Accepted Manuscript Delayed cerebral ischemia predicts neurocognitive impairment following aneurysmal subarachnoid hemorrhage Martin N. Stienen , MD Nicolas R. Smoll , MD Rahel Weisshaupt , MSc Javier Fandino , MD Gerhard Hildebrandt , MD Aline Studerus-Germann , MSc Bawarjan Schatlo , MD PII:
S1878-8750(14)00470-7
DOI:
10.1016/j.wneu.2014.05.011
Reference:
WNEU 2366
To appear in:
World Neurosurgery
Received Date: 30 August 2013 Revised Date:
21 October 2013
Accepted Date: 6 May 2014
Please cite this article as: Stienen MN, Smoll NR, Weisshaupt R, Fandino J, Hildebrandt G, StuderusGermann A, Schatlo B, Delayed cerebral ischemia predicts neurocognitive impairment following aneurysmal subarachnoid hemorrhage, World Neurosurgery (2014), doi: 10.1016/j.wneu.2014.05.011. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Abstract Objective: Prior studies have shown that the incidence of neuropsychological deficits (NPD) after aneurysmal subarachnoid hemorrhage (aSAH) is high despite excellent outcome
RI PT
according to neurological grading scales. Delayed cerebral ischemia (DCI) occurs in 30% of patients after aSAH and significantly contributes to the mortality and morbidity of aSAH. We tested the hypothesis that DCI is associated with neuropsychological outcome.
SC
Methods: Files of patients treated between 01-2009 and 08-2012 at two neurovascular centers were reviewed. Neuropsychological outcome was assessed in a 2-2.5h face-to-face-interview
M AN U
and graded as no (regular), minimal, moderate or severe deficit by an experienced, independent neuropsychologist according to normative population data. The test battery was applied considering the patients’ individual pre-morbid level of workload and social activities and accounted for the following cognitive domains: (A) Memory; (B) Alertness; (C)
TE D
Executive Function; (D) Visual and spatial perception; (E) Language/Calculation; (F) Behaviour.
Results: Of 226 patients treated at the two centers, 187 were discharged alive. Full
EP
neuropsychological outcome assessment was available in 92 patients, which served as the
AC C
study population. Twenty-eight (30.4%) patients developed DCI, of which twenty-four patients (85.7%) revealed moderate to severe NPD. From a univariate perspective, patients suffering from DCI were 6.38 times as likely to suffer moderate to severe NPD after aSAH (OR; 95% CI: 1.98 – 20.50; p=0.002), which remained statistically significant after correction for admission WFNS and Fisher-grades, patient age, hydrocephalus, and further potential confounders (OR 4.9; 95% CI 1.26 – 19.58; p=0.022). Of all factors analyzed, DCI was the strongest predictor of NPD in the multivariate analysis, followed by chronic hydrocephalus (OR 4.85; 95% CI 1.26 – 18.63; p=0.022) and patient age ≥ 50 years (OR 4.06; 95% CI 1.39 – 11.92; p=0.001).
ACCEPTED MANUSCRIPT Conclusion: Patients with evidence of DCI during the hospital course have a six-fold increased risk of suffering from moderate to severe neuropsychological deficits than patients who do not suffer from DCI after aSAH. Secondary events occurring during the acute hospitalization may be more important to the overall neuropsychological outcome than
RI PT
hemorrhage grading (Fisher) and clinical severity (WFNS) scores at admission.
