Article Text

Original research
Ignoring floor and ceiling effects may underestimate the effect of carotid artery stenting on cognitive performance
  1. Martin Scherr1,2,3,6,
  2. Alexander Kunz1,6,
  3. Anselm Doll3,
  4. Johannes Sebastian Mutzenbach1,6,
  5. Erasmia Broussalis1,4,6,
  6. Hans Jürgen Bergmann5,6,
  7. Margarita Kirschner5,6,
  8. Eugen Trinka1,6,
  9. Monika Killer-Oberpfalzer1,4,6
  1. 1Department of Neurology, Christian Doppler Klinik, Paracelsus Medical University, Salzburg, Austria
  2. 2Department of Psychiatry, Klinikum rechts der Isar, Technische Universität München, Munich, Germany
  3. 3TUM Neuroimaging Center, Klinikum rechts der Isar, Technische Universität München, Munich, Germany
  4. 4Research Institute of Neurointervention, Paracelsus Medical University Salzburg, Salzburg, Austria
  5. 5Neuroscience Institute, Christian Doppler Klinik, Paracelsus Medical University Salzburg, Salzburg, Austria
  6. 6Centre for Cognitive Neuroscience, University of Salzburg, Salzburg, Austria
  1. Correspondence to Dr Martin Scherr, Universitätsklinik für Neurologie, Paracelsus Medical University, Ignaz Harrer Straße 79, Salzburg 5020, Austria; martin_scherr{at}


Introduction Data on neuropsychological outcome after carotid artery stenting (CAS) remain inconsistent, furthermore cognitive outcome seems to be unpredictable in the individual case. Previous studies reporting improvement or decline might be due to ceiling and floor effects of the applied cognitive tests. We applied cognitive testing before and after CAS, avoiding the pitfall of ceiling and floor effects.

Methods In our prospective database, we identified 72 patients free of clinical stroke with ≥70% carotid artery stenosis, who were treated with CAS. They were administered a neurocognitive test battery before and 3 months after CAS to compare cognitive performance before and after CAS. To avoid ceiling and floor effects of test performances, we additionally analysed subgroups of patients without baseline floor and ceiling cognitive performance.

Results Pre-interventional to post-interventional cognitive performance improved significantly in the subtests measuring verbal episodic memory; deterioration was observed in spatial memory. The subgroups of patients without baseline floor and ceiling cognitive performance improved in measures of global cognition, verbal episodic memory (patients with left-sided CAS) and divided attention (patients with right-sided CAS); we observed no significant effects in the other domains.

Conclusions Ignoring floor and ceiling effects may underestimate the impact of CAS on cognitive performance.

  • Angiography
  • Atherosclerosis

Statistics from


Carotid artery stenting (CAS) and carotid endarterectomy (CEA) are well established in the therapy of symptomatic carotid artery stenosis.1 ,2

Though results still remain inconsistent, patients with stenosis of the internal carotid artery seem to have impaired cognitive performance compared with control groups.3 Cognitive deficits may not be as apparent as neurological deficits, but they can substantially impair the quality of life of affected patients.4 Mlekusch et al4 suggest that improvement of cognitive performance should be considered as a major therapeutic goal in patients with cerebral atherosclerosis. Beyond prevention of cerebrovascular events, CAS and CEA have been discussed to improve cognitive performance by enhancing cerebral perfusion. Knowledge about the effect of these treatments on cognitive performance may therefore play a role in assessing the risk/benefit ratio of the treatment choice.1

Studies investigating the influence of CAS on cognitive performance are few and led to inconsistent results.1 ,2 ,5–8 Thus, the factors influencing the post-interventional cognitive outcome remain unclear.6 Antonopoulos et al9 conclude in their review that CAS is associated with improvement in at least some cognitive domains (global cognition, memory and attention/psychomotor speed) and has no negative effect on cognition. However, the heterogeneity of study designs makes comparisons difficult. Critically, the relevant studies may have underestimated the effect of baseline cognitive performance on the post-interventional outcome. Specifically, patients with baseline ‘floor’ performance can either improve or hold their level without any option for a further significant decline. However, patients with baseline ‘ceiling’ performance can either deteriorate or hold their level, without the opportunity for further improvement.

