From heart to mind: A narrative review of incidence, risk factors, and mechanisms of dementia after cardiac surgery

Abstract

As cardiac surgery becomes more common, concerns about postoperative cognitive decline and the potential progression to dementia have grown. While postoperative cognitive dysfunction (POCD) is well documented, its relationship to long‐term dementia risk remains unclear. We narratively summarize evidence on dementia incidence, diagnostic approaches, risk factors, and pathophysiological mechanisms. Across the literature, dementia incidence after cardiac surgery varied between 1.5% and 30.8% over 1 to 24 years of follow‐up, reflecting heterogeneous patient populations, surgical procedures, follow‐up durations, and diagnostic criteria. Consistent risk factors included older age, preoperative mild cognitive impairment, cardiovascular comorbidities, and postoperative delirium. Proposed mechanisms involved cerebral microembolism, systemic inflammation, hypoxia‐ischemia, and impaired cerebral autoregulation. Future research would benefit from standardized diagnostic criteria, consistent incidence reporting, and long‐term, multicenter studies to refine risk estimates. Identifying modifiable perioperative factors through targeted screening and prevention strategies may preserve cognition and improve long‐term outcomes after cardiac surgery.

Highlights

• This narrative review examines the epidemiology of dementia following cardiac surgery.

• Dementia incidence after cardiac surgery was 1.5% to 30.8% over 1 to 24 years.

• Heterogeneity exists due to varying populations, surgeries, and follow‐up durations.

• Consistent risk factors included age, cognitive impairment, and cardiac comorbidities.

1. INTRODUCTION

The increasing prevalence of cardiac surgery procedures, especially among older adults, has attracted considerable attention of healthcare professionals to associated postoperative cognitive complications. ¹ While the risk of postoperative cognitive dysfunction (POCD) is well documented following cardiac surgery, the relationship between this initial decline and the development of long‐term cognitive deficits, including dementia, remains unclear. ² Understanding this potential connection is essential due to its significant implications for healthcare expenditures and patient outcomes, particularly quality of life. Elucidating the underlying mechanisms that may contribute to dementia in this vulnerable population is critical for developing effective prevention strategies and optimizing clinical management.

Several biologically plausible mechanisms have been proposed to link cardiac surgery with subsequent cognitive decline, including cerebral microembolism, inflammation, hypoxic, and ischemic injury. ³ , ⁴ These perioperative insults may compound patient‐specific risk factors such as pre‐existing cognitive impairment, genetic susceptibility, and comorbidities. However, interpretation of the literature is limited by heterogeneity in cognitive assessment tools, definitions of POCD and dementia, specific surgical techniques, and follow‐up duration, contributing to inconsistent findings across studies.

This review synthesizes current evidence regarding the incidence, predictors, and mechanisms of dementia after cardiac surgery. We first summarize how POCD and dementia are defined and measured and review the epidemiology of post‐cardiac surgery dementia. We then examine mechanistic pathways, including embolic, inflammatory, and hypoxic‐ischemic processes. Finally, we review clinical and procedural predictors of long‐term cognitive decline, including biomarkers, demographics, comorbidities, medications, baseline cognition, and surgical characteristics.

By integrating epidemiologic, mechanistic, and clinical perspectives, this review aims to inform strategies to preserve cognitive function, including improved postoperative screening and monitoring, targeted interventions, and evidence‐based patient selection for cardiac surgery versus less invasive alternatives.

2. DEFINITIONS

The Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM‐5) classifies dementia under the broader category of major neurocognitive disorder (NCD). ⁵ According to DSM‐5 criteria, major NCD involves significant cognitive decline in at least one cognitive domain (such as attention, memory, or language). This decline must be evident through concerns reported by the individual or others and confirmed by objective cognitive assessments, ideally neuropsychological testing. Additionally, the cognitive impairment must interfere with daily functioning and cannot be better explained by delirium or another mental disorder. Major NCD encompasses various etiologies, including Alzheimer's disease and vascular dementia, each with specific diagnostic criteria and distinct clinical trajectories. ⁶ Depending on the underlying cause, cognitive impairment may remain stable following an initial event, fluctuate over time, or progressively worsen. ⁶

In this review, it is important to distinguish dementia from both POCD and postoperative delirium (POD), as most of the existing work focuses on POCD and/or POD rather than dementia specifically. Varpaei et al. ⁷ defined POCD as cognitive changes affecting one or more domains that occur after major surgery. Unlike dementia, POCD is typically transient and potentially reversible, with durations ranging from weeks to years. POD is defined in the DSM‐5 as an acute, fluctuating disturbance in attention and awareness accompanied by a change in cognition not better explained by a pre‐existing neurocognitive disorder. ⁵ POD typically occurs within days after surgery and represents a distinct postoperative neurocognitive disorder.

3. DIAGNOSIS AND COGNITIVE TESTING

There is considerable variability in diagnostic criteria across studies investigating dementia following cardiac surgery, complicating the comparison and interpretation of results. Some studies rely on clinical diagnoses, while others utilize cognitive screening instruments or administrative data, resulting in inconsistent definitions of dementia outcomes. Standardizing diagnostic approaches would improve comparability and clarity of findings in this field.

