital's institutional review board (IRB), with exemption from informed consent, and was conducted between 2010 and 2011.

Study Setting and Patient Population

The setting was a single urban academic hospital with over 90,000 annual ED visits, which serves as a tertiary care referral center for neurosurgical emergencies. A list of all ED patients who were older than 16 years of age and incurred a charge for an LP between 2001 and 2009 was generated and reviewed for 3 inclusion criteria: 1) a complaint of headache, 2) > 5 RBCs/mm3 CSF in all LP tubes, and 3) cerebrovascular imaging (CTA or MRA) within 2 weeks of the ED visit.

Exclusion criteria included: depressed consciousness such that the patient was unable to indicate having a headache, ED fever with temperature above 100.5°F, unsuccessful LP, LP performed at another facility, LP primarily performed for cytology evaluation, LP with culture-positive meningitis, neurovascular imaging with evidence of intracranial hemorrhage (ICH), aneurysm, or vascular lesion, presence of a ventriculoperitoneal shunt, head trauma within 2 weeks of the ED visit, neurosurgery within 4 weeks prior to the ED visit, or less than 6 months of clinical follow-up available after the initial ED visit.

Fever was an exclusion criterion because although patients with SAH may have secondary fever as part of their clinical presentation, we wanted to eliminate patients who might have meningitis. Those who had a history concerning for meningitis would clearly need further evaluation or admission and would not be low-risk patients who could safely be discharged.

Imaging

Brain imaging was performed in the ED with 16- or 64-slice CT scanners (GE Medical Systems) and a 1.5-T MR scanner (GE Medical Systems).

Study Protocol

Screening of all computer records generated by a search for billing codes for LP were performed by 4 nonblinded reviewers to determine whether patients met the inclusion criteria. Exclusion criteria were assessed in 2 tiers. Most exclusion criteria were assessed with this initial screening of the computer record. A few exclusion criteria—including presence of ED fever with temperature above 100.5°F; culture-proven meningitis; evidence of ICH, aneurysm, or vascular lesion; and less than 6 months of follow-up—required further detailed review of the medical record, and these cases were assessed upon comprehensive medical record review.

Comprehensive medical record review of all patients who met the initial inclusion and exclusion criteria was performed by 2 unblinded emergency physician reviewers, who abstracted data by consensus using a standardized protocol and data collection instrument. Data collected included: date and time of ED presentation; demographic characteristics; duration, quality, and circumstances of headache, and associated symptoms; medical history (hypertension, chronic headache disorders, history of SAH, cerebrovascular disease or malignancy, polycystic kidney disease, vasculitis, smoking history, warfarin, aspirin, clopidogrel use); relevant family history; physical examination findings of neck stiffness or neurological abnormalities; maximum ED temperature; date and time of LP performance; results of CSF analysis; and all neuroimaging results. If a finding such as photophobia was not specifically mentioned in the record, it was noted as such, but ultimately treated as absent.

Standardized follow-up consisting of 3 questions was performed for all patients without exclusion criteria: 1) the presence of new atraumatic ICH subsequent to the visit, 2) the presence of a new cerebral vascular lesion subsequent to the visit, and 3) any adverse events related to the ED visit including complications of SAH such as new seizures or hydrocephalus and complications post-LP or CTA.

The follow-up process was performed according to the IRB-approved protocol in 2 phases. 1) Patients who presented for any follow-up care to one of over 20 local hospitals/clinics within the “Partners” network, the largest such network in the state, which shares an electronic health information system (longitudinal medical record [LMR]) with the study institution, had systematic computerized medical record review of all visit notes, admission/discharge summaries, neuroimaging studies, and operative reports to determine responses to the above questions. 2) Patients who did not receive any documented follow-up in the electronic Partners LMR were contacted by telephone. First, a letter, along with a self-addressed return card, was sent to the patient signed by the ED vice chairman and administrative director and study investigator, informing the patient about the study objective and requesting a response to accept or decline contact. If the patient approved contact or no response was received within 2 weeks, then up to 3 attempts were made to contact the patient by telephone to obtain answers to the 3 follow-up questions listed above using a standardized telephone script.

If the patient could not be reached by telephone, the following strategies were used sequentially to attempt follow-up: 1) calling the emergency contact number, 2) calling the primary care physician to find a valid patient contact number, 3) performing a Google search with the patient's name (and sometimes also home state in cases of multiple results), and finally 4) searching the US Social Security Death Index.

Because our objective was to determine long-term clinical outcome for patients presenting with headache and abnormal CSF results with normal vascular imaging, patients who did not have follow-up available for at least 6 months after their ED visit were considered to have inadequate follow-up and were excluded. These patients were still followed up to the maximum number of months available to exclude SAH/vascular lesion in that period of less than 6 months, and they were assessed for death in the Social Security Death Index. All patient deaths were investigated to determine if the death may have been related to an acute SAH.

Two blinded emergency physicians each reviewed a different half of all of the charts that were subjected to comprehensive review to assess for reliability of data extraction. They collected data on the following variables: elements related to the headache presentation, physical examination, medical history, and family history, as listed previously, as well as time from symptom onset to LP, CSF results (number of RBCs/mm3 in the first tube, number of RBCs/mm3 in the last tube, xanthochromia), maximum ED temperature, all neuroimaging results (head CT/CTA, neck CTA, head MRI/MRA, neck MRI/MRA, conventional angiogram), and follow-up information available by medical record review (subsequent ICH, subsequent aneurysm/neurovascular lesion).