We hypothesised that patients undergoing CAS significantly improve in cognitive performance, even after accounting for bottom and ceiling effects of baseline test scores. We compared cognitive performance before and after CAS in a sample of patients with no history of clinical stroke. To test our hypothesis, patients with baseline and ceiling test scores were excluded in every cognitive domain in a second step. The influence of possible confounders (peri-interventional ischemia, atrophy, white matter lesion load, age, contralateral stenosis) on changes in cognitive performance was also tested.

Materials and methods

Subjects and study design

In our prospective institutional stroke database, we identified 72 consecutive asymptomatic (ie, no history of clinical stroke) patients (24 women, 33.3%) with the diagnosis of severe internal carotid artery stenosis. They were treated with CAS and underwent neuropsychological testing, carotid ultrasound scan, MRI and clinical examination following a standardised protocol the day before and 3 months after CAS.

Clinical records were screened by a neurologist to identify possible psychiatric diagnosis that could act as confounders to cognitive performance (eg, depression, schizophrenia, substance abuse).

Clinical examination

Clinical examination was conducted by an independent neurologist, blinded to the neuropsychological data before and shortly after the CAS procedure, as well as at 3 months’ follow-up.

Carotid ultrasound scan

All patients were referred after diagnosis of >70% of the carotid artery on duplex sonography. Grade of stenosis was quantified according to the European Carotid Surgery Trial (ECST10).

Carotid stenosis was determined by colour flow duplex ultrasound scan (ATL iU22 ultrasound machine; Philips Medical System) with a L9-3 (3–9 MHz) linear array transducer. Measurements were performed and read by board-certified ultrasound operators following a standardised protocol. Peak systolic velocities of >215 cm/s were graded as stenosis of ≥70%. Near occlusion was defined as visible plaque that led to marked narrowing of the lumen, and occlusion was defined as no detectable patent lumen seen on colour Doppler ultrasound.

CAS procedure

All patients were given acetylsalicylic acid (100 mg/day) and clopidogrel (75 mg/day) at least 6 days prior to intervention. Whole blood impedance platelet aggregometry was performed to test the response to acetylsalicylic acid and clopidogrel. Only responders were eligible for CAS. All patients were recommended a lifelong application of 100 mg acetylsalicylic acid and 75 mg clopidogrel for a minimum of 12 weeks after CAS.

All CAS procedures were carried out using self-expanding stents, following application of distal filter protection devices, via puncture of the right or left femoral artery. Atropine (0.5–1 mg) was given intravenously to reduce bradycardia and hypotension, potentially associated with carotid dilatation.

Magnetic resonance imaging

All patients underwent MRI within 3 months before and the day after CAS. MRI was performed on a 3 Tesla whole body MRI scanner (Achieva, Philips, The Netherlands). Coronar T1-weighted images (36 slices, TR 451 ms, TE 13 ms, thickness 4 mm/L, flip 80, Turbo factor 1, matrix 512, time 2:31 min), axial T2 images (28 slices, TR 3048 ms, TE 80 ms, thickness 4 mm/L, Turbo factor 15, EPI factor 1, matrix 512, time 1:19 min) and fluid attention inversion recovery (FLAIR) axial sequences (28 slices, TR 10 000 ms, TE 125 ms, TI 2800 ms, thickness 4 mm/L, Turbo factor 32, EPI factor 1, matrix 560, time 3:30 min) were acquired before CAS. Axial T2 images, FLAIR axial sequences and diffusion weighted axial images (DWIs; b value 0 and 1000 s/mm2, TR 2500 ms, echo time 55 ms, field of view 230×230 mm, matrix size 256×256, 28 slices, thickness 5 mm) were assessed.

MRI rating

For the clinical evaluation, MR images were read by neuroradiologists with at least 10 years of experience. Post-interventional DWI-positive lesions were considered as acute infarctions.

Pre-existing changes in cerebral white matter were rated by a neurologist (MS) according to the revised version of the Fazekas scale11 and classified as ‘none’ (=0; no WML), ‘mild’ (=1; single lesions below 10 mm and grouped lesions less than 20 mm in any diameter), ‘moderate’ (=2; single lesions 10–20 mm, grouped lesions more than 20 mm with connecting bridges only between the individual lesions) or ‘severe’ (=3; confluent lesions with more than 20 mm in any diameter).