Researchers have employed diverse methods to assess cognitive impairment following cardiac surgery. For example, Whitlock et al. ⁸ estimated dementia probability using a memory score derived from questionnaire items calibrated against the Aging, Demographics, and Memory Study sample from the Health and Retirement Study (HRS), along with the Telephone Interview for Cognitive Status and serial‐7 subtraction tasks. When participants became too cognitively impaired to respond directly, proxy reports were utilized. Although these methods are practical for large‐scale assessments, they may be less sensitive in detecting deficits in executive functioning and processing speed. In contrast, Burker et al. ⁹ conducted comprehensive neuropsychological assessments covering five cognitive domains – attention, processing speed, executive function, verbal and visual memory, and language – using standardized instruments such as the Trail Making Test, the Color‐Word Interference Test, and the Hopkins Verbal Learning Test.

A few studies have employed comprehensive approaches to diagnosing dementia, combining cognitive testing with clinical judgment. For instance, Evered et al. ² utilized multiple assessment tools, including the Clinical Dementia Rating scale (CDR), the Informant Questionnaire on Cognitive Decline completed by relatives, the Mini‐Mental State Examination (MMSE), the Consortium to Establish a Registry for Alzheimer's Disease–Alzheimer's Disease Assessment Scale‐Verbal Learning Test (CERAD‐AVLT), and evaluations of activities of daily living (ADLs). Similarly, Greaves and Kuzma integrated clinical examinations, cognitive assessments, questionnaires, and, in Kuzma's case, magnetic resonance imaging (MRI) brain imaging to more accurately classify the dementia subtypes. ¹⁰ , ¹¹

In contrast, other studies employed simpler or administrative data‐based methods. Hata et al. ¹² and Lingehall et al. ¹³ relied exclusively on the MMSE, while several others, including Pan et al., ¹⁴ Giang et al., ¹⁵ Lee et al., ¹⁶ and Mutch et al. ¹⁷ used International Classification of Diseases (ICD) codes extracted directly from medical records. In ICD‐9, dementia is primarily coded under 290 to 294; for example, code 290.0 indicates senile dementia, uncomplicated. With the transition to ICD‐10, dementia coding became more detailed and specific, primarily classified under codes F01‐F03. Specifically, F01 denotes vascular dementia, F02 dementia associated with other diseases (e.g., Alzheimer's or Parkinson's disease), and F03 unspecified dementia.

4. EPIDEMIOLOGY

The age‐standardized prevalence of dementia among individuals aged 60 years and older is estimated at approximately 5% to 7% across most global regions. ¹⁸ In contrast, our review identified considerable variability in reported dementia incidence following cardiac surgery, ranging from 1.5% to 30.8% over follow‐up periods of 1 to 24 years (Table 1). This discrepancy likely reflects differences in study populations, surgical procedures, cohort sizes, follow‐up durations, and diagnostic criteria.

TABLE 1.

Incidence of dementia following cardiac surgery across 13 studies.

Study	Year	Procedure	N	Dementia incidence	Age (mean)	Follow‐up duration	Notes
Whitlock et al. 8	2021	CABG PCI	665 1015	10.5% 9.6%	74.8 75.0	10 years a 10 years a	No significant difference in long‐term dementia probability.
Bürker et al. 9	2017	Heart transplant	37	2.7%	37.2	20.3 years a	n/a
Greaves et al. 10	2020	CABG	91 829	7%	64.86	7.5 years a	n/a
Kuźma et al. 11	2017	CABG PCI	161 2994	All‐cause dementia: 22.9% Alzheimer's dementia: 10.1% Vascular dementia: 3.9% Mixed dementia: 12.2% All‐cause dementia: 13.4% Alzheimer's dementia: 7.5% Vascular dementia: 1.9% Mixed dementia: 4.5%	74.9 74.5	5.9 years c 6.0 years c	No significant difference between groups for Alzheimer's or vascular dementia; significant difference between groups for all‐cause dementia and mixed dementia.
Mutch et al. 17	2011	CABG PCI Medical management	629 1124 34,508	7.0% 4.9% 12.1%	66.9 67.3 74.1	8 years a 8 years a 8 years a	Statistically significant difference between PCI and medical management; no significant difference between CABG and medical management.
Hata et al. 12	2017	Distal limited open stenting in distal aortic arch aneurysm	24	9.5%	81.6	68 months a 2.59 years b	n/a
Harmand et al. 20	2024	Cardiac surgery	48	7.1% of non‐MCI patients at baseline developed MCI, and 10% of MCI patients at baseline developed dementia	73.42	1 year a	n/a
Lingehall et al. 13	2017	Cardiac surgery with CBP	114	26.3%	76.5	5 years a	n/a
Evered et al. 2	2016	CABG	326	30.8% a (prevalence)	75.2	7.5 years a	n/a
Pan et al. 14	2022	CABG	35,193	3.3%	68.1	5.3 years c 8.2 years a	n/a
Giang et al. 15	2021	CABG Control population	111,335 222 396	5.88 per 1000 person‐years 5.37 per 1000 person‐years	66.3 66.3	10.6 years c 24 years a 10.6 years c 24 years a	No difference in the risk for all‐cause dementia between CABG patients and control subjects.
Lee et al. 16	2005	CABG PTCA	5216 3954	Alzheimer's dementia: 1.5% Alzheimer's dementia: 1.0%	67.2 66.6	4.88 years b 4.76 years b	No statistically significant difference between groups.
Knopman et al. 21	2005	CABG	557 dementia cohort 557 control cohort	4.3% of dementia cohort underwent CABG 5.0% of control cohort underwent CABG	74 73	5.5 years c 3.9 years c	No statistically significant difference between dementia and control cohorts.