When there were discrepancies between the blinded and nonblinded review, a third emergency physician reviewed the data to determine the final outcome.

Outcome Measures

The primary outcomes were 1) any subsequent nontraumatic SAH and 2) any new aneurysm, arteriovenous malformation, or other cerebrovascular lesion. We also made note of all cases of subsequent ICH. Secondary outcomes were any reported complications from SAH in addition to rebleeding and procedural complications within 2 weeks of the LP (headache, back pain, infection, neurological symptoms, other) or CTA (allergic reaction, contrast nephropathy, other) requiring medical care.

Data Analysis

The study was powered for a 95% confidence interval span of no more than 2% (0% to 2%) if no primary outcome events were observed. Summary statistics were computed with SPSS-17 (IBM, Inc.). For binomial results, exact confidence intervals were computed with StatXact-3 (Cytel Software). Reliability analyses were computed using unweighted kappa statistics and intraclass correlation coefficients for nominal and interval data, respectively.

Results

Characteristics of Study Subjects

Computer records of 4462 consecutive ED patients older than 16 years of age who had an LP performed between 2001 and 2009 were initially reviewed. Of these 4462 patients, 1193 had no headache. Of the remaining 3269 patients, 2133 did not have bloody CSF on LP. Of the 1136 patients with headache and bloody CSF, 824 received no angiographic examination, and 32 were excluded from the study for other reasons. Two hundred eighty patients with headache, > 5 RBCs/mm3 in the final CSF tube, neurovascular imaging (CTA, MRA), and none of the first-tier exclusion criteria were eligible for comprehensive chart review (Fig. 1). Of these, 181 patients were included who met all inclusion/exclusion criteria and had long-term follow-up (Table 1). Five patients were lost to follow-up and none were identified in the Social Security Death Index. Patients presented at a median time of 2 days after their headache onset. Eighty-one percent of patients had neurovascular imaging within 1 calendar day of the ED visit and LP (median 0 days, interquartile range [IQR] 0–1 day).

Fig. 1.
Fig. 1.

Study flow diagram. *Other exclusion factors: ventriculoperitoneal shunt, trauma within 2 weeks of ED visit, neurosurgery within 4 weeks of ED visit, LP primarily for cytology. §Some patients had more than 1 reason for exclusion. #Some patients had more than 1 type of cerebrovascular imaging.

TABLE 1:

Summary of characteristics of 181 patients*

CharacteristicValue
demographics
 mean age in yrs ± SD42.4 ± 13.9
 female107 (59.1)
historical factors
 sudden or thunderclap headache54 (30.2)
 worst headache of life39 (21.8)
 hours from headache onset
  median48
  IQR8.1–120.0
 exertion20 (11.2)
 syncope9 (5)
 subjective fever23 (12.8)
 neck stiffness57 (31.8)
 photophobia53 (29.6)
 vomiting48 (26.8)
 altered mental status13 (7.3)
 seizure7 (3.9)
 any neurological complaint36 (20.1)
past medical history
 hypertension35 (19.6)
 chronic headache disorder
  migraine24 (13.4)
  other14 (7.8)
 current smoker33 (18.4)
 history of SAH7 (3.9)
 cerebrovascular disease9 (5)
 brain tumor3 (1.7)
 vasculitis/autoimmune disease4 (2.2)
 polycystic kidney disease0 (0)
 family history of aneurysm15 (8.4)
medications
 warfarin2 (1.1)
 aspirin26 (14.5)
 clopidogrel1 (0.6)
physical exam findings
 neck stiffness13 (7.3)
 any neurological abnormality30 (16.8)
LP results
 first tube RBCs/mm3
  median600
  IQR110–2185
 last tube RBCs/mm3 (n = 171)
  median72
  IQR12–663
 final tube WBCs/mm3
  median2
  IQR1–4
 xanthochromia in at least 1 tube13 (7.2)

Values represent numbers of patients (%) unless otherwise indicated.

Analysis is based on a total of 179 patients for historical factors and physical examination variables because 2 paper charts were missing.

Neurological abnormalities: unilateral weakness (13), unilateral sensory loss (4), disorientation (3), aphasia (2), visual deficit (7), unsteady gait (1).

Main Results

Over a median follow-up of 53 months (IQR 27.5–79.0 months), 0 of 181 patients (0%, 95% CI 0%–2.0%) had a primary outcome of subsequent nontraumatic SAH. There were 4 cases of subsequent new intraparenchymal hemorrhage in follow-up, 3 related to trauma. The single nontraumatic case involved a patient who returned 2 weeks after the index ED visit with new hemiparesis and was found to have vasculitis, which was subsequently complicated by petechial intraparenchymal hemorrhage. The other primary outcome of a subsequent new vascular lesion was found in 0 of 181 patients (0%, 95% CI 0%–2.0%).