Global cerebral atrophy was rated by a neurologist on a four-point Likert scale and classified as ‘none’ (0), ‘mild’ (1), ‘moderate’ (2) and ‘severe’ (3) using the approach of Leys et al.12

Absence of cortical and/or corticosubcortical non-lacunar infarcts, watershed infarcts, haemorrhages and signs of normal pressure hydrocephalus was required.

Neuropsychological examination

All patients performed a neuropsychological examination the day before the intervention and a neuropsychological follow-up-examination 3 months after CAS. Each patient completed the Mini-Mental Status Exam (MMSE13) and the neuropsychological test battery established by the Consortium to Establish a Registry for Alzheimer’s Disease (CERAD-Plus)14 ,15 at baseline and at 3 months’ follow-up. The CERAD is composed of several subtests: Semantic Verbal Fluency (SVF), Phonemic Verbal Fluency (PVF), Boston Naming Test (BNT), Word List Learning (WLL), Word List Delayed Recall (WLDR), Word List Recognition (WLR), Discriminability (D), Figure Copy (FC), Figure Recall (FR), Trail Making Test A and B (TMT-A and TMT-B).

The test battery assesses verbal fluency (SVF, PVF), constructional praxis (FC), figurative memory (FR), verbal short-term and long-term memory (WLL, WLDR, WLR), verbal recognition memory (D), semantic processing (BNT), speed of cognitive processing (TMT A) and divided attention (TMT B).

Scores of each subtest were population-based scores adjusted for age, gender and level of education.

Pre-interventional and post-interventional test scores were compared via two-sided one-sample t-tests for the whole sample. To test lateralisation, subgroup analysis was calculated for patients with right-sided and left-sided CAS separately. To account for possible effects of peri-interventional infarction, a two-sided one-sample t-test (pre-interventional vs post-interventional test scores) was performed with the subgroup of patients with DWI-positive MRI lesions.

To test for possible confounders, the predictive value of age, global atrophy, degree of white matter lesion load and degree of contralateral stenosis (ie, stenosis contralateral to the side where CAS was performed) on changes in cognitive performance (=zpost-interventional–zpre-interventional) was calculated.

Subgroups of patients with baseline test scores ranging from z≤−1 to z≥1 in each of the subtests were excluded to avoid potential influence of floor and ceiling effects. Pre-interventional and post-interventional test scores were compared again in the identified subgroup.

Statistical analysis

SPSS V.16.0 was used for computation. Paired samples t-test was performed to compare pre-interventional with post-interventional neuropsychological test results in whole and subgroup analyses. Multivariate regression analysis was performed to calculate the predictive value of age, global atrophy, degree of white matter lesions and contralateral stenosis on changes in cognitive performance.

Significance was established at p values <0.05.


Demographic characteristics

Data of 72 (24 women, 33.33%) asymptomatic patients with a stenosis of the internal carotid artery, who were treated with CAS, were analysed. CAS was carried out on 33 patients (45.83%) on the left and 39 patients (54.17%) on the right. Mean age was 68.4 years (SD 9.31, range 48–88).

Two patients (2.78%; both women) had a history of depression and were under antidepressant medication during the assessment period. Clinical records revealed no further history of psychiatric comorbidities (eg, substance abuse, schizophrenia, bipolar disorder).

Clinical results

Clinical examination did not reveal a neurological deficit immediately after the intervention and at 3-month follow-up examination in all patients. This accounted for patients with and without DWI-positive lesions in post-interventional MRI.

Carotid ultrasound scan

To control for re-stenosis, all patients were re-evaluated by duplex sonography of the carotid artery the day after CAS and 3 months after CAS. No residual stenosis or re-stenosis at follow-up was detected. Within-stent flow velocities were <120 cm/s in all post-interventional ultrasound scans.


After the CAS stenting procedure, MRI was performed to detect potential peri-interventional infarction. DWI-positive lesions were found in six patients (8.33%). Lacunar infarction occurred in the cerebellum (two patients, 2.78%), the precentral gyrus (two patients, 2.78%), the basal ganglia (one patient, 1.39%) and the parietal subcortical white matter (one patient, 1.39%).

All patients had detectable white mater lesions in pre-interventional MRI scans. The mean white matter lesion load score was 1.30 (SD 0.55).

The mean atrophy score was 1.03 (SD 0.92).

Neuropsychological results

In our collective, pre-interventional to post-interventional cognitive performance significantly improved in the subtests measuring verbal long-term memory; significant deterioration was observed in constructional praxis (see online supplementary table S1).

