Article Text

Original article
Cognitive outcomes after acute coronary syndrome: a population based comparison with transient ischaemic attack and minor stroke
  1. Irene Volonghi1,
  2. Sarah T Pendlebury1,2,
  3. Sarah J V Welch1,
  4. Ziyah Mehta1,
  5. Peter M Rothwell1
  1. 1Stroke Prevention Research Unit, Nuffield Department of Clinical Neurology, John Radcliffe Hospital, and the University of Oxford, Oxford, UK
  2. 2NIHR Biomedical Research Centre, John Radcliffe Hospital, Oxford, UK
  1. Correspondence to Dr Sarah Pendlebury, Stroke Prevention Research Unit, Nuffield Department of Clinical Neurology, Level 6 West Wing, John Radcliffe Hospital, Oxford OX3 9DU, UK; sarah.pendlebury{at}


Objective Acute coronary syndrome (ACS) is associated with increased risk of cognitive decline when compared with controls, but case:control studies are subject to selection bias. We therefore compared cognitive outcomes in ACS with transient ischaemic attack (TIA) and minor stroke, diseases with similar risk factor burden generally considered to be at high risk of cognitive decline.

Design Prospective population based cohort study

Setting Oxford Vascular Study (OXVASC) carried out within a defined population of 91 000 in Oxfordshire, UK.

Patients 614 in total: 216 ACS, 182 TIA, 216 minor (non-disabling) stroke.

Outcome measures Mini-Mental-State-Examination (MMSE), Telephone Interview for Cognitive Status-modified (TICSm), and Montreal Cognitive Assessment (MoCA) at 1 and 5 years.

Results Overall risk factor burden was similar across groups but ACS patients had more smoking (27% vs 14%, p<0.001) and less hypertension (45% vs 53%, p<0.01) and atrial fibrillation (6% vs 14%, p<0.001). Cognitive outcomes were worse at 1 year in ACS versus TIA patients: mean±SD MMSE 26.6±2.7 vs 27.6±2.5, p<0.0001; OR=2.14, 95% CI 1.11 to 4.13 for moderate/severe cognitive impairment (MMSE <24) with a similar trend at 5 years, and ACS outcomes were more similar to minor stroke. Memory and language versus frontal/executive subtests were relatively more impaired in ACS than TIA and minor stroke patients.

Conclusions Risk of cognitive impairment after ACS is similar to minor stroke and higher than TIA with implications for clinical practice including consent and adherence with medication. Differences in cognitive domain performance suggest a greater role for degenerative brain pathology in ACS which may be linked to vascular risk profile and cardiac factors.

Statistics from

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.


Available data suggest worse cognitive performance and increased risk of cognitive decline in patients with coronary vascular disease in comparison to controls.1–3 However, a recent review showed that data were limited to a few case:control and cohort studies of stable coronary vascular disease and all but one study did not exclude co-existent stroke.3 Case:control studies may be subject to selection bias in that controls may be particularly healthy volunteers or be selected using inappropriately strict criteria, and cases may themselves be unrepresentative of the total population with the disease. Cohort studies may be similarly affected by volunteer bias. Furthermore, the high prevalence of vascular risk factors in patients with coronary vascular disease should result in an association between acute coronary syndrome (ACS) and cognitive dysfunction over and above any effects of the event itself since vascular risk factors are well recognised predictors of cognitive decline.2 None of the previous studies specifically matched cases and controls for vascular risk factors.3

It might therefore be more informative to compare cognitive outcomes in ACS patients with transient ischaemic attack (TIA) and stroke patients in whom there is a similar high prevalence of vascular risk factors. Although ACS patients have been thought to be at risk of poor cognitive outcomes through indirect effects of cardiac dysfunction on the brain including hypoperfusion,3 ,4 cognition might be expected to be worse after TIA and stroke owing to the powerful effect of the direct brain insult5 ,6 and more cerebrovascular pathology.

We therefore performed a study of cognitive outcome including cognitive domain profile >1 year after ACS versus TIA and minor stroke in a population based study of all acute vascular events.


Patients were participants in the Oxford Vascular Study (OXVASC 2002–2007), a prospective population based cohort study of all acute vascular events occurring within a defined population of approximately 91 000.7 ,8 OXVASC was approved by the local ethics committee and informed consent or assent was obtained from all participants or relatives.