We did not identify any patients with new seizures or hydrocephalus related to their ED visit. With respect to the secondary procedural outcomes, 18 (9.9%) of 181 patients (95% CI 6.0%–15.3%) had an LP-related complication, including headache (in 15 cases), neck/back pain (in 2), or leg pain (in 1), and 0 (0%) of 181 patients (95% CI 0%–2.0%) had an angiography-related complication such as allergic reaction or contrast nephropathy.

There were 5 patients who were found to be deceased during follow-up. The cause of death in the single patient who died within the 6-month follow-up period was thought to be complications from an ischemic stroke. He presented with mild headache and confusion, and an inpatient MRI demonstrated multiple nonhemorrhagic infarcts after the patient's initial ED presentation. CTA showed a right M2 inferior branch cutoff and no aneurysms. Stroke service evaluation included an LP with 6 RBCs/mm3 and 2 white blood cells (WBCs)/mm3 in the final tube. A transesophageal echocardiogram demonstrated cardiomyopathy with left ventricular thrombus, which was thought to be the source of embolic strokes. The patient's hospital course was complicated by witnessed aspiration and sepsis, and the patient died within 2 months of the initial ED visit. The other 4 deaths were not related to SAH.

Data reliability analysis was performed on all reported variables for all 280 patients, except for the follow-up results, which were collected for only 137 patients with long-term follow-up by computer record. All reliability values were greater than 0.80 except for those for the following variables: findings on conventional angiogram (κ = 0.74), evidence of subsequent ICH on medical record follow-up (κ = 0.39), evidence of subsequent aneurysm/neurovascular lesion on medical record follow-up (κ unable to be calculated because no positive results), and evidence of post-LP complications (κ = 0.71).

Discussion

The results of this study support the common clinical assumption that patients with bloody CSF but negative CTA/MRA generally have a good prognosis, do not require further emergent diagnostic evaluation, and can safely be discharged home, if other emergent workup is not being considered. Although we expected CTA to be more commonly used, the neurovascular imaging in this study is distributed almost evenly between CTA and MRA. Comparable sensitivity between CTA and MRA is assumed.4,16,30,33

Although not the main focus of our study, among over 3000 patients who received an LP for headache after negative findings on head CT over 8 years in our ED, ultimately only 4 patients were diagnosed with SAH based on LP results alone after negative results on noncontrast head CT. Diagnosis was made in conjunction with neurovascular imaging. This illustrates a low yield of the current CT/LP strategy at our institution.

In essence, whether the LP demonstrates bloody CSF, the disposition and prognosis seem to be most strongly determined by the results of the noncontrast head CT and CTA. Thus, it seems logical to consider an alternative diagnostic strategy. For patients with suspected SAH and negative findings on noncontrast CT, such a strategy might be to perform primary CTA in lieu of LP.

Our result is also in concurrence with a recent study that estimated the risk of a missed aneurysmal SAH after a CTA with negative findings to be 1% or less by mathematical modeling.22 A single prospective study of patients with clinical presentations suggestive of SAH has been performed directly comparing CT/CTA with CT/LP in 106 patients.6 Based on strict criteria, 2 patients had positive results of LP, and in both cases aneurysms were revealed by CTA and confirmed by DSA. In 100 patients, the results of both LP and CTA were negative, and it is unknown if any of these patients suffered subsequent bleeds.

There are caveats to the use of the proposed alternative diagnostic strategy. Physicians cannot forgo LP if other conditions are being considered for which LP is crucial for diagnosis, such as meningitis, idiopathic intracranial hypertension, or intracranial hypotension.

If CTA is used after CT in lieu of LP, then nonaneurysmal SAH with negative findings on head CT would not be identified. For patients with nonaneurysmal SAH identified on CT, prognosis seems inversely associated with the volume of hemorrhage.2,19,27 Although not previously proven, considering these data, it seems reasonable that patients with CSF hemorrhage identified only by LP without hemorrhage or vascular lesion visualized by head CT/CTA also have a good long-term prognosis.

CTA interpretation is important for sensitivity and depends on the experience of the radiologist.17 Although not found in this study, the use of CTA may be associated with complications from the intravenously administered contrast agent, such as allergic reaction or contrast agent–induced nephropathy, and those patients with low glomerular filtration rates may require exclusion from a pathway featuring angiography. The risk of ionizing radiation and comparative costs are also factors to consider.

It is estimated that about 2% of the general population harbors cerebral aneurysms.26 When an aneurysm is detected on CTA with the CT/CTA approach, further workup must be performed to determine whether the aneurysm is ruptured, and this may include LP to determine whether there is bloody CSF and/or xanthochromia. If the aneurysm is unruptured, then an incidental finding may lead to additional evaluation with attendant risk and expense. However, there may be some benefit to identifying the aneurysm if it were deemed to be a symptomatic lesion, because in these cases, there is an 8-fold risk of rupture compared with asymptomatic incidental aneurysms.26

Despite caveats, the concept of alternative diagnostic strategies merits consideration. Implementing this approach would be anticipated to lead to more CTAs and fewer LPs among patients with headache who present to EDs. Avoiding LPs may mean less pain, less evaluation time for the patient and physician, less horizontal time (patients may receive ongoing evaluation and care without requiring a stretcher), improved flow in the ED, fewer neurosurgical consultations, and fewer cases of post-LP headache or other rare but serious LP complications.