Age, degree of white matter lesion load, degree of global atrophy and contralateral stenosis did not predict changes in cognitive performance (see online supplementary table S2 for detailed statistics). The subgroup analysis with DWI-positive MRI lesions (indicating peri-interventional embolism; n=6) revealed no significant decline in cognitive performance (see online supplementary table S3).

The subgroup analyses of patients with left-sided and right-sided CAS (tables 1 and 2) revealed that patients with left-sided CAS significantly improved in verbal episodic memory; patients undergoing right-sided CAS declined in spatial memory performance.

Table 1

Pre-interventional and post-interventional neurocognitive performance in patients with left-sided CAS

Table 2

Pre-interventional and post-interventional neurocognitive performance in patients with right-sided CAS

When analysing subgroups of patients without floor and ceiling performance, patients undergoing left-sided CAS showed significant improvement in global cognition, verbal memory and verbal recognition performance (table 3). Patients with right-sided CAS improved in measures of divided attention but showed no significant decline in spatial memory (although a non-significant trend was observed) (table 4).

Table 3

Pre-interventional and post-interventional neurocognitive performance in patients with left-sided CAS

Table 4

Pre-interventional and post-interventional neurocognitive performance in patients with right-sided CAS


In our sample of patients with asymptomatic (ie, without prior clinical stroke) and severe carotid artery stenosis, significant improvements after CAS were found in verbal memory (left-sided CAS); decline was found in spatial memory (right-sided CAS). The subgroups of patients without ceiling or floor performance at baseline showed improvement in global cognition and pronounced improvements in verbal memory (left-sided CAS). Patients undergoing right-sided CAS showed improvements in divided attention and no significant decline in spatial memory.

Studies investigating the influence of carotid artery stenosis on cognitive performance have led to inconsistent results.1 ,3 ,6 ,16–20 Tiemann et al6 concluded that neuropsychological outcome after CAS is unpredictable in the individual case. It is unclear if there is a significant effect on cognition and in which cognitive domains this might occur. Furthermore, specific subgroups might profit from carotid revascularisation therapies.

Direct comparison of study results is difficult since they differ in cognitive tests, tested cognitive domains, severity and side of pre-interventional carotid artery stenosis and time of follow-up. Furthermore, the role of potential confounders remains unclear, for example, age, symptomatic status, contralateral carotid artery stenosis or the use of protection devices.9

Antonopoulos et al9 concluded in their meta-analysis that CAS may be associated with improvement in global cognition (measured by MMSE13 among others), memory and attention/psychomotor speed. No significant improvements were found in other domains (executive functions (measured by TMT A and B among others), language (measured by BNT among others), functional ability). Significant deterioration could not be observed in any of the cognitive domains tested in the included studies. These results are partly consistent with our study. Lin et al21 found post-interventional improvements in global cognitive function (measured by MMSE13), and attention and psychomotor processing. Xu et al7 demonstrated improvements in verbal memory 12 weeks after intervention compared with a control group with similar clinical condition. The latter studies are in line with our results.

Grunwald et al22 differentiated between speed and memory tests. They found a significant increase in cognitive speed but could not detect any change in memory function 3 months after the procedure, contradicting our results. Mendiz et al23 found a beneficial effect on executive function and memory 3 months after CAS among their prospective cohort of consecutive patients with unilateral and asymptomatic carotid artery stenosis of 60% or more. This also supports our findings.

Little is known about the pathophysiological mechanism of how carotid atherosclerosis affects cognitive performance.

Carotid atherosclerosis may cause microembolisation and consecutive clinically silent cerebral infarcts;24 the accumulation of these lesions may cause cognitive impairment. Furthermore it has been proposed that carotid atherosclerosis is simply a marker for intracerebral atherosclerosis.3 Intracerebral atherosclerosis may cause microcirculatory disturbances and impaired vasoreactivity, which in turn might cause cognitive impairment. This hypothesis has not been tested so far.