Abbreviations: CABG, coronary artery bypass grafting; CAD, coronary artery disease; CBP, cardiopulmonary bypass; PCI, percutaneous coronary intervention; PTCA, percutaneous transluminal coronary angioplasty.

^a Maximal duration.

^b Mean duration.

^c Median duration.

In large cohorts, dementia incidence after CABG ranged from 3.3% over 5.3 years ¹⁴ to 7% over 7.5 years, ¹⁰ while population‐based analyses estimate rates of approximately 5.88 cases per 1000 person‐years, comparable to non‐surgical controls. ¹⁵ Other studies reported higher cumulative risks, including 10.5% over 10 years after CABG ⁸ and 26.3% over 5 years following cardiac surgery with cardiopulmonary bypass (CPB). ¹³ Procedure‐specific analyses demonstrated an incidence of 2.7% over 20 years after heart transplantation ⁹ and 9.5% over 5.7 years after distal aortic arch aneurysm stenting. ¹² Of note, the meta‐analysis by Greaves et al. ¹⁰ synthesized many of the individual studies discussed in this review and therefore represents a higher level of evidence. Findings from individual studies should be interpreted within this broader context to avoid over‐weighing duplicated data.

4.1. Variations in inclusion criteria

Pan et al., ¹⁴ Mutch et al., ¹⁷ Giang et al., ¹⁵ and Burker et al. ⁹ included relatively younger populations (mean age 37.2 to 68.1 years), which may partly explain their lower reported dementia incidence (2.7% to 7.0%). In contrast, Lingehall et al. ¹³ restricted their study population to older individuals (>70 years of age) and subsequently reported a dementia incidence of 26.3% over 5 years. ¹³ Demographic differences also limit generalizability, as some studies drew from single‐center experiences, while others focused on Medicare ⁸ or Veteran Affairs populations. ¹⁹

4.2. Type of cardiac surgery

Nine of the 13 included studies specifically examined isolated CABG, while the remaining studies examined other procedures, such as distal aortic aneurysm repair, ¹² heart transplantation, ⁹ or cardiac surgery more broadly. ¹³ , ²⁰ Several studies, including those by Evered et al., ² Pan et al., ¹⁴ and Giang et al., ¹⁵ limited their cohort to patients undergoing elective, first‐time CABG. Notably, there is a paucity of research examining dementia incidence following valvular or combined cardiac surgical procedures.

4.3. Cohort size and follow‐up duration

Studies with smaller sample sizes, such as those by Burker et al. ⁹ (n = 24) and Harmand et al. ²⁰ (n = 97), reported relatively low dementia incidences (2.7% over 20.3 years and 2.1% over 1 year, respectively). In contrast, larger cohort studies generally reported higher dementia incidences (ranging from 7% to 22.3% over 5.9 to 10 years), ⁸ , ¹⁰ , ¹¹ although this pattern was not consistent across all studies. ¹⁴ , ¹⁵ , ¹⁶ Because most published studies did not standardize dementia incidence using person‐years of follow‐up, shorter follow‐up periods likely underestimated dementia incidence by missing cases with delayed onset.

4.4. Diagnostic criteria

Some studies employed limited or less rigorous diagnostic methods. For instance, Hata et al. ¹² relied exclusively on MMSE scores, and Giang et al. ¹⁵ identified dementia cases solely using ICD codes. In contrast, other studies utilized more comprehensive diagnostic approaches, including clinical interviews, detailed neuropsychological testing across multiple cognitive domains, ¹¹ , ²⁰ , ²¹ and MRI to differentiate dementia subtypes. ¹¹ For studies relying on administrative or electronic record data, employing validated dementia algorithms would improve the accuracy and reliability of dementia ascertainment.

The wide variability observed in study design elements and reported dementia incidence highlights the need for additional research examining dementia risk after cardiac surgery that prioritizes large, multicenter cohorts with extended follow‐up periods to accurately evaluate long‐term cognitive outcomes across diverse patient populations and surgical procedures. Additionally, standardized diagnostic criteria, consistent methodological approaches, and dementia incidence reporting are essential to enhance the comparability and generalizability of findings across studies.

