Contemporary Incidence of Cognitive Impairment or Dementia in Patients Undergoing Coronary Artery Bypass Grafting: A Systematic Review and Meta-Analysis

Coronary artery bypass grafting (CABG) is one of the most common cardiac surgeries performed worldwide with almost 400,000 surgeries performed annually [1]. CABG is indicated in patients with high-grade blockages in major coronary arteries and/or failed percutaneous coronary intervention [2]. While perioperative and technical advances in CABG have resulted in lower mortality and complication rates, neurological complications such as cognitive impairment and dementia are commonly reported [3, 4]. It is worth noting that some patients may have preexisting mild cognitive impairment or dementia that is not caused by CABG.

There are a few forms of cognitive dysfunction that can occur post-surgery. Delirium is a temporary condition that occurs 24–96 h post-CABG. Cognitive impairment could follow in the short and long-term period beyond 6 to 12 months. Even though this is potentially reversible, there has been increasing concerns regarding links between cognitive impairment and dementia [5, 6]. Complex attention, executive functioning, learning and memory, language, perceptuomotor functioning, and social cognition are affected modestly and substantially in cognitive impairment and dementia, respectively. However, dementia also involves a component of lack of independence in everyday activities and is a permanent condition [7].

Recent studies have reported that cognitive decline was observed in 54% of patients 5 to 7 years post-CABG [8], and dementia was observed in 30.8% of patients’ seven-and-a-half years post-CABG [6]. This has been postulated to be due to chronic cerebral hypoperfusion causing metabolic dysfunction in the brain which ultimately results in brain ischemia and cognitive dysfunction. This is especially so in patients with accompanying carotid atherosclerosis [9, 10]. Previous research has also shown that metabolic diseases such as hypertension (HTN), diabetes mellitus (DM) and hyperlipidemia (HLD) are associated with increased risk of postoperative cognitive dysfunction [11]. HTN can result in vascular remodeling and decreased blood flow to the brain which greatly increases the risk of stroke and cognitive dysfunction [12]. Other risk factors include advanced age, fewer years of education, systemic inflammation, depression, anxiety, duration of anesthesia and surgery [13, 14]. Subtle cognitive impairment tends to precede dementia and is also associated with similar risk factors, including cardiovascular diseases [6].

Postoperative cognitive dysfunction and dementia worsens patients’ quality of life, lengthens hospital stay and recovery post-operation, and is a major cause of increased morbidity and mortality [15]. Such neuropsychiatric sequelae may also lead to increased healthcare costs [16‒18]. Hence, understanding the relationship between cognitive impairment and dementia in patients undergoing CABG may allow for earlier detection of at-risk patients, including patients with preexisting mild cognitive dysfunction, and for potential targeted intervention.

A recent meta-analysis has been performed to examine cognitive outcomes post-CABG from 1983 to 2017 [3]. However, as studies analyzed spanned across 3 decades, there are several limitations present. Improvements in surgical technology may have impacted cognitive outcomes post-CABG and this is evident with the reduced incidence of cognitive outcomes post-CABG compared to previous years [3]. Furthermore, the associated risk factors contributing to cognitive outcomes post-CABG were not investigated. In addition, since delirium is a temporary condition, we decided to focus solely on cognitive impairment and dementia in our paper as these have longer term repercussions. Hence, our paper aims to investigate the contemporary incidence and risk factors associated with cognitive impairment or dementia in patients undergoing CABG. Identifying this would highlight a more relevant burden of morbidity in these patients and hence allow for further research and targeted interventions toward at-risk groups.

This meta-analysis was registered on PROSPERO (CRD42023401155) and was reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines [19]. Searches of three databases (PubMed, SCOPUS, Embase) were conducted for articles published from 2013 to March 01, 2023. Papers published more than 10 years ago were excluded to prevent analyses of outdated material due to improvements in technology and surgical technique. The search strategy included only terms relating to or describing the intervention or outcome. Search terms included: “coronary artery bypass” OR CABG AND cognitive impair* OR Dementia* OR postoperative cognitive OR post-operative cognitive OR POCD OR cognit* OR mild-cognitive impairment* OR mild cognitive impairment* OR MCI OR neurocognit* OR “neurocognitive decline.” The search terms were adapted for use with other bibliographic databases in combination with database-specific filters for randomized-controlled trials and cohort studies, where these were available. There were no language restrictions. The search strategies are found in Supplementary Table 1 (for all online suppl. material, see https://doi.org/10.1159/000540450).