A third hypothesis suggests that cerebral hypoperfusion may be accountable for cognitive impairment caused by carotid atherosclerosis.16 ,25–28 If hypoperfusion is an independent factor of cognitive performance, an improvement of cerebral perfusion might be associated with improvement of cognitive functioning.17 ,29 Carotid artery stenosis may cause cerebral hypoperfusion and—as a consequence—cognitive impairment.16 ,28 Due to sufficient revascularisation and an improvement in cerebral perfusion, interventions like CAS, CEA and percutaneous transluminal angioplasty might lead to an improvement in cognitive functioning. In fact, Chen et al30 found increased cerebral perfusion and an improvement in cognitive performance after CAS. Mlekusch et al4 found a significant cognitive improvement after CAS in a subgroup of patients with angiographic filling of the ipsilateral anterior cerebral artery 6 months after CAS. This might explain why an improvement in frontal lobe perfusion accounts for the improvement in cognition.6

However, cerebral perfusion relies rather on collateralisation and cerebral vasoreactivity of the small vessels than on vascularisation of the large vessels (ie, carotid arteries, vertebral arteries). Following this concept, the impact of a stenosis of the carotid artery on cerebral perfusion plays an ambiguous role.

Our data and the studies finding positive effects of carotid revascularisation on cognitive performance appear to support the hypothesis that hypoperfusion is an independent mechanism of cognitive impairment caused by carotid artery stenosis: hemodynamically relevant carotid artery stenosis is treated while leaving white matter lesion load and intracerebral atherosclerosis as they were; as a result cognitive performance improves. Maybe just a subgroup of patients undergoing CAS with poor collaterals and impaired vasoreactivity and therefore low or lacking vascular reserve capacity improves in cognition and this explains the positive results. This hypothesis has not been tested so far.

Peri-interventional embolisation due to guide wire, catheter manipulation, stent placement and balloon deflation has been suspected to negatively influence cognitive outcome.31 ,32 There is evidence that the presence of lesions as shown in post-interventional diffusion-weighted MRI does not affect neuropsychological performance.6 ,33 Ogasawara et al34 hypothesised that a reduction in cerebral blood flow due to clamping or ballooning might cause post-interventional cognitive impairment. Tiemann et al6 found no association between lesion quantity and cognitive change. This is in line with our results in the subgroup of patients with post-interventional DWI-positive lesions (most likely due to peri-interventional embolisation) in which no decline in cognitive performance was observed.

To the best of our knowledge, this is the first study to account for baseline floor and ceiling effects when calculating the impact of CAS on cognitive performance. Floor and ceiling performance could bias the results in two ways, either by overestimating the impact of CAS (ie, the significant negative effect is positively biased by the patients with floor performance, whose condition cannot decline any further) or by underestimating the impact of CAS (ie, the significant positive effect is negatively biased by the patients with ceiling performance, whose condition cannot improve anymore). Our sample suggests that the latter assumption applies.

In conclusion, our data suggest that ignoring floor and ceiling effects might underestimate the positive effect of CAS on cognitive performance. When accounting for these effects, our sample showed improvements in global cognition, divided attention and verbal memory. Future and ongoing studies (eg, the Carotid Revascularisation and Medical Management for Asymptomatic Carotid Stenosis Trial35) may focus on this issue.

Our study has several limitations. The neuropsychological test battery established by the CERAD-Plus14 ,15 was designed for Alzheimer's disease. Therefore it might not be sensitive for vascular cognitive impairment. However, no consensus has yet been reached on which tests to use in vascular cognitive impairment. The applied neurocognitive tests could have been influenced by the effect of learning, since we did not use parallel forms. Handedness of the participants was not assessed. We did not control for multiple testing. Furthermore, we did not compare our results with a control group. Due to the small sample size, conclusions from the subgroup analyses should be drawn with caution.


Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.


  • Contributors All authors made substantial contributions to either the conception or the design of the research. Martin Scherr, Alexander Kunz and Monika Killer-Oberpfalzer drafted the work, all other authors revised the draft critically for important intellectual content. All authors gave final approval of the version to be published and agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

  • Funding This research received no specific grant from any funding agency in the public, commercial or not-for-profit-sector

  • Competing interests ET has acted as a paid consultant for Eisai, Ever Neuropharma, Biogen Idec, Medtronics, Bial, and UCB and has received speakers’ honoraria from Bial, Eisai, GL Pharma, GlaxoSmithKline, Boehringer, Viropharma, Actavis, and UCB Pharma in the past 3 years. ET has received research funding from UCB Pharma, Biogen-Idec, Red Bull, Merck, the European Union, FWF Österreichischer Fond zur Wissenschaftsförderung, and Bundesministerium für Wissenschaft und Forschung.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.