5. PATHOPHYSIOLOGIC MECHANISMS

Ongoing research seeks to clarify the association between cardiac surgery and dementia. The underlying mechanisms are complex and multifactorial; some have historically been linked to cardiopulmonary bypass (CPB). However, comparative studies of on‐ versus off‐pump techniques have yielded mixed results, and some have shown no consistent difference in postoperative cognitive outcomes such as POCD. ²² , ²³

5.1. Cerebral microembolism

During cardiac surgery, especially procedures involving CPB, multiple factors can lead to the formation and systemic dissemination of embolic materials. These emboli typically include dislodged fragments of atherosclerotic plaques, gaseous microemboli (GME), and microthrombi. ⁴ , ²⁴ , ²⁵

Surgical manipulation of the aorta and major vessels – such as cannulation for CPB, placement and removal of aortic cross‐clamps, and partial occlusion clamps – can disrupt existing atherosclerotic lesions. ³ This mechanical disruption may release fragments (atheromatous debris) into the bloodstream, potentially embolizing to the cerebral circulation. Another embolic risk arises from GME, which can form within the CPB circuit due to inadvertent air entrainment or the release of dissolved gases from blood. ³ Additionally, microthrombus formation occurs through several mechanisms. Blood contact with artificial CPB surfaces activates the coagulation cascade and promotes platelet aggregation. ²⁶ Surgical trauma and systemic inflammation further amplify this prothrombotic state, resulting in the formation of small fibrin‐platelet clots. ²⁶ Intraoperative hypotension may also contribute by reducing cerebral prefusion and slowing blood flow in small vessels, promoting stasis, a critical factor in thrombus formation. ²⁷ Furthermore, hypoperfusion and tissue hypoxia can cause endothelial injury, leading to increased expression of procoagulant factors and suppression of antithrombotic mechanisms, thereby favoring local thrombosis. ²⁸

These embolic materials can enter the cerebral circulation, causing localized ischemia and subsequent cerebral microinfarctions. Although individual microinfarctions are often clinically silent, their cumulative burden is thought to contribute to long‐term cognitive decline, especially among older or vulnerable patient populations. This embolic phenomenon is considered a key potential mechanism underlying POCD. However, the strength of this association remains uncertain. While some studies have identified correlations between cerebral emboli and postoperative neurocognitive deficits, other large‐scale investigations and systematic reviews have not consistently demonstrated statistically significant relationships. ²⁹ , ³⁰ , ³¹ , ³²

To reduce embolic burden during cardiac surgery, a multifaceted approach should be adopted. Incorporating arterial line filters into the CPB circuit can effectively capture particulate and gaseous emboli before they enter the cerebral circulation. ⁴ , ³³ Meticulous de‐airing techniques ³⁴ and carbon dioxide field flooding ³⁵ are essential to remove residual intracardiac air and prevent air embolism. Minimizing aortic manipulation and employing epiaortic ultrasound to identify and avoid heavily calcified regions can significantly decrease the release of atheromatous debris. ³⁶ Additionally, in selected cases, off‐pump CABG may be considered to avoid CPB use entirely, thereby reducing exposure to embolic sources associated with extracorporeal circulation.

5.2. Inflammatory response

Cardiac surgery and CPB can trigger a systemic inflammatory response due to tissue injury, organ ischemia, and mechanical trauma from the bypass circuit. ³⁷ This involves activation of complement pathways, leukocytes, platelets, and coagulation factors, resulting in a pronounced systemic inflammatory response syndrome characterized by elevated cytokine levels. Such inflammation can lead to neuroinflammation and compromise the integrity of the blood‐brain barrier (BBB). ³⁸ Ultimately, this inflammatory cascade contributes to microglial activation, synaptic injury, neuronal dysfunction and death, and impaired neurogenesis, ³⁹ , ⁴⁰ , ⁴¹ all of which represent key pathophysiological mechanisms underlying both acute and chronic cognitive impairment. ⁴²

Excessive microglia activation, such as that occurring after surgery or in the context of aging or neurodegenerative susceptibility, can lead microglia to damage surrounding healthy neural tissue. ⁴³ Injured neurons subsequently release signals that further stimulate microglia activation, creating a self‐sustaining cycle of neuroinflammation. This persistent inflammatory environment contributes to ongoing neuronal injury long‐term cognitive decline ⁴² via accelerated neuropathological changes associated with dementia, particularly Alzheimer's disease. ⁴⁴ Indeed, repeated or chronic inflammatory insults may serve as iatrogenic risk factors for the development of neurodegenerative disorders.

5.3. Hypoxia and ischemia

Rios‐Monterrosa and Sun ⁴⁵ described the multiple factors contributing to cerebral ischemia risk during cardiac surgery. Patient‐related factors include chronic hypertension, which impairs cerebral autoregulation and heightens vulnerability to hypotensive episodes, and atherosclerosis, which promotes vascular occlusion and increases susceptibility to hypoperfusion. Procedure‐related factors also play a critical role. CPB induces a prothrombotic state, generates non‐pulsatile blood flow that can impair cerebral microcirculation, and disrupts CBF autoregulation, collectively compromising cerebral perfusion.