Table 1 describes the Population, Intervention, Comparison, Outcomes and Study inclusion and exclusion criteria used for study selection. We included peer-reviewed and full-text studies which investigated patients undergoing all forms of CABG surgery or reported results of those who had undergone CABG surgery. Studies had to include a cognitive outcome (cognitive impairment or intact cognition), using an objective test. Duplicate studies of the same cohort were excluded.

Study titles retrieved using the search strategy were screened by two review authors to identify studies that potentially met the inclusion criteria outlined above. The full text of these potentially eligible studies was retrieved and independently assessed for eligibility by the two review authors. Data extraction of articles included in this systematic review and meta-analysis was performed by two review authors. A standardized data collection sheet was used to extract data from the included studies for assessment of study quality and evidence synthesis. Extracted information included those relating to study information: country, study type, time period, outcome assessment criteria for cognitive impairment and dementia; study population and participant demographics: sample size, age, gender, and cognitive impairment; study methodology; study outcomes; and suggested mechanisms of intervention action. Subgroup analyses and meta-regression by risk factors were performed for patients with cognitive impairment or dementia to determine their influence on the results. The categories for the categorical variables were divided based on an equal distribution to ensure statistical power and methodological rigor. As different studies had varying time periods, our data were further divided into three different timeframes – less than 1 month, 2 to 6 months, more than 12 months. Separate analyses for incidence and prevalence of cognitive impairment or dementia were also done based on individual studies’ inclusion and exclusion criteria. A study was determined to be studying prevalence of cognitive impairment or dementia if they did not screen and/or test for cognitive dysfunction prior to CABG or included patients with mild cognitive impairment in their study. Studies used in the analysis of incidence of cognitive impairment or dementia must screen and/or test for cognitive dysfunction prior to CABG and exclude patients with any form of cognitive impairment in their study. A supplementary analysis was also done on studies that were published in the last 5 years only.

A narrative synthesis of the findings from the included studies, structured around the presence or absence of cognitive impairment or dementia as well as target population characteristics were provided. A meta-analysis was conducted, as sufficient data were available. All analyses were conducted using meta and metafor packages on R (version 4.1.0), in accordance with statistical approaches laid out by the Cochrane Handbook [20]. Metaprop was used to meta-analyze the proportions/prevalence under a generalized linear mixed model for number of events every 100 observations. Maximally covariate-adjusted estimates were favored for observational studies. Sensitivity analyses were conducted using the leave-one-out analysis, random effects and identification, and exclusion of outliers.

For studies that used an analytical method that is incompatible for synthesis with a majority of the other studies, the effect estimate was converted to an appropriate ratio for synthesis, or the study was excluded from the meta-analysis. Between-study heterogeneity was assessed using the Q-test or the I² statistic. I² of <30% indicated low heterogeneity between studies, 30–60% indicated moderate heterogeneity, and >60% indicated substantial heterogeneity [21]. Two-sided p values of <0.05 were regarded as statistically significant for the purpose of these analyses. Finally, publication bias was assessed. Quantitative bias was assessed using Egger’s test [22] and qualitative bias was assessed via visual inspection of funnel plot asymmetry. Sensitivity analyses using the trim-and-fill method [23] (fixed-random effects models and R0 estimator) to reestimate the pooled effect size after imputing potential studies were done if publication bias was suspected. There will be a normal distribution of effect sizes around the center of the funnel plot if there is no publication bias [24].

Quality control of included studies was performed by two review authors who independently assessed the risk of bias using the Cochrane Risk of Bias tool [25, 26] in randomized-controlled trials and the Newcastle-Ottawa Scale [27] in cohort studies. As per the NOS grading in past reviews, studies were graded as having a high (<5 stars), moderate (5–7 stars) or low risk of bias (≥8 stars). A Preferred Reporting Items of Systematic Reviews and Meta-analyses checklist [19] is included in online supplementary Table 2.

The PRISMA flowchart is presented in Figure 1. The search identified 1,353 studies, of which 489 articles were screened by title and abstract, following duplicate removal and exclusion of articles that were published more than 10 years ago. Therefore, 23 studies [6, 8, 12, 14, 28‒46] between 2013 and 2023, with a total of 2,620 patients met the inclusion criteria for meta-analysis. 19 studies [8, 12, 14, 28‒43] (twelve cohort studies [8, 12, 14, 28, 32‒34, 36, 37, 40, 42, 43] and seven randomized-controlled trials [29‒31, 35, 38, 39, 41]), with 2,043 patients studied incidence of cognitive impairment or dementia post-CABG. Four studies (two cohort [6, 44] and two randomized-controlled trials [45, 46]), with 577 patients studied the prevalence of cognitive impairment or dementia post-CABG.