SC
Key words:
AC C
EP
TE D
M AN U
subarachnoid hemorrhage; cognitive impairment; aneurysm; neuropsychology; vasospasm
Stienen
ACCEPTED MANUSCRIPT Delayed cerebral ischemia predicts neurocognitive impairment following aneurysmal subarachnoid hemorrhage
Martin N. Stienen, MD1; Nicolas R. Smoll, MD2; Rahel Weisshaupt, MSc3; Javier Fandino,
MD3,6
RI PT
MD4; Gerhard Hildebrandt, MD1; Aline Studerus-Germann, MSc5 and Bawarjan Schatlo,
Department of Neurosurgery, Kantonsspital St.Gallen, Switzerland
2
Department of Neurosurgery, University Hospital Geneva, Switzerland
3
Neuropsychology Unit, Department of Neurology, Kantonsspital Aarau, Switzerland
4
Department of Neurosurgery, Kantonsspital Aarau, Switzerland
5
Neuropsychology Unit, Department of Neurology, Kantonsspital St.Gallen, Switzerland
6
Department of Neurosurgery, Georg-August-University, Göttingen, Germany
Martin N. Stienen, MD Department of Neurosurgery
Rorschacher Str. 95
EP
Kantonsspital St.Gallen
TE D
Correspondence and reprints:
M AN U
SC
1
AC C
9007 St.Gallen, Switzerland Phone +41 71 494 1111 Fax +41 71 494 2883
Email:
[email protected] 1
Stienen
ACCEPTED MANUSCRIPT Introduction Aneurysmal subarachnoid hemorrhage (aSAH) is associated with high mortality and morbidity, which occurs during the early and later stages of onset.(19) Early brain injury can be attributed to the initial bleeding, re-bleeding before occlusion therapy and as a
RI PT
complication of aneurysm treatment itself, amongst others. The subacute phase after aSAH carries the risk of delayed neurological deterioration due to cerebral vasospasm (CVS) and delayed cerebral ischemia (DCI). CVS is the delayed narrowing of one or more basal arteries
SC
of the brain after aSAH, which may go in line with radiographic evidence of diminished cerebral perfusion distal of the affected arteries.(4) DCI may occur as a consequence of CVS
M AN U
and may present as focal neurological deficit and/or as worsening on the GCS scale.(24) After aSAH, angiographic CVS is seen in 30 – 70% of patients, typically occurs 3 to 5 days after the bleed, with maximal vessel narrowing between days 5 – 14 and a gradual resolution over 2 – 4 weeks.(4) While some patients with CVS will not develop symptoms, about half of
TE D
cases will deteriorate and develop neurological symptoms, which later may resolve or progress to cerebral infarction with equal likelihood. The incidence of CVS and DCI is higher in patients with worse admission scores according to the Fisher-grade, WFNS- and GCS
EP
score, and a higher amount of intracranial blood.(24)
AC C
Both CVS and DCI are associated with the presence of infarcts and worse neurological outcome.(17) As a matter of fact, 15 – 20% of patients with CVS suffer stroke or die despite maximal therapy. CVS and DCI thus account for nearly 50% of deaths in aSAH patients surviving until after the aneurysm repair.(4) Still, an increasing number of patients survive aSAH, and therefore functional aspects of the clinical outcome become increasingly important. Neuropsychological deficits (NPD) in patients surviving aSAH are frequent,(14) even though often ignored in clinical studies and outcome reports.(21) Reasons for impaired neurocognitive function remain to be identified and must be confirmed by further
2
Stienen
ACCEPTED MANUSCRIPT investigations. Therefore, it was our aim to test the hypothesis that DCI has a negative impact on the neuropsychological outcome in aSAH patients.
RI PT
Methods In a retrospective two-centre approach, a cohort study was performed analyzing aSAH patients that were treated at the Kantonsspital St.Gallen (KSSG) or Kantonsspital Aarau
SC
(KSA) between 01-2009 and 08-2012. The clinical data, including the presence or absence of CVS and DCI, was extracted from the Swiss SOS database (19) by two neurosurgeons of both
M AN U
centres according to predefined criteria. Both local ethical committees consented to the study. The study conformed to the Declaration of Helsinki and written informed consent was obtained from all participants or their next of kin.