Managing intraoperative hypotension is essential for maintaining adequate cerebral perfusion during the perioperative period. Additionally, the “washout” hypothesis suggests that maintaining higher mean arterial pressure (MAP) may improve emboli clearance, thereby reducing the risk of ischemic stroke. ⁴⁶

Several studies have investigated intraoperative hypotension during cardiac surgery; however, clinical studies and randomized trials have yielded conflicting findings regarding the impact of different MAP levels on neurological outcomes. ⁴⁷ , ⁴⁸ , ⁴⁹ , ⁵⁰ , ⁵¹ While some studies identified a clear association between prolonged or severe hypotension and increased stroke risk, others failed to identify statistically significant relationships. ⁴⁷ , ⁴⁹ , ⁵⁰ , ⁵¹

Two large studies involving over 12,000 patients undergoing cardiac surgery with CPB evaluated the relationship between MAP and stroke risk. ²⁷ , ⁵² The first study demonstrated that persistent MAP below 65 mmHg for more than 10 min was associated with a dose‐dependent increase in stroke risk. ²⁷ Similarly, the second study found that a longer duration of MAP below 65 mmHg was significantly associated with higher risk of stroke and other adverse outcomes. ⁵²

Maintaining adequate cerebral perfusion during cardiac surgery is essential, given the numerous patient‐ and procedure‐related factors involved. Although managing intraoperative hypotension is central to this goal, historical clinical studies have presented conflicting evidence on optimal MAP targets. However, recent large‐scale studies have provided clearer evidence, demonstrating that sustained periods of MAP below certain thresholds (e.g., 65 mmHg or relative decreases of ≥10% for more than 10 min) significantly increase stroke risk. ²⁷ , ⁵² These findings highlight the importance of ongoing efforts to refine perioperative blood pressure management guidelines during CPB, as well as further research on effective cerebral protection strategies in this population.

Cerebral oximetry, a non‐invasive monitoring technique utilizing near‐infrared spectroscopy, provides real‐time measurements of regional cerebral oxygen saturation (rSO₂). This technology has shown considerable promise in predicting postoperative cognitive complications. Systematic reviews and meta‐analyses by Zorrilla‐Vaca et al. ⁵³ and Tian et al. ⁵⁴ have demonstrated that intraoperative reductions in rSO₂ – either below absolute thresholds (e.g., 50% to 60%) or significant declines from baseline (20% to 25%) – are strongly associated with increased risks of POCD and POD. Specifically, Tian et al. reported that episodes of cerebral desaturation, defined as a drop below 70% to 90% of baseline or below 50% to 60% absolute rSO₂ – more than double the odds of POD. Importantly, their findings also suggest that interventions guided by intraoperative rSO₂ monitoring can significantly reduce the incidence of POCD. Thus, cerebral oximetry may serve not only as a monitoring tool but also as an early warning system for cognitive complications following surgery.

6. PREDICTING CARDIAC SURGERY‐ASSOCIATED DEMENTIA

6.1. Preoperative cognitive impairment

Harmand et al. ²⁰ investigated the prognostic impact of mild cognitive impairment (MCI) prior to cardiac surgery on clinical outcomes at 12 months in patients aged 65 years and older. In their study, MCI was diagnosed based on patient history and neuropsychological assessment, following the criteria established by Winblad et al.: ⁵⁵ cognitive performance below normal but not meeting criteria for dementia, evidence of cognitive decline, and preserved or minimally impaired ADLs. They found that patients with preoperative MCI had a 10% risk of progressing to dementia after cardiac surgery, compared to 0% among patients without MCI. ²⁰

In another study, multivariable logistic regression analysis demonstrated that lower preoperative MMSE scores were significantly associated with dementia after cardiac surgery (odds ratio [OR] 0.68, 95% confidence interval [CI] 0.54 to 0.84.) ¹³ The authors hypothesized that cardiac surgery patients often have multiple dementia risk factors, such as advanced age, hypertension, diabetes, or prior stroke. Surgery may exacerbate these underlying vulnerabilities, accelerating cognitive decline, particularly in patients already experiencing MCI.

6.2. Biomarkers

Identifying specific biomarkers predictive of POCD could improve preoperative risk stratification and inform therapeutic strategies. Several studies have explored biomarkers associated with dementia as potential predictors of POCD following cardiac surgery. Palotas et al. ⁵⁶ conducted a prospective cohort study evaluating cognitive function and cerebrospinal fluid (CSF) biomarkers associated with POCD – specifically amyloid beta peptide, tau, and S100β – in patients undergoing off‐pump CABG. They found that POCD correlated with changes in biomarkers typically associated with Alzheimer's disease, suggesting a shared underlying pathophysiological mechanism in these conditions. However, this study had notable limitations, including substantial loss to follow‐up, with only 14 of 30 patients completing cognitive assessments at all three time points (preoperatively and 1 week and 6 months postoperatively). Additionally, the study lacked appropriate control groups, such as non‐surgical coronary artery disease (CAD) patients or other surgical cohorts. Although cognitive decline was assessed up to 6 months postoperatively, the study did not specifically evaluate participants for dementia diagnoses. Furthermore, reliance on CSF biomarkers limits clinical practicality due to the invasive nature of sample collection.

In another study, researchers examined genetic biomarkers, specifically the PlA2 allele of the platelet glycoprotein‐IIIa gene – associated with increased coagulability – and the apolipoprotein E ε4 allele to dementia risk and severe atherosclerosis. ⁵⁷ They found that the PlA2 allele was more common among patients experiencing POCD 1 year after cardiac surgery, whereas the apolipoprotein E ε4 allele showed no significant association. Similar to the previous study, this investigation did not specifically assess dementia outcomes.