Patients with ACS had ST elevation and non-ST elevation myocardial infarction, or unstable angina, defined by currently accepted criteria.9 ,10 Consecutive patients with ACS recruited from April 2002 to April 2005 had follow-up at 1 year with the Mini-Mental State Examination11 (MMSE, 2002–2003 recruits) and/or Telephone Interview for Cognitive Status-modified12 (TICSm, all those recruited 2003 onwards), and at 5 years with MMSE and Montreal Cognitive Assessment (MoCA).13

TIA and stroke were defined clinically by WHO criteria14 with strokes dichotomised into minor (National Institute of Health Stroke Scale (NIHSS) ≤ 3) and major (NIHSS > 3) events.15 Both primary intracerebral haemorrhage and ischaemic stroke were included. We excluded major stroke since untestability from lesional effects including dysphasia and hemiparesis is common, early mortality is high, and such patients are thus not comparable to those with ACS.5 ,6 Consecutive minor stroke and TIA patients recruited from April 2002 to 2007 had similar cognitive assessments to ACS subjects with MMSE and TICSm at 1 year and MoCA and MMSE at 5 years.

All cognitive testing was performed by trained research nurses (led by SJVW) with several years experience of following up OXVASC patients including performing the cognitive assessments. Cognitive data were checked and cleaned under the supervision of STP. Reasons for not being tested were recorded for all the cohorts. Concurrent treatment for depression was recorded at the time of follow-up. The MMSE, TICSm, and MoCA have all been used previously in studies of cognition in vascular disease,6 ,16–18 and the MoCA has previously been shown to have good sensitivity and specificity for mild cognitive impairment in subjects with TIA and stroke (see online supplementary table S1 and figure S1)16 and detects differences in cognitive domain profile between cerebrovascular and memory research subjects.19

The number of patients in each group undergoing coronary artery bypass graft surgery (CABG) was prospectively recorded since CABG is known to impact on cognitive outcome.20 Subjects with previous stroke or any recurrent ischaemic event during follow-up were excluded from both the cardiac and cerebrovascular cohorts in order to restrict the cohorts to those with first-ever event, either ACS or TIA or minor stroke.

Vascular risk factors were recorded at study entry on the basis of structured questionnaire and general practitioner records submitted at the time of patient referral.7 ,8 Hypertension, diabetes, and hypercholesterolaemia were defined on the basis of history and/or use of relevant medication and smoking as current smoking at the time of study entry.7 ,8 Since previous studies have highlighted the importance of major geriatric syndromes and comorbidity on outcome after acute cardiac conditions,21 the following were also recorded: functional dependency using the modified Rankin Scale (mRS)22 and history of treated depression, cardiac failure, end stage renal disease requiring dialysis or transplant, chronic liver disease, cancer, chronic obstructive pulmonary disease, and asthma.

Statistical analysis

Demographic differences between tested and not tested patients were compared using t test or Fisher exact test. Demographic and vascular risk factor characteristics between tested ACS, TIA, and minor stroke patients were compared using analysis of variance or χ2 test as appropriate. According to published recommendations, any cognitive impairment was defined as TICSm <2512 ,17 and MoCA <26 (an additional point being added to the total score for patients with <12 years of education6 ,13 ,16), while MMSE <24 was taken to indicate moderate/severe cognitive impairment.23 Non-tested patients with dementia were assumed to have scores below the cut-offs—that is, MMSE <24, TICSm <25, and MoCA <26. Although MMSE data were more skewed than MoCA or TICSm data, results using both parametric and non-parametric tests were similar. For between group comparisons of mean cognitive outcomes and cognitive subscores, linear or logistic regression analyses were used as appropriate, adjusted for age, hypertension, and sex. Odds ratio (OR) for likelihood of having cognitive impairment were calculated using TIA as the reference group versus ACS and minor stroke.


At 1 year, n=614/817 subjects (216 (74.2%) ACS, 182 (76.4%) TIA, and 216 (75.0%) minor stroke) completed valid face-to-face cognitive assessment. Numbers were fewer at 5 years owing to death and the fact that some patients had not yet reached 5 year follow-up at the time of data analysis (see online supplementary figure S2). At 1 year, not tested versus tested ACS patients were more likely to be female, but education was similar between tested and not tested patients in all the cohorts. Reasons for not completing cognitive assessment are given in online supplementary figure S2.

Demographic characteristics, risk factor burden, functional status, and comorbidities of tested patients are shown in table 1. Tested ACS patients were younger and more often male compared to the cerebrovascular patients. Although overall vascular risk factor burden in ACS and cerebrovascular patients was similar, ACS patients were more often smokers whereas hypertension and atrial fibrillation (AF) were more frequent in cerebrovascular patients (table 1). Hyperlipidaemia and statin prescription were most common in ACS and TIA subjects. CABG was more likely to be performed in ACS versus TIA and stroke patients (19 (9%) vs 1 (<1%) and 0 (0%), and 16 (12%) vs 0 (0%) and 0 (0%) for 1 and 5 year cohorts, respectively) (table 1). Functional status and comorbidities were similar in the three patient groups (table 1).