Recent evidence even suggests that within 6 hours of headache onset, noncontrast head CT alone may be used to rule out SAH.25 Ultimately, clinical judgment will also need to be used to determine whether a patient is safe for discharge or further testing is warranted.

This retrospective study suggests that a subsequent catastrophic SAH is unlikely in patients who present to an ED with headache and have normal findings on head CT/CTA. Thus, there may be sufficient safety and clinical equipoise to merit a prospective trial comparing a CT/LP versus a CT/CTA diagnostic approach.

If a CT/CTA diagnostic approach proves efficacious, future studies are needed to better delineate the patient population in which such a strategy might be employed. We would not view it as necessarily a better diagnostic strategy, but rather an apparently safe alternative approach, useful in certain situations. Such an approach might be particularly useful for patients who refuse LP, for patients with a critical requirement for anticoagulation (such as those with left-sided mechanical valves), or in those with spinal disease in whom an LP might be technically difficult. Clinical criteria for ruling out meningitis and identifying patients in whom LP can be omitted without missing infection represent an important area for further clarification. Finally, the optimal management of incidentally found aneurysms and whether such an aneurysm is symptomatic requires further study.

Limitations

Some patients included in this study may not seem to be at high risk for SAH. The included patients had a median of 600 RBCs/mm3 (IQR 110–2185 RBCs/mm3) and 72 RBCs/mm3 (IQR 12–663 RBCs/mm3) in the first and final LP tubes, respectively. Based on previous work that included this data set, a result of < 100 RBCs/mm3 in the final tube decreases the odds of true SAH by a point estimate likelihood ratio or factor of 0 (95% CI 0–0.3).11 In this study, 140 (77%) of 181 and 81 (45%) of 181 first and final tubes, respectively, had > 100 RBCs/mm3. The median absolute difference divided by the mean RBC counts of the first and final tubes was 1.45 (IQR 0.61–1.86). A value greater than 0.63 decreases the likelihood of true SAH.11 Thirteen patients had xanthochromia. Xanthochromia may not develop for 12 hours after headache onset, and this test alone may be insufficiently sensitive for SAH.21 Thirty patients (17%) demonstrated neurological abnormalities (Table 1). Overall, the 181-patient group appeared to be at moderate risk for SAH. While patients with headache and bloody LP warranting neurovascular imaging represent the ED group at highest risk for SAH despite negative findings on head CT, this underscores the difficulty in identifying a truly high-risk group in the modern era of sensitive multidetector head CT imaging.

Conversely, 824 patients with > 5 RBCs/mm3 in the final tube received no neurovascular imaging and were thus excluded from further analysis. This raises the possibility of selection bias and missed aneurysmal hemorrhage. Analyses of LP data were obtained on a random sample of 100 of these excluded patients. The first and final tube median RBC counts were 311 cells/mm3 (IQR 74–924 cells/mm3) and 20 cells (IQR 10–94 cells/mm3), respectively. The median absolute difference divided by the mean of the values for the first and final tubes was 1.67 (IQR 0.80–1.89). In 31 of the 100 samples selected for further analysis, the additional analysis showed an elevated WBC count consistent with meningitis. Overall, these LP characteristics suggest the patients were at low risk for aneurysmal SAH and neurovascular imaging was appropriately deferred.

There are several other limitations of our study related to its retrospective nature. It is possible an LP may have been performed but no LP billing code was generated in a small number of cases, and the patients involved would not have been identified by our search. For some of the variables characterizing the patient's history and nature of headache, we are limited by the paucity of information reported in some medical records; however this does not affect our objective criteria of LP and CTA or MRA findings. While CTA and MRA have been shown to have high sensitivity and specificity for detecting aneurysms, this was not specifically tested in our study.

Both primary and secondary outcomes could have been underestimated if patients presented to facilities outside of the study institution's network of hospitals and the event was not reported in the Partners medical record. Furthermore, patients may not be able to remember adverse events from LP/CTA with such delayed follow-up leading to recall bias.

Although our reliability analysis revealed good agreement between reviewers on almost all variables, for the variable of subsequent ICH on medical record follow-up, κ was especially low (0.39) because of 3 cases of disagreement. All reviewers agreed on the 135 cases negative for ICH and on 1 case positive for ICH. However, there were 2 additional cases of ICH found by the blinded reviewer (the vasculitis case mentioned above and 1 trauma-related case), and 1 traumatic ICH case found by the nonblinded reviewer, which was missed by the blinded reviewer. The primary outcome, nontraumatic SAH, was not identified by any reviewer. It is known that the kappa statistic can be inordinately decreased when the marginal values are imbalanced.14 The overall agreement proportion was 98% (136 of 139 cases).

Conclusions

Patients who present to the emergency department with a headache and have a normal head CT but bloody CSF on LP are at low risk for subsequent SAH if they also have normal neurovascular imaging. These patients are likely safe for discharge unless they need further evaluation for other conditions. Further prospective study is warranted to determine whether cerebrovascular imaging, such as CTA alone after CT, may be a safe and efficacious alternative approach to rule out suspected aneurysmal SAH at risk for subsequent bleeding.

Acknowledgments

We would like to thank John T. Nagurney, M.D., and Clif Callaway, M.D., Ph.D., for their support and assistance with this study.

Disclosure

Dr. Marill reports an ownership interest in General Electric.