6.3. Demographic factors and comorbidities

Giang et al. ¹⁵ conducted a population‐based cohort study in Sweden, specifically examining risk factors for dementia among 111,335 patients who had undergone CABG between 1992 and 2015, compared to 222,396 matched controls. They identified older age (adjusted hazard ratio [aHR] 1.12, 95% CI 1.12 to 1.13) and divorced status (aHR 1.16, 95% CI 1.06 to 1.27) as significant risk factors for dementia. Comorbidities significantly associated with increased dementia risk included heart failure (aHR 1.20, 95% CI 1.11 to 1.29), hypertension (aHR 1.15, 95% CI 1.08 to 1.21), diabetes (aHR 1.62, 95% CI 1.52 to 1.73), renal failure (aHR 1.38, 95% CI 1.52 to 1.73), history of stroke (aHR 1.47, 95% CI 1.33 to 1.62), atrial fibrillation (aHR 1.14, 95% CI 1.04 to 1.25), and depression (aHR 1.58, 95% CI 1.29 to 1.93). Higher income levels were associated with lower dementia risk compared to the lowest income quintile, with adjusted HRs for quintiles 2 to 5 of 0.97 (95% CI 0.91 to 1.04), 0.92 (95% CI 0.86 to 0.99), 0.91 (95% CI 0.84 to 0.99), and 0.82 (95% CI 0.74 to 0.90), respectively. Similarly, advanced age, ⁸ , ¹¹ , ¹³ , ¹⁵ , ¹⁹ pre‐existing cognitive impairment, ² , ¹⁰ , ¹³ and cardiovascular comorbidities (i.e., hypertension, diabetes, atrial fibrillation, prior stroke) have consistently been identified as dementia risk factors following cardiac surgery across several studies. ⁸ , ¹¹ , ¹⁵ , ¹⁷ Conditions not significantly associated with dementia risk included education level, sex, myocardial infarction, peripheral vascular disease, chronic respiratory disease, history of cancer, and hyperlipidemia. ¹⁵

6.4. Medications

A large nationwide population‐based study (n = 35,193) conducted in Sweden from 2006 to 2017 found that ongoing statin use, irrespective of dose, significantly reduced dementia risk among patients undergoing first‐time CABG (aHR 0.74, 95% CI 0.65 to 0.85). ¹⁴ This study utilized data from five national registries. Similarly, Lingehall et al. ¹³ reported in a single‐center cohort study (n = 153) based in Sweden in 2017 that anti‐lipemic agents were protective against dementia in multivariate analysis (OR 0.31, 95% CI 0.11 to 0.90). In the same study, diuretic use was associated with increased dementia risk after cardiac surgery in univariable analysis (OR 2.55, 95% CI 1.09 to 5.99); however, this association was not evaluated in multivariable analysis.

6.5. Procedure type

Much of the existing literature evaluating dementia risk following cardiac surgery has primarily focused on CABG. There are relatively few studies evaluating other cardiac surgical procedures.

6.5.1. CABG

Several studies have reported no statistically significant increase in the risk of all‐cause dementia ¹³ , ²¹ or Alzheimer's disease ²¹ following CABG. On the other hand, the US‐based Cardiovascular Health Study, ¹¹ which enrolled participants between 1991 and 1994, found that a history of CABG was associated with a two‐fold increase in the risk of all‐cause dementia (adjusted HR 1.93 [95% CI 1.36 to 2.74]) compared to those without CABG. However, findings may not be generalizable to more diverse populations as the cohort was limited to White and African American participants.

Studies comparing the risk of dementia after CABG versus percutaneous coronary intervention (PCI) have yielded mixed results (Table 1). One retrospective cohort study found no significant difference in dementia incidence 5 years after revascularization (difference, 0.9%, 95% CI 2.0 to 3.8). ⁸ Conversely, Kuźma et al. ¹¹ reported a trend toward higher risk of all‐cause dementia following CABG compared to PCI (HR 1.47, 95% CI, 0.98 to 2.98). Another study found PCI to be associated with a 1.71‐fold higher risk of Alzheimer's disease compared to CABG (95% CI 1.02 to 2.87). ¹⁶ Mutch et al. ¹⁷ observed fewer new dementia diagnoses after PCI (4.9%) compared to CABG (7.0%) and medical management (12.1%). PCI demonstrated a significant advantage over medical management (HR = 0.65, 95% Cl 0.46 to 0.84), whereas CABG did not differ significantly from medical management (HR = 0.90, 95% Cl 0.72 to 1.12). Taken together, these findings suggest that CABG carries a similar or potentially higher risk of postoperative dementia compared to PCI.

6.5.2. Valvular and combined procedures

Among the 13 studies reviewed, Lingehall et al. ¹³ specifically evaluated dementia risk associated with intraoperative factors (e.g., operative duration, CPB time, and blood loss) in patients undergoing valvular and combined CABG and valvular procedures. In this single‐center study, none of the intraoperative variables significantly increased the risk of dementia within 5 years after cardiac surgery. Given the limited and inconclusive current evidence, further research is necessary to clarify the impact of intraoperative factors on dementia risk across a broader scope of cardiac surgery.