Clinical characteristics of each study were collated and are presented in Table 2. The cohort comprises approximately 69% males of mean age 63.4 ± 7.4. Important cardiovascular risk factors like HTN, HLD, DM, and smoking status were included from studies with available data. Table 3 summarizes the cognitive assessments conducted and diagnostic criteria for cognitive impairment and dementia in the studies.

Meta-analyses of sixteen studies [12, 14, 28‒30, 33‒43] showed that the incidence of cognitive impairment or dementia less than 1 month post-CABG was 35.96% (95% confidence interval [CI]: 28.22–44.51, I² = 87%) (Fig. 2a). Subgroup analyses revealed that HLD, HTN, DM, gender, smoking status, left ventricle ejection fraction, age and presence of carotid artery stenosis were not statistically significant (online suppl. Table 3). Meta-regression revealed that studies with more than 80% of the cohort diagnosed with HTN were significantly associated with incidence of cognitive impairment or dementia less than 1 month post-CABG. Other variables including gender, smoking status, left ventricle ejection fraction, and age were not statistically significant (online suppl. Table 4).

Meta-analyses of ten studies [12, 14, 28, 31, 32, 36, 37, 39, 40, 42] showed that the incidence of cognitive impairment or dementia 2 to 6 months post-CABG was 21.33% (95% CI: 13.44–32.15, I² = 88%) (Fig. 2b). Subgroup analyses revealed that HTN, DM, gender, smoking status, and age were not statistically significant (online suppl. Table 5). Meta-regression revealed that HTN, gender, smoking status, DM, and age were not statistically significant (online suppl. Table 6).

Meta-analyses of two studies [8, 36] showed that the incidence of cognitive impairment or dementia more than 12 month post-CABG was 39.13% (95% CI: 21.72–58.84, I² = 84%) (Fig. 2c).

Meta-analyses of three studies [44‒46] showed that prevalence of cognitive impairment or dementia less than 1 month post-CABG was 36.31% (95% CI: 24.79–49.64, I² = 86%) (Fig. 2d). Assessment of the quality of the 23 studies included in the meta-analysis was done in accordance with the JBI checklist and is presented in online supplementary Tables 7 and 8. Only one paper [40] was considered a poor-quality study due to the lack of controls in the study. Overall, no significant risk of bias was identified in the rest of the studies.

There were some publication biases visualized in the sensitivity analyses, funnel plots, trim-and-fill, and Egger’s test (online suppl. Fig. 1–10). The supplementary analysis on studies that were published in the last 5 years showed that incidence of cognitive impairment or dementia less than 1 month post-CABG was 25.16% (95% CI: 21.97–28.64, I² = 51%) (online suppl. Fig. 11) and incidence of cognitive impairment or dementia 2 to 6 months post-CABG was 19.51% (95% CI: 8.53–38.64, I² = 51%) (online suppl. Fig. 12). This follows a similar trend to our original analysis where incidence of cognitive impairment or dementia less than 1 month and 2 to 6 months post-CABG were 35.96% (95% CI: 28.22–44.51, I² = 87%) and 21.33% (95% CI: 13.44–32.15, I² = 88%), respectively. In both analyses, there was a decrease in incidence of cognitive impairment or dementia. As there was only one study investigating the incidence of cognitive impairment or dementia more than 12 months post-CABG, an analysis was unable to be completed.

In this systematic review and meta-analysis of 23 contemporary studies, we demonstrated that there was a high incidence of cognitive impairment or dementia in patients’ post-CABG, particularly in the short-term follow-up period of less than 1 month. Short-term (less than 1 month) incidence of cognitive impairment or dementia was 35.96%. This incidence decreased at 2 to 6 months follow-up to 21.33% and increased to 39.13% at more than 12 months post-CABG. The prevalence of cognitive impairment or dementia had a similar increasing trend of 36.31% (short term follow-up period of less than 1 month) and 43.50% (long term follow-up period of more than 12 months) [6]. For a follow-up duration of less than 1 month post-CABG, we also illustrated that a high prevalence of HTN (HTN in more than 80% of the cohort) had a significant association with cognitive impairment or dementia.