This study sought to assess the relationship between DCI (independent variable) and
TE D
neuropsychological deficits (dependant variable and primary outcome). An additional analysis was performed with respect to the occurrence of CVS. CVS was defined as mean blood flow velocity (Vmean) > 140 cm/sec or increase in Vmean > 50cm/sec/24h or as a Lindegaard-
EP
Index > 3 in transcranial Doppler sonography or confirmation by CT-angiography or digital subtraction angiography. DCI was defined as occurrence of focal neurological impairment, a
AC C
decrease of at least 2 points on the GCS after ruling out of other causes (hydrocephalus, electrolyte disturbance, epilepsy, infection). According to a recent consensus paper suggesting occurrence of new cerebral infarctions on imaging at discharge as surrogate marker, this was also regarded as DCI.(24) Neurological deterioration within 24h following aneurysm occlusion therapy was only considered as DCI if the condition was not a therapy complication of aneurysm occlusion. CVS and DCI were treated by induced arterial hypertension and normovolemia; patients with DCI received additional endovascular angioplasty or chemical
3
Stienen
ACCEPTED MANUSCRIPT vessel dilation, if feasible and deemed meaningful by the treating neurosurgeons and neuroradiologists.(6) Acute hydrocephalus was treated with ventriculostomy (external ventricular drainage = EVD)
RI PT
or a lumbar drain. Chronic hydrocephalus was suspected if patients worsened > 2 points on the GCS and revealed dilated ventricles after clamping of the EVD. All cases were treated by
SC
ventriculoperitoneal shunt placement.
M AN U
Neuropsychological assessment
Only surviving patients with available full neuropsychological assessment (face-to-face interview, usually 2-2.5 hours / patient) by an independent neuropsychologist were selected for analysis. The neuropsychological assessment was performed at a median of 45 days after aneurysm occlusion, representing the time of return home/to a nursing home after completion
TE D
of the stationary rehabilitation. The assessment was not standardized, but tailored to the individual patient by the neuropsychologist performing the assessment. For this, a brief
EP
dialogue at the beginning of the assessment and certain patient-specific factors (e.g. age, premorbid occupation and social responsibilities) were taken into account in order to get a
AC C
preliminary estimation of expected deficits and optimize the subsequent choice of tests during the assessment. The test battery accounted for the following cognitive domains: (A) Memory; (B) Attention; (C) Executive Function; (D) Visual and spatial perception; (E) Language/Calculation; (F) Behaviour and included a combination of the following tests: Alertness (Testbatterie zur Aufmerksamkeitsprüfung, TAP 2.2, (28)), Go/Nogo (TAP 2.2, (28)), Geteilte Aufmerksamkeit (TAP 2.2, (28)), Deux Barrage ((2002), (27)), Farbe-WortInterferenztest (FWIT, after J.R. Stroop, 1985, (3)), Regensburger Wortflüssigkeitstest (RWT (2000), (2)), 5-Punkte-Test (HAMASCH, H5PT-R, (11)), Frontal Assessment Battery 4
Stienen
ACCEPTED MANUSCRIPT Bedside (FAB, (7)), Verbaler Lern- und Merkfähigkeitstest (VLMT, (13)), Rey Complex Figure Test (RCFT (1995), (18)), Tiere-Wörter-Test of the test battery Consortium to Establish A Registry for Alzheimer (CERAD, (8)), Boston Naming Test (CERAD, (8)), MiniMental-Status-Examination (CERAD, (8)), Trail-Making-Test A (CERAD, (8)), Trail-
RI PT
Making-Test B (CERAD), (8)), S-Wörter-Test (CERAD, (8)), Apraxie-Prüfung (Goldenberg, (10)). For the applied tests raw data was compared to normative data, some of which were normative data for a selected age population (Deux Barrage; VLMT), a selected age and pre-
SC
morbid education population (FAB), or age, gender and pre-morbid education population (Alertness; Go/Nogo; Divided Attention; FWIT; RWT; H5PT; RCFT; CERAD). Based on
M AN U
normative population data of the test results and the neuropsychologist’s impression (for “behaviour”, reading, writing and calculation) of the patient during the assessment, the patient’s cognitive status was graded as no (regular), or as minimal, moderate or severe
TE D
disabled. In addition, the affected domains were recorded for each patient.