6.5.3. Heart transplant

Burker et al. ⁹ evaluated postoperative cognitive function in heart transplant recipients at a national transplant center in Norway over a 16‐year period following transplantation. Cognitive function was evaluated using a neuropsychological test battery and adapted criteria for MCI. In their 2017 study, MCI was defined by the following criteria: (1) no pre‐existing dementia diagnosis, (2) self‐reported subjective cognitive difficulties, and (3) cognitive performance at least 1.5 SD below normative mean in at least one cognitive domain. Among 37 participants (mean age 57.5 years, 18.9% female), 38.9% demonstrated impaired cognitive performance, particularly in processing speed, executive function, memory, and language. Overall, 30.1% met criteria for postoperative MCI, and one subject (2.7%) was diagnosed with dementia.

6.5.4. Endovascular procedures for aortic arch aneurysm

A 2017 Japanese study introduced a novel surgical approach for distal aortic arch aneurysm (DAAA) in octogenarians that involved distal limited open stenting via median sternotomy and CPB. The study included 27 patients (mean age 81.6, 20.8% female; mean follow‐up 31.1 months), of whom three were excluded due to pre‐existing dementia. ¹² All patients had large DAAAs (mean maximum diameter 60.6 ± 5.5 mm) with mural thrombi and atherosclerotic plaques identified on imaging. At postoperative follow‐up, no patients experienced delirium or stroke, but 9.5% demonstrated new‐onset dementia based on neuropsychological testing. However, the study lacked detailed inclusion criteria, comprehensive baseline patient characteristics, and consideration of potential confounding factors. Additionally, concomitant CABG was performed in three cases (12.5%), further confounding the observed relationship.

6.6. Postoperative neurocognitive disorders

POCD and POD are both postoperative neurocognitive disorders that share certain risk factors and may precipitate dementia. Despite the overlap, they are clinically and pathophysiologically distinct, differing in onset, duration, and diagnostic criteria.

POCD is characterized by a measurable decline in cognitive function relative to a patient's preoperative baseline following exposure to surgery and anesthesia, typically detected through standardized neuropsychological testing before and after surgery. ² , ⁵⁸ , ⁵⁹ A decline of one to two standard deviations in test scores across domains such as memory, learning, attention, and, in more severe cases, alterations in mood or personality are used as a diagnostic threshold. ² , ⁵⁸ However, there is no universally accepted definition of POCD, and several other diagnostic approaches have also been used in the literature. These include the “20/20” method (≥20% decline in a participant's postoperative test score compared with their preoperative test score, on at least 20% of the tests conducted) ⁶⁰ , reliable change index (RCI)‐based methods that account for practice effects and test‐retest variability, ⁶¹ the 1 SD method (≥1 SD decline on at least 20% of tests), ⁶² and cut‐off methods that use population norms to define cognitive impairment on specific tests. ⁶³ This heterogeneity in definitions contributes to variability in reported incidence and complicates comparisons across studies. POCD usually becomes apparent later in the postoperative course – often toward the end of the first week – and may persist for months or, in some cases, more than a year. ⁶⁴ The reported incidence of POCD following cardiac surgery demonstrates considerable variability across studies, with estimates ranging from 9% to 54% of patients when assessed prior to hospital discharge, ² , ⁵⁸ , ⁵⁹ , ⁶⁴ and approximately 15% to 30% of patients with pre‐discharge POCD continue to exhibit cognitive deficits up to 3 months postoperatively. ² , ⁵⁸ , ⁶⁵

In contrast, according to the DSM‐5, POD is characterized by an acute onset or fluctuating disturbances in attention and awareness that cannot be better explained by pre‐existing or evolving neurocognitive disorders or severely reduced levels of consciousness (e.g., coma). ⁵ It typically manifests within the first 1 to 3 days after surgery. ⁶⁴ , ⁶⁵ POD can be triggered by surgical or anesthesia factors, pain, sedative medications, or the patient's overall medical conditions.

Multiple risk factors are common to both POCD and POD. Advanced age remains the strongest factor, likely reflecting age‐related declines in neuroplasticity, cerebral perfusion, and autoregulation. ⁵⁸ , ⁶⁴ , ⁶⁶ , ⁶⁷ Pre‐existing cognitive impairment, frailty, and multimorbidity further increase susceptibility. For POCD specifically, procedure type has been implicated, with cardiac surgeries carrying a higher incidence compared to non‐cardiac surgeries. ⁶⁴ , ⁶⁵ Longer surgery and anesthesia durations – possibly due to prolonged physiological stress and inflammation – as well as lighter depths of anesthesia and certain anesthetic agents (e.g., hypnotics and sedatives) are associated with an elevated risk of POCD. ⁶⁴ , ⁶⁵ Neuroinflammation is an increasingly recognized mechanism in both POCD and POD, as perioperative inflammatory surges can disrupt the BBB and contribute to cognitive decline. ⁶⁴ , ⁶⁷ , ⁶⁸ Moreover, additional perioperative factors such as hypoxia have also been implicated in increasing the risk of POCD and POD. ⁶⁴ , ⁶⁹ A systematic review and meta‐analysis of 97 studies including 60,479 patients undergoing CABG identified several risk factors for POD and cognitive decline, with substantial overlap. ⁷⁰ Delirium risk factors included preoperative cognitive impairment, depression, and prior stroke, as well as longer intubation time and prolonged intensive care unit (ICU) stay. Risk factors for cognitive decline included older age and preoperative depression, longer intubation time, and POD and longer ICU stay. In this meta‐analysis, POD after CABG was associated with an approximately 5‐fold increased risk of acute cognitive decline at 1 month. ⁷⁰ However, evidence linking delirium to long‐term cognitive decline after cardiac surgery remains limited.