The project described was supported by Award Number 5K12HL109068 from the National Heart, Lung, and Blood Institute (K.A.M.). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Heart, Lung, and Blood Institute or the National Institutes of Health.

Author contributions to the study and manuscript preparation include the following. Conception and design: Thomas, Czuczman, Marill. Acquisition of data: all authors. Analysis and interpretation of data: Thomas, Czuczman, Marill. Drafting the article: Thomas, Marill. Critically revising the article: all authors. Reviewed submitted version of manuscript: all authors. Statistical analysis: Marill. Administrative/technical/material support: Marill. Study supervision: Marill.

This article contains some figures that are displayed in color online but in black-and-white in the print edition.

A summary of this work was previously presented at the annual meeting of the Society for Academic Emergency Medicine in Boston, Massachusetts, June 1–5, 2011.

References

  • 1

    Agid RLee SKWillinsky RAFarb RIterBrugge KG: Acute subarachnoid hemorrhage: using 64-slice multidetector CT angiography to “triage” patients' treatment. Neuroradiology 48:7877942006

    • Search Google Scholar
    • Export Citation
  • 2

    Alexander MSDias PSUttley D: Spontaneous subarachnoid hemorrhage and negative cerebral panangiography. Review of 140 cases. J Neurosurg 64:5375421986

    • Search Google Scholar
    • Export Citation
  • 3

    Andaluz NZuccarello M: Yield of further diagnostic workup of cryptogenic subarachnoid hemorrhage based on bleeding patterns on computed tomographic scans. Neurosurgery 62:104010472008

    • Search Google Scholar
    • Export Citation
  • 4

    Bederson JBConnolly ES JrBatjer HHDacey RGDion JEDiringer MN: Guidelines for the management of aneurysmal subarachnoid hemorrhage: a statement for healthcare professionals from a special writing group of the Stroke Council, American Heart Association. Stroke 40:99410252009

    • Search Google Scholar
    • Export Citation
  • 5

    Canhão PFerro JMPinto ANMelo TPCampos JG: Perimesencephalic and nonperimesencephalic subarachnoid haemorrhages with negative angiograms. Acta Neurochir (Wien) 132:14191995

    • Search Google Scholar
    • Export Citation
  • 6

    Carstairs SDTanen DADuncan TDNordling OBWanebo JEPaluska TR: Computed tomographic angiography for the evaluation of aneurysmal subarachnoid hemorrhage. Acad Emerg Med 13:4864922006

    • Search Google Scholar
    • Export Citation
  • 7

    Chappell ETMoure FCGood MC: Comparison of computed tomographic angiography with digital subtraction angiography in the diagnosis of cerebral aneurysms: a meta-analysis. Neurosurgery 52:6246312003

    • Search Google Scholar
    • Export Citation
  • 8

    Chen YCSun ZKLi MHLi YDWang WTan HQ: The clinical value of MRA at 3.0 T for the diagnosis and therapeutic planning of patients with subarachnoid haemorrhage. Eur Radiol 22:140414122012

    • Search Google Scholar
    • Export Citation
  • 9

    Chen WWang JXin WPeng YXu Q: Accuracy of 16-row multislice computed tomographic angiography for assessment of small cerebral aneurysms. Neurosurgery 62:1131222008

    • Search Google Scholar
    • Export Citation
  • 10

    Connolly ES JrRabinstein AACarhuapoma JRDerdeyn CPDion JHigashida RT: Guidelines for the management of aneurysmal subarachnoid hemorrhage: a guideline for healthcare professionals from the American Heart Association/American Stroke Association. Stroke 43:171117372012

    • Search Google Scholar
    • Export Citation
  • 11

    Czuczman ADThomas LEBoulanger ABPeak DASenecal ELBrown DF: Interpreting red blood cells in lumbar puncture: distinguishing true subarachnoid hemorrhage from traumatic tap. Acad Emerg Med 20:2472562013

    • Search Google Scholar
    • Export Citation
  • 12

    Edlow JAPanagos PDGodwin SAThomas TLDecker WW: Clinical policy: critical issues in the evaluation and management of adult patients presenting to the emergency department with acute headache. Ann Emerg Med 52:4074362008

    • Search Google Scholar
    • Export Citation
  • 13

    El Khaldi MPernter PFerro FAlfieri ADecaminada NNaibo L: Detection of cerebral aneurysms in nontraumatic subarachnoid haemorrhage: role of multislice CT angiography in 130 consecutive patients. Radiol Med (Torino) 112:1231372007

    • Search Google Scholar
    • Export Citation
  • 14

    Feinstein ARCicchetti DV: High agreement but low kappa: I. The problems of two paradoxes. J Clin Epidemiol 43:5435491990

  • 15

    Flaherty MLHaverbusch MKissela BKleindorfer DSchneider ASekar P: Perimesencephalic subarachnoid hemorrhage: incidence, risk factors, and outcome. J Stroke Cerebrovasc Dis 14:2672712005

    • Search Google Scholar
    • Export Citation
  • 16

    Gibbs GFHuston J IIIBernstein MARiederer SJBrown RD Jr: 3.0-Tesla MR angiography of intracranial aneurysms: comparison of time-of-flight and contrast-enhanced techniques. J Magn Reson Imaging 21:971022005