Table 1

Demographic details, risk factor burden, risk factor profile and rates of CABG and comorbidities for 1 and 5 years tested ACS, TIA, and minor stroke survivors

Looking at all the cohorts combined, increasing vascular risk factor burden was associated with higher rates of cognitive impairment (TICSm <25 or MoCA <26, table 2). However, the relationship between vascular burden and cognition became slightly less apparent when adjusted for age and education.

Table 2

Cognitive function by number of vascular risk factors for ACS, TIA and minor stroke survivors combined

ACS patients did worse on MMSE at 1 year compared to those with TIA (p<0.0001) and minor stroke (p=0.05) (table 3), and were more likely to have moderate/severe cognitive impairment (MMSE <24): OR=2.14 (95% CI 1.11 to 4.13) for ACS versus TIA at 1 year, compared to OR=1.48 (95% CI 0.62 to 3.50) for minor stroke versus TIA, with a similar trend at 5 years. ACS patients were also more likely to have low MoCA scores at 5 year follow-up compared to those with TIA: OR=2.16 (95% CI 1.01 to 4.61) (table 3). Results were similar when patients with CABG were excluded from analyses (see online supplementary table S2).

Table 3

Comparison of mean cognitive scores and rates and ORs for moderate/severe (MMSE <24) and mild+moderate/severe (TICSm <25, MoCA <26) cognitive impairment across ACS, TIA (reference cohort for ORs) and minor stroke cohorts

Performance on MMSE, TICSm and MoCA subtests is presented in figure 1 and the online supplementary table S3. ACS patients did worse than TIA and minor stroke patients on MMSE recall at both 1 and 5 years and MMSE repetition at 1 year with relatively better performance on drawing (copying pentagons) and subtraction. For all patients, low MMSE recall subscores were associated with reduced likelihood of survival to 5 year follow-up (OR=1.28 (95% CI 1.01 to 1.62) p=0.04 for death before 5 year follow-up for each item not recalled, adjusted for age and education).

Figure 1

Mean 1 year Mini-Mental-State-Examination (MMSE), 1 year Telephone Interview for Cognitive Status-modified (TICSm), 5 year MMSE, and 5 year Montreal Cognitive Assessment (MoCA) subscores shown as a percentage of the maximum subtest score for acute coronary syndrome (solid line), transient ischaemic attack (dashed and dotted line), and minor stroke (dashed line). A=sustained attention, F=verbal fluency.

ACS patients also performed worse than those with TIA and minor stroke on TICSm orientation and repetition although TICSm delayed recall was worse after minor stroke. At 5 years MoCA, ACS patients did relatively better on sustained attention (tapping at the letter A), verbal fluency, and abstraction and subtraction, and had relatively greater impairment of recall and language than cerebrovascular patients (figure 1, see online supplementary table S3).


Our study is the first direct comparison of cognition after ACS versus TIA and minor stroke, diseases in which the vascular risk factor burden is similar. ACS outcomes were comparable to minor stroke and, in fact, worse than after TIA. Relatively more memory and language impairment, and less frontal/executive impairment, was seen in ACS versus cerebrovascular patients, suggesting a greater role for degenerative pathology in ACS.

An increasing number of vascular risk factors was associated with worse cognition in our cohorts overall, in keeping with the known link between vascular risk factors and cognitive decline and dementia,2 but overall risk factor burden was similar across ACS, TIA, and minor stroke groups. One might have expected therefore to see worse global cognitive outcomes after TIA and minor stroke compared to ACS owing to the well recognised impact of symptomatic cerebrovascular disease on cognition,5 whereas we observed that impairment after ACS was greater than after TIA and as poor as after minor stroke.

Although we included patients with acute coronary vascular events, cognitive testing was performed after an interval and thus our patients are likely to be largely similar to those included in studies of stable coronary disease. Our findings suggest that factors other than vascular risk burden are important in the cognitive decline noted in previous case:control and cohort studies.3 Various mechanisms have been proposed including cardiac dysfunction2–4 and CABG in some patients.20 CABG was more prevalent in the ACS patients in our study, but the numbers undergoing surgery were relatively small and excluding these patients did not change our findings. Comorbidity burden, which is known to affect outcome in older cardiac patients,21 was similar between ACS, TIA, and stroke groups so it seems unlikely that this explains the relatively poor cognitive performance of the ACS patients.