The consequences of POCD and POD extend well beyond the immediate perioperative period. In the short term, POCD within the first 30 days after cardiac surgery correlates with longer hospital stays and complication rates. ⁵⁹ Some studies indicate that POCD identified beyond 30 days may compromise functional recovery. ⁵⁹ Furthermore, certain functional assessment tools may lack the sensitivity required to detect subtle but clinically relevant deficits in executive function, potentially underestimating the real impact of POCD. Over the long term, POCD has also been associated with increased mortality. ⁵⁹ Meanwhile, for POD, short‐term consequences frequently include falls, aspiration pneumonia, emotional distress, and an elevated risk of short‐term mortality. ⁶⁸ POD may also predispose individuals to depression, anxiety, and post‐traumatic stress disorder, as well as an overall reduction in quality of life. ⁶⁸

Although the link between perioperative cognitive disorders and dementia is complex, POCD and POD may differ in their roles in dementia risk. Persistent POCD has prompted concerns that it could represent an early stage of neurodegeneration; however, a long‐term follow‐up by Steinmetz et al. ⁷¹ over 11 years found no significant association between POCD and dementia. By contrast, a growing body of literature demonstrates a strong association between delirium and long‐term cognitive decline. Leighton et al. reported a 5‐year cumulative incidence of dementia of 31% among 12,949 patients with incident delirium. ⁷² Using the Vantaa 85+ cohort, Davis et al. ⁷³ found delirium was associated with increased odds of dementia (OR 8.7; 95% CI 2.1 to 35). These studies, however, were conducted in general populations rather than cardiac surgery cohorts. Evidence in cardiac surgical populations is more limited, but Lingehall et al. ¹³ observed that 26.3% of patients without baseline dementia developed dementia within 5 years of cardiac surgery with CPB; among those who developed dementia, 87% had experienced POD (OR 7.57; 95% CI 2.15 to 26.65). Taken together, POD may reflect underlying vulnerability and may also contribute to accelerated cognitive decline.

7. FUTURE DIRECTIONS

Current literature on dementia after cardiac surgery is limited by substantial heterogeneity in study design, diagnostic criteria, procedural types, patient populations, and follow‐up duration, along with non‐standardized incidence reporting. These variations limit interpretability and generalizability. The advancement of this field necessitates coordinated efforts to address these fundamental inconsistencies in future research.

Standardizing diagnostic criteria for both POCD and postoperative dementia is essential. Consensus criteria – such as selecting validated cognitive tests and clear diagnostic thresholds – would increase the comparability of findings across studies, enabling more precise incidence estimates and stronger meta‐analyses. Given that advanced age is a key risk factor for dementia, studies should either use age‐standardized incidence or stratify by age across various cardiac procedures with longer follow‐up periods. Finally, because dementia often manifests long after surgery, incidence rates expressed in person‐years are preferable to simple percentages.

Beyond methodological consistency, future research should prioritize identifying modifiable risk factors and developing effective preventive strategies. These may include optimizing surgical, anesthetic, and perfusion techniques, employing advanced intraoperative neuromonitoring and personalized hemodynamic targets, and implementing cognitive prehabilitation and POD prevention protocols.

From a practical standpoint, progress could likely be achieved through a stepwise and collaborative approach that balanced scientific rigor with clinical feasibility. An initial goal could include the integration of standardized cognitive assessments into existing cardiac surgery registries and quality‐improvement programs, which would allow for large‐scale data collection while minimizing disruption to clinical workflows. In parallel, multidisciplinary specialties involving cardiac surgeons, anesthesiologists, neurologists, and geriatricians could develop consensus “minimum standards” for cognitive outcome reporting, acknowledging the diversity across the distinct cardiovascular and medical specialties. Over time, these pragmatic standards could serve as a foundation for more detailed phenotyping, longer‐term follow‐up, and targeted interventional trials aimed at modifiable perioperative risk factors. Such an incremental, cooperative strategy could offer a realistic path toward addressing key questions surrounding dementia following cardiac surgery.

8. CONCLUSIONS

The findings of this narrative review indicate that older age, baseline MCI, and the occurrence of POD are associated with an increased risk of dementia following cardiac surgery. Thorough preoperative assessments and individualized postoperative care protocols may help protect cognitive function in vulnerable patients. Further research will be necessary to identify additional modifiable risk factors and effective preventive strategies. Protecting cognitive health in cardiac surgery patients is critical, as it significantly impacts individual quality of life and may reduce healthcare expenditures related to dementia and institutionalized care.

CONFLICT OF INTEREST STATEMENT

The authors declare that they have no conflicts of interest relevant to this manuscript. Author disclosures are available in the Supporting Information.

FUNDING INFORMATION

The authors received no financial support or funding for the research, authorship, and/or publication of this manuscript.

Supporting information

Supporting Information