    • Search Google Scholar
    • Export Citation
  • 17

    Hochberg ARRojas RThomas AJReddy ASBhadelia RA: Accuracy of on-call resident interpretation of CT angiography for intracranial aneurysm in subarachnoid hemorrhage. AJR Am J Roentgenol 197:143614412011

    • Search Google Scholar
    • Export Citation
  • 18

    Jung JYKim YBLee JWHuh SKLee KC: Spontaneous subarachnoid haemorrhage with negative initial angiography: a review of 143 cases. J Clin Neurosci 13:101110172006

    • Search Google Scholar
    • Export Citation
  • 19

    Kang DHPark JLee SHPark SHKim YSHamm IS: Does non-perimesencephalic type non-aneurysmal subarachnoid hemorrhage have a benign prognosis?. J Clin Neurosci 16:9049082009

    • Search Google Scholar
    • Export Citation
  • 20

    Kokkinis CVlychou MZavras GMHadjigeorgiou GMPapadimitriou AFezoulidis IV: The role of 3D-computed tomography angiography (3D-CTA) in investigation of spontaneous subarachnoid haemorrhage: comparison with digital subtraction angiography (DSA) and surgical findings. Br J Neurosurg 22:71782008

    • Search Google Scholar
    • Export Citation
  • 21

    MacDonald AMendelow AD: Xanthochromia revisited: a re-evaluation of lumbar puncture and CT scanning in the diagnosis of subarachnoid haemorrhage. J Neurol Neurosurg Psychiatry 51:3423441988

    • Search Google Scholar
    • Export Citation
  • 22

    McCormack RFHutson A: Can computed tomography angiography of the brain replace lumbar puncture in the evaluation of acute-onset headache after a negative noncontrast cranial computed tomography scan?. Acad Emerg Med 17:4444512010

    • Search Google Scholar
    • Export Citation
  • 23

    Menke JLarsen JKallenberg K: Diagnosing cerebral aneurysms by computed tomographic angiography: meta-analysis. Ann Neurol 69:6466542011

    • Search Google Scholar
    • Export Citation
  • 24

    Nijjar SPatel BMcGinn GWest M: Computed tomographic angiography as the primary diagnostic study in spontaneous subarachnoid hemorrhage. J Neuroimaging 17:2952992007

    • Search Google Scholar
    • Export Citation
  • 25

    Perry JJStiell IGSivilotti MLBullard MJEmond MSymington C: Sensitivity of computed tomography performed within six hours of onset of headache for diagnosis of subarachnoid haemorrhage: prospective cohort study. BMJ 343:d42772011

    • Search Google Scholar
    • Export Citation
  • 26

    Rinkel GJDjibuti MAlgra Avan Gijn J: Prevalence and risk of rupture of intracranial aneurysms: a systematic review. Stroke 29:2512561998

    • Search Google Scholar
    • Export Citation
  • 27

    Rinkel GJvan Gijn JWijdicks EF: Subarachnoid hemorrhage without detectable aneurysm. A review of the causes. Stroke 24:140314091993

    • Search Google Scholar
    • Export Citation
  • 28

    Rinkel GJWijdicks EFHasan DKienstra GEFranke CLHageman LM: Outcome in patients with subarachnoid haemorrhage and negative angiography according to pattern of haemorrhage on computed tomography. Lancet 338:9649681991

    • Search Google Scholar
    • Export Citation
  • 29

    Rinkel GJWijdicks EFVermeulen MHageman LMTans JTvan Gijn J: Outcome in perimesencephalic (nonaneurysmal) subarachnoid hemorrhage: a follow-up study in 37 patients. Neurology 40:113011321990

    • Search Google Scholar
    • Export Citation
  • 30

    Sato MNakano MSasanuma JAsari JWatanabe K: Preoperative cerebral aneurysm assessment by three-dimensional magnetic resonance angiography: feasibility of surgery without conventional catheter angiography. Neurosurgery 56:9039122005

    • Search Google Scholar
    • Export Citation
  • 31

    Westerlaan HEvan der Vliet AMHew JMMetzemaekers JDMooij JJOudkerk M: Magnetic resonance angiography in the selection of patients suitable for neurosurgical intervention of ruptured intracranial aneurysms. Neuroradiology 46:8678752004

    • Search Google Scholar
    • Export Citation
  • 32

    White PMWardlaw JMEaston V: Can noninvasive imaging accurately depict intracranial aneurysms? A systematic review. Radiology 217:3613702000

    • Search Google Scholar
    • Export Citation
  • 33

    Wilms GGuffens MGryspeerdt SBosmans HMaaly MBoulanger T: Spiral CT of intracranial aneurysms: correlation with digital subtraction and magnetic resonance angiography. Neuroradiology 38:Suppl 1S20S251996

    • Search Google Scholar
    • Export Citation
  • 34

    Yoon DYLim KJChoi CSCho BMOh SMChang SK: Detection and characterization of intracranial aneurysms with 16-channel multidetector row CT angiography: a prospective comparison of volume-rendered images and digital subtraction angiography. AJNR Am J Neuroradiol 28:60672007

    • Search Google Scholar
    • Export Citation

If the inline PDF is not rendering correctly, you can download the PDF file here.