The worse than expected global cognitive outcomes after ACS were accompanied by interesting differences in performance on individual cognitive subtest items. Compared to the cerebrovascular patients, the cognitive profile of the ACS patients was more similar to that of memory clinic subjects19 and those with Alzheimer's disease,24 ,25 with relatively greater impairments in memory and language domains versus frontal/executive deficits, the latter being prominent in patients with TIA and stroke.26 However, frontal executive performance was probably not entirely normal in ACS patients as suggested by previous studies,1 since it was similar to that seen in TIA patients who perform worse in this domain than memory clinic subjects.19 The relatively less poor performance of ACS (and cerebrovascular patients) on the MMSE recall subtest at 5 year follow-up may be partly explained by the high mortality rate in those with low MMSE recall scores which has been observed previously.27

The differences in cognitive domain profiles between ACS and cerebrovascular patients suggest that Alzheimer-type pathology may be a relatively more important mediator of cognitive decline in ACS, at least in those without comcomitant symptomatic cerebrovascular disease. Rates of cognitive impairment without dementia are high in older survivors of acute myocardial infarction and heart failure is known to be associated with cognitive impairment and may potentiate Alzheimer-type changes.2–4 While cardiac dysfunction may have an impact on the brain, differences in risk factor profile may also be important: more smoking and less hypertension and AF were seen in ACS patients compared to TIA and stroke patients. Smoking is a risk factor for memory impairment and Alzheimer's disease more than vascular cognitive impairment,28 ,29 whereas hypertension and AF are associated with stroke, small vessel disease, and vascular dementia.

The lack of significantly different performance between ACS and cerebrovascular groups on the recall (and repetition) subtests of the MoCA is at first surprising. However, fewer patients completed the MoCA than the MMSE at 5 year follow-up, resulting in less power to show between group differences, and the MoCA pattern of deficits in ACS versus cerebrovascular patients was similar. Furthermore, the MoCA recall task is harder than the MMSE, resulting in a floor effect and inability to differentiate between subjects with more impaired recall.30 Overall, recent studies suggest that the MoCA would appear to be a better test than the MMSE for cognitive outcomes in cardiovascular disease since the MMSE is insensitive to mild cognitive impairment and has less good coverage of frontal/executive domains.6 ,16

Our study has some limitations. First, we used short cognitive tests (MoCA, MMSE, TICSm) since this was a large population based study of all acute vascular events which included clinical, physiological and other assessments, and an extensive neuropsychological battery was not practical. Furthermore, we have shown that the MoCA, MMSE, and TICSm are valid in our TIA and stroke population.6 ,16–19 Secondly, the MoCA was not available at the outset of the study hence it was only performed at 5 year follow-up. Before that, the TICSm was used as a more sensitive test for milder cognitive impairment than the MMSE.12 ,17 Thirdly, the 5 year survivor cohorts were smaller than the 1 year cohorts, mainly through death and the fact that many cerebrovascular patients had not yet reached 5 year follow-up at the time of data analysis. Detailed examination of the reasons for lack of valid test (see online supplementary figure S2) showed no major bias between tested and untested patients. Finally, comparative neuroimaging data were not available so we could not assess asymptomatic brain lesions or white matter disease, although conventional neuroimaging does not in any case demonstrate neurodegenerative pathology. We excluded patients with symptomatic vascular disease in other territories to try and obtain ‘pure’ cohorts and avoid confounding.

In conclusion, global cognitive outcomes were worse after ACS than after TIA, being similar to that of minor stroke despite comparable overall risk factor burden. Cognitive profile differences suggest that degenerative brain pathology may be relatively more important in ACS subjects and may be linked to risk factor profile (more smoking and less hypertension and AF) and cardiac factors. Our findings have some implications for clinical practice in ACS including the need for care when consenting patients for procedures, and ensuring adherence with medications, particularly given the prominence of memory impairment. Further longitudinal studies are required together with cardiac outcomes and neuroimaging data to determine further the factors impacting on cognitive function in subjects with symptomatic coronary vascular disease.


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.

    Files in this Data Supplement:


  • IV and STP contributed equally to this study.

  • Contributors IV cleaned data, contributed to analyses and drafted the manuscript. STP contributed to analyses and wrote the manuscript. SJVW interviewed patients and collected data. ZM performed statistical analyses, and PMR conceived and directs the OXVASC study and co-wrote the manuscript.

  • Funding The Oxford Vascular Study is funded by the the Wellcome Trust, the UK Stroke Association, the Dunhill Medical Trust, the National Institute of Health Research (NIHR), the Medical Research Council, the NIHR Oxford Biomedical Research Centre, and the Wolfson Foundation. STP is supported by the NIHR Oxford Biomedical Research Centre, UK and PMR has NIHR and Wellcome Trust Senior Investigator Awards.

  • Competing interests None.

  • Ethics approval Local ethics committee.

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