The proposed mechanisms for post-CABG cognitive impairment or dementia include hypoperfusion during CABG leading to brain ischemia particularly in patients with cerebrovascular risk factors and inflammatory responses involving cytokines post-CABG which has been postulated to affect postoperative cognitive decline [9, 47]. Risk of stroke following intensive CABG operation and general anesthesia may also contribute to the neurological sequelae [34]. Other studies have shown that modifiable risk factors such as the presence of preoperative depression, DM, and HTN were found to increase the risk for delirium and cognitive impairment post-CABG [48], and hence should be targeted preoperatively to improve neurological outcomes. This is also supported by our meta-regression which shows that studies with more than 80% of the cohort diagnosed with HTN were significant compared to studies with less than 80% of the cohort diagnosed with HTN. Cerebral autoregulation impairment is associated with cardiac surgery. This causes the brain to be less able to adapt to changes in blood pressure, which increases the risk of damage to the brain due to brain oligemia and hyperemia [49, 50]. Serial follow-up is necessary to truly establish the causal links between the risk factors and cognitive decline [48].

Cognitive impairment or dementia post-CABG is an important public health issue as cognitive decline post-CABG has been associated with increased risk of depression, decreased quality of life, functional capacity, and the ability to perform activities of daily living [51]. Thus, to reduce the economic burden of post-CABG cognitive impairment and dementia, we propose that further research will be needed to allow for earlier identification of risk factors and preoperative, perioperative, and postoperative interventions to reduce the neurological burden post-CABG [48].

In a recent study published by Evered et al. [6], the prevalence of dementia was 30.8% seven-and-a-half years post-CABG and postoperative cognitive dysfunction was found in 32.8% of patients. In this meta-analysis, we demonstrated that the incidence of cognitive impairment or dementia was high in the immediate post-CABG period of less than 1 month, followed by a decrease in incidence in the interim 2 to 6 months, and an increase in incidence in the longer follow-up period of more than 12 months post-CABG. In patients who have cognitive impairment post-CABG, the gradual decline in incidence of cognitive impairment could be due to the biological recovery of the brain after a significant surgery. There will be reduced surgery-induced inflammation and neurodegeneration caused by anesthesia [52]. However, some patients may continue to have cognitive impairment or dementia in the long-term. There has been an ongoing debate regarding early cognitive impairment as a risk factor for late cognitive impairment. However, results have been varied. A few studies showed that long-term cognitive impairment was predicted by early cognitive impairment [53, 54], while others showed no difference [55]. Furthermore, it has been suggested that the long-term prevalence of cognitive impairment or dementia post-CABG is due to the presence of cardiovascular and cerebrovascular diseases rather than anesthesia or the surgery itself. The severity of the patients’ underlying cardiovascular and cerebrovascular diseases is associated with increasing risks of postoperative cognitive dysfunction in aged patients [47]. This could also be precipitated by patients’ pre-morbid cognitive status, which is affected by every individual’s cognitive reserve. Cognitive reserve is a theoretical construct used to describe the brain’s ability to cope with different insults due to age or pathology [52]. Lower cognitive reserve will result in a weaker ability to cope with the damage caused by CABG surgery. Hence, the cerebrovascular system may not be able to adjust appropriately, resulting in long-term cognitive dysfunction [52]. Future studies are required to better delineate the relationship and pathophysiological mechanisms between time course and the risk of developing cognitive impairment or dementia post-CABG.

While risk factors for post-CABG cognitive impairment and dementia have been described [48], moving forward, both risk factors and protective factors should be integrated into existing clinical tools and play a part in the development of prediction tools for postoperative cognitive decline [56, 57]. Such early identification of at-risk patients would allow for timely and targeted intervention to occur. For example, cognitive training has been shown to be effective at modifying cognition, and hence may be carried out preoperatively and perioperatively [58‒61]. Computerized cognitive training is one example of cognitive training, and it involves videogames or virtual reality that targets specific cognitive processes [58, 61]. Furthermore, for patients with early cognitive impairment post-CABG, it would be prudent to provide cognitive training to prevent late cognitive deterioration while research continues in this space [54]. Preoperative cognitive screening is also a cost-effective way to prevent postoperative cognitive dysfunction, as optimization of cognitive status would allow patients to be better able to adapt to the post-operative environment [47]. Other studies have reported that anesthetic preconditioning methods may improve postoperative cognitive dysfunction [62]. Postoperatively, therapeutic agents with neuroprotective effects for the treatment of postoperative cognitive dysfunction may be considered. These medications fall into three broad categories – anti-inflammatory, anti-oxidative, and pro-neuronal, which includes COX-2 inhibitors, statins and dexmedetomidine [63‒66]. However, further studies are required to determine the long-term efficacy of such medications [67, 68].