Article Information

Address correspondence to: Lisa E. Thomas, M.D., Department of Emergency Medicine, UM UCMC, 500 Upper Chesapeake Dr., Bel Air, MD 21014. email: lisaethomas@gmail.com.

Please include this information when citing this paper: published online April 18, 2014; DOI: 10.3171/2014.3.JNS132239.

© AANS, except where prohibited by US copyright law.

Headings

Figures

  • View in gallery

    Study flow diagram. *Other exclusion factors: ventriculoperitoneal shunt, trauma within 2 weeks of ED visit, neurosurgery within 4 weeks of ED visit, LP primarily for cytology. §Some patients had more than 1 reason for exclusion. #Some patients had more than 1 type of cerebrovascular imaging.

References

  • 1

    Agid RLee SKWillinsky RAFarb RIterBrugge KG: Acute subarachnoid hemorrhage: using 64-slice multidetector CT angiography to “triage” patients' treatment. Neuroradiology 48:7877942006

    • Search Google Scholar
    • Export Citation
  • 2

    Alexander MSDias PSUttley D: Spontaneous subarachnoid hemorrhage and negative cerebral panangiography. Review of 140 cases. J Neurosurg 64:5375421986

    • Search Google Scholar
    • Export Citation
  • 3

    Andaluz NZuccarello M: Yield of further diagnostic workup of cryptogenic subarachnoid hemorrhage based on bleeding patterns on computed tomographic scans. Neurosurgery 62:104010472008

    • Search Google Scholar
    • Export Citation
  • 4

    Bederson JBConnolly ES JrBatjer HHDacey RGDion JEDiringer MN: Guidelines for the management of aneurysmal subarachnoid hemorrhage: a statement for healthcare professionals from a special writing group of the Stroke Council, American Heart Association. Stroke 40:99410252009

    • Search Google Scholar
    • Export Citation
  • 5

    Canhão PFerro JMPinto ANMelo TPCampos JG: Perimesencephalic and nonperimesencephalic subarachnoid haemorrhages with negative angiograms. Acta Neurochir (Wien) 132:14191995

    • Search Google Scholar
    • Export Citation
  • 6

    Carstairs SDTanen DADuncan TDNordling OBWanebo JEPaluska TR: Computed tomographic angiography for the evaluation of aneurysmal subarachnoid hemorrhage. Acad Emerg Med 13:4864922006

    • Search Google Scholar
    • Export Citation
  • 7

    Chappell ETMoure FCGood MC: Comparison of computed tomographic angiography with digital subtraction angiography in the diagnosis of cerebral aneurysms: a meta-analysis. Neurosurgery 52:6246312003

    • Search Google Scholar
    • Export Citation
  • 8

    Chen YCSun ZKLi MHLi YDWang WTan HQ: The clinical value of MRA at 3.0 T for the diagnosis and therapeutic planning of patients with subarachnoid haemorrhage. Eur Radiol 22:140414122012

    • Search Google Scholar
    • Export Citation
  • 9

    Chen WWang JXin WPeng YXu Q: Accuracy of 16-row multislice computed tomographic angiography for assessment of small cerebral aneurysms. Neurosurgery 62:1131222008

    • Search Google Scholar
    • Export Citation
  • 10

    Connolly ES JrRabinstein AACarhuapoma JRDerdeyn CPDion JHigashida RT: Guidelines for the management of aneurysmal subarachnoid hemorrhage: a guideline for healthcare professionals from the American Heart Association/American Stroke Association. Stroke 43:171117372012

    • Search Google Scholar
    • Export Citation
  • 11

    Czuczman ADThomas LEBoulanger ABPeak DASenecal ELBrown DF: Interpreting red blood cells in lumbar puncture: distinguishing true subarachnoid hemorrhage from traumatic tap. Acad Emerg Med 20:2472562013

    • Search Google Scholar
    • Export Citation
  • 12

    Edlow JAPanagos PDGodwin SAThomas TLDecker WW: Clinical policy: critical issues in the evaluation and management of adult patients presenting to the emergency department with acute headache. Ann Emerg Med 52:4074362008

    • Search Google Scholar
    • Export Citation
  • 13

    El Khaldi MPernter PFerro FAlfieri ADecaminada NNaibo L: Detection of cerebral aneurysms in nontraumatic subarachnoid haemorrhage: role of multislice CT angiography in 130 consecutive patients. Radiol Med (Torino) 112:1231372007

    • Search Google Scholar
    • Export Citation
  • 14

    Feinstein ARCicchetti DV: High agreement but low kappa: I. The problems of two paradoxes. J Clin Epidemiol 43:5435491990

  • 15

    Flaherty MLHaverbusch MKissela BKleindorfer DSchneider ASekar P: Perimesencephalic subarachnoid hemorrhage: incidence, risk factors, and outcome. J Stroke Cerebrovasc Dis 14:2672712005

    • Search Google Scholar
    • Export Citation
  • 16

    Gibbs GFHuston J IIIBernstein MARiederer SJBrown RD Jr: 3.0-Tesla MR angiography of intracranial aneurysms: comparison of time-of-flight and contrast-enhanced techniques. J Magn Reson Imaging 21:971022005