Statistical Methods
EP
Baseline data was described using frequencies and percents, and imbalances between patients
AC C
with and without DCI were tested for using the Pearson Chi2 test or the Fisher Exact test if a single cell value was less than or equal to 5. Mann-Whitney U-test was used to check for imbalances in groups with and without DCI, for values that are continuous. Logistic regression was used then to assess the effect size of the relationship between DCI and neuropsychological deficits. First, a univariable model was built to assess relationships without adjustment, and then a multivariable model was built using forced-entry methodology. No stepwise or other automated variable selection methods were used due to
5
Stienen
ACCEPTED MANUSCRIPT their well known flaws.(15) Relationships before and after were then observed for changes in relationships. The software used for the statistical analysis was Stata v11.2 (College Station, Texas).
RI PT
Probability values (p-value) < 0.05 were considered statistically significant.
Results
SC
Between 01-2009 and 08-2012 226 aSAH patients were treated at both KSA and KSSG. Of these, 39 died (17.2%) and 187 surviving patients were identified for possible enrolment. Full
M AN U
neuropsychological assessment was available in 92 patients (40.7%; 35 male and 57 female; mean age 51.4 ± 11.6 years (± SD)), which thus served as study population (Figure 1). The median WFNS grade of the total cohort was 2.0, with 62 patients (67%) presenting with a low (1-2) and 30 patients (33%) presenting with a high (3-5) WFNS grade. The median Fisher
TE D
score of the cohort was 3.0, with 66 patients (72%) presenting with Fisher 3-4 hemorrhage and 26 patients (28%) with Fisher 1-2 hemorrhage. Of note, Fisher and WFNS scores were imbalanced between the study groups (table 1): Ninety-six percent of patients with DCI had a
EP
high Fisher grade (3 or 4) hemorrhage and only one patient in the low Fisher grade category
AC C
developed DCI (p=0.001). Patients with a high WFNS score (3 – 5) were also more likely to have DCI compared to patients with a low WFNS score (p=0.001). Aneurysm distribution was as follows: 41 aneurysms of the anterior communicating artery (Acom)/anterior cerebral artery (ACA), 24 aneurysms of the middle cerebral artery (MCA), 17 aneurysms of the internal carotid artery (ICA)/posterior communicating artery (Pcom) and 10 aneurysms of the posterior circulation. The distribution of aneurysms size was similar in both study groups. Information on patient and aneurysm-related parameters of the respective study groups are depicted in table 1. 6
Stienen
ACCEPTED MANUSCRIPT In total, 58 patients (63%) received clipping therapy and 34 patients (37%) were subject to endovascular coiling. Fifty-three patients (58%) received EVD-placement for acute hydrocephalus and a ventriculo-peritoneal shunt had to be implanted in 38 patients (41%) of the total group. CVS without clinical symptoms was detected in 39 patients (42%), 28
RI PT
patients (30%) developed DCI. Table 2 provides information on the clinical course and the treatment performed in patients with and without evidence of DCI. A more aggressive CSF diversion therapy had to be performed in patients of the DCI group with higher rates of EVD-
SC
placement (p=0.007) and decompressive hemicraniectomy (DHC) surgery (p=0.040). A similar amount of patients from both groups received either surgical or endovascular
M AN U
aneurysm occlusion therapy (p=0.270). CVS without clinical symptoms was recorded in a higher fraction of patients who developed DCI at a later time (p=0.023).
TE D
Relationship of DCI to Neuropsychological and Neurological Deficits Neurological and neuropsychological outcome of patients with and without DCI is depicted in table 3. We observed a higher neurological morbidity associated with DCI as demonstrated by
EP
significantly worse modified Rankin scores at discharge, lower proportion of patients with
AC C
good outcome, and higher rates of neurological deficits (p