To the best of our knowledge, this is the first study with a comprehensive and systematic approach to evaluating published studies regarding cognitive impairment or dementia post-CABG in the contemporary era, as well as to examine the associated risk factors. Greaves et al.’s [3] paper in 2019 provided a thorough analysis with 215 studies from 1983 to 2017. However, our paper aims to provide a contemporary approach to this topic by analyzing only recent studies as surgical methods have evolved dramatically since the 1980s, which could have resulted in skewed data as patients with cognitive impairment or dementia post-CABG are comparably smaller compared to in the past [3]. In addition, we did further analyses regarding the association of risk factors with cognitive impairment or dementia post-CABG. Nonetheless, our study should be interpreted in the context of known and potential limitations. First, among statistically significant outcomes, there was substantial heterogeneity for incidence of cognitive impairment or dementia less than 1 month or 2 to 6 months and more than 12 months post-CABG (I² = 87%, I² = 88%, I² = 84%). Furthermore, operation data such as duration of surgery, type of graft used, and presence of concurrent cardiac surgery were not available for most of the included studies, and hence we were unable to determine the effect of operation differences apart from on-pump and off-pump CABG. Even so, only one cohort study compared post-CABG risk of cognitive impairment in patients undergoing off-pump versus on-pump CABG [14]. Future studies should report operation data to allow for better understanding of the relationship between technicalities of the surgery and subsequent risk of developing cognitive impairment and/or dementia. In addition, different studies adopted different definitions and diagnostic criteria for cognitive impairment and dementia. This might lead to either underreporting or overreporting of cognitive impairment or dementia. Hence, there is a need to establish a consistent measure of cognitive performance as part of routine practice. In addition, as there was only one paper that investigated the incidence of dementia in patients after CABG separately from cognitive dysfunction [6], future research should be conducted to examine this in order to better delineate their relationship. Lastly, there were no studies that investigated the relationship of risk factors to the occurrence of cognitive impairment or dementia. Thus, future studies could include the correlation between the risk factors and cognitive impairment or dementia, which would allow the detection of at-risk patients and potential altered management.

In this systematic review and meta-analysis, we demonstrated that there was a significant incidence of cognitive impairment and dementia in patients’ post-CABG, particularly in the immediate follow-up period of less than 1 month post-CABG and more than 12 months post-CABG. In the immediate follow-up period, we also found that HTN in more than 80% of the cohort was significant compared to those with HTN in less than 80% of the cohort. Besides HTN, there are many other risk factors for cognitive decline and dementia which have been highlighted in our review. Strategies moving forward include developing a cognitive decline risk score for CABG patients as well as target modifiable risk factors to reduce patients’ risk of postoperative cognitive decline.

We thank Mr. William Kin for his help in data extraction for this project.

A statement of ethics is not applicable because this study is based exclusively on published literature.

The authors have declared that no competing interests exist.

This work was supported by the National University of Singapore Yong Loo Lin School of Medicine’s Junior Academic Fellowship Scheme and the Singapore Ministry of Health National Medical Research Council’s Clinician Scientist Individual Research Grant New Investigator Grant (MOH-001080-00) and Transition Award (MOH-001368-00).

H.Z.L., C.F.W., C.E.L., Y.H.T., and C.H.S. designed the study and developed the study protocol and tools. H.Z.L., C.F.W., C.E.L., W.K., and C.H.S. were responsible for data collection. H.Z.L., C.F.W., C.E.L., Y.H.T., Y.N.T., C.Y.Y., N.L.S., B.Y.Q.T., P.C., L.L.L.Y., T.‐C.Y., Y.F.C., K.‐K.P., W.K.F.K., R.C.C.W., M.Y.C., and C.‐H.S. analyzed data and wrote the manuscript. All authors contributed to the conceptualization of the research questions, interpretation of the results, and manuscript writing. All authors read and approved the final manuscript.

All data generated or analyzed during this study are included in this article and its supplementary material files. Further inquiries can be directed to the corresponding author.