    • Search Google Scholar
    • Export Citation
  • 17

    Hochberg ARRojas RThomas AJReddy ASBhadelia RA: Accuracy of on-call resident interpretation of CT angiography for intracranial aneurysm in subarachnoid hemorrhage. AJR Am J Roentgenol 197:143614412011

    • Search Google Scholar
    • Export Citation
  • 18

    Jung JYKim YBLee JWHuh SKLee KC: Spontaneous subarachnoid haemorrhage with negative initial angiography: a review of 143 cases. J Clin Neurosci 13:101110172006

    • Search Google Scholar
    • Export Citation
  • 19

    Kang DHPark JLee SHPark SHKim YSHamm IS: Does non-perimesencephalic type non-aneurysmal subarachnoid hemorrhage have a benign prognosis?. J Clin Neurosci 16:9049082009

    • Search Google Scholar
    • Export Citation
  • 20

    Kokkinis CVlychou MZavras GMHadjigeorgiou GMPapadimitriou AFezoulidis IV: The role of 3D-computed tomography angiography (3D-CTA) in investigation of spontaneous subarachnoid haemorrhage: comparison with digital subtraction angiography (DSA) and surgical findings. Br J Neurosurg 22:71782008

    • Search Google Scholar
    • Export Citation
  • 21

    MacDonald AMendelow AD: Xanthochromia revisited: a re-evaluation of lumbar puncture and CT scanning in the diagnosis of subarachnoid haemorrhage. J Neurol Neurosurg Psychiatry 51:3423441988

    • Search Google Scholar
    • Export Citation
  • 22

    McCormack RFHutson A: Can computed tomography angiography of the brain replace lumbar puncture in the evaluation of acute-onset headache after a negative noncontrast cranial computed tomography scan?. Acad Emerg Med 17:4444512010

    • Search Google Scholar
    • Export Citation
  • 23

    Menke JLarsen JKallenberg K: Diagnosing cerebral aneurysms by computed tomographic angiography: meta-analysis. Ann Neurol 69:6466542011

    • Search Google Scholar
    • Export Citation
  • 24

    Nijjar SPatel BMcGinn GWest M: Computed tomographic angiography as the primary diagnostic study in spontaneous subarachnoid hemorrhage. J Neuroimaging 17:2952992007

    • Search Google Scholar
    • Export Citation
  • 25

    Perry JJStiell IGSivilotti MLBullard MJEmond MSymington C: Sensitivity of computed tomography performed within six hours of onset of headache for diagnosis of subarachnoid haemorrhage: prospective cohort study. BMJ 343:d42772011

    • Search Google Scholar
    • Export Citation
  • 26

    Rinkel GJDjibuti MAlgra Avan Gijn J: Prevalence and risk of rupture of intracranial aneurysms: a systematic review. Stroke 29:2512561998

    • Search Google Scholar
    • Export Citation
  • 27

    Rinkel GJvan Gijn JWijdicks EF: Subarachnoid hemorrhage without detectable aneurysm. A review of the causes. Stroke 24:140314091993

    • Search Google Scholar
    • Export Citation
  • 28

    Rinkel GJWijdicks EFHasan DKienstra GEFranke CLHageman LM: Outcome in patients with subarachnoid haemorrhage and negative angiography according to pattern of haemorrhage on computed tomography. Lancet 338:9649681991

    • Search Google Scholar
    • Export Citation
  • 29

    Rinkel GJWijdicks EFVermeulen MHageman LMTans JTvan Gijn J: Outcome in perimesencephalic (nonaneurysmal) subarachnoid hemorrhage: a follow-up study in 37 patients. Neurology 40:113011321990

    • Search Google Scholar
    • Export Citation
  • 30

    Sato MNakano MSasanuma JAsari JWatanabe K: Preoperative cerebral aneurysm assessment by three-dimensional magnetic resonance angiography: feasibility of surgery without conventional catheter angiography. Neurosurgery 56:9039122005

    • Search Google Scholar
    • Export Citation
  • 31

    Westerlaan HEvan der Vliet AMHew JMMetzemaekers JDMooij JJOudkerk M: Magnetic resonance angiography in the selection of patients suitable for neurosurgical intervention of ruptured intracranial aneurysms. Neuroradiology 46:8678752004

    • Search Google Scholar
    • Export Citation
  • 32

    White PMWardlaw JMEaston V: Can noninvasive imaging accurately depict intracranial aneurysms? A systematic review. Radiology 217:3613702000

    • Search Google Scholar
    • Export Citation
  • 33

    Wilms GGuffens MGryspeerdt SBosmans HMaaly MBoulanger T: Spiral CT of intracranial aneurysms: correlation with digital subtraction and magnetic resonance angiography. Neuroradiology 38:Suppl 1S20S251996

    • Search Google Scholar
    • Export Citation
  • 34

    Yoon DYLim KJChoi CSCho BMOh SMChang SK: Detection and characterization of intracranial aneurysms with 16-channel multidetector row CT angiography: a prospective comparison of volume-rendered images and digital subtraction angiography. AJNR Am J Neuroradiol 28:60672007

    • Search Google Scholar
    • Export Citation

TrendMD

Metrics

Metrics

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 327 291 5
PDF Downloads 221 192 6
EPUB Downloads 0 0 0

PubMed

Google Scholar