Abstract

Background:

Schizophrenia is a severe psychiatric disorder characterized by persistent cognitive impairment across multiple domains and is increasingly associated with elevated risk of late-life dementia. However, the nature of this association and its underlying mechanisms remain unclear.

Objective:

This mini-review synthesizes current evidence on dementia risk in schizophrenia, focusing on epidemiology, cognitive trajectories, biological mechanisms, and differential relationships with Alzheimer’s disease (AD), vascular dementia (VaD), and frontotemporal dementia (FTD).

Results:

Epidemiological studies consistently indicate a two- to threefold increased risk of dementia among individuals with schizophrenia, although estimates vary due to diagnostic and ascertainment biases. Cognitive trajectories are heterogeneous: many patients remain cognitively stable over time, while subgroups demonstrate gradual or accelerated decline associated with negative symptoms, medical comorbidities, and social factors. Current evidence does not support a uniform progression toward Alzheimer-type neurodegeneration. Biomarker, neuropathological, and neuroimaging findings suggest distinct biological profiles, with reduced cognitive reserve, neurodevelopmental vulnerability, accelerated aging processes, and vascular and metabolic burden contributing to dementia risk. Genetic overlap between schizophrenia and AD appears modest, whereas partial clinical and molecular convergence is observed with FTD. Screening tools such as MMSE and MoCA may overestimate dementia prevalence due to longstanding baseline cognitive deficits. Sex differences, late-onset psychosis, and cardiometabolic comorbidities further modify risk trajectories.

Conclusion:

Dementia risk in schizophrenia likely reflects the interaction of lifelong neurodevelopmental vulnerability with aging-related and modifiable factors rather than a disorder-specific neurodegenerative pathway. Longitudinal biomarker-informed studies and tailored diagnostic frameworks are needed to improve differentiation between chronic cognitive impairment and true neurodegeneration.

Introduction

Schizophrenia is a severe psychiatric disorder marked by chronic functional impairment and cognitive deficits constitute a core and enduring feature of the illness, affecting domains such as processing speed, attention, executive functions, memory, and social cognition, and they significantly contribute to disability across the lifespan (1–3). In older adults, these impairments become more pronounced and are increasingly recognized as a major driver of the elevated dementia risk observed in schizophrenia (4). This mini-review summarizes current evidence on dementia risk in schizophrenia encompassing Alzheimer’s disease (AD), vascular dementia (VaD), and frontotemporal dementia (FTD), with a focus on epidemiology, cognitive trajectories, and disorder-specific patterns across different dementia types. Literature searches were conducted primarily in PubMed, with a deliberate focus on longitudinal studies, meta-analyses, and review articles to capture long-term cognitive trajectories and dementia risk in schizophrenia (publications from 1994 to 2025). A more detailed description of the search strategy, including search terms, selection criteria, and study screening procedures, is provided in Supplementary material.

Schizophrenia, adulthood, and the risk of dementia

Severe mental illnesses are associated with elevated dementia risk—highest in bipolar disorder, followed by schizophrenia and major depression (5, 6). Analysis of 86 million patients revealed age- and sex-specific comorbidity patterns in schizophrenia, including increased risks of delirium, dementia, and various somatic conditions in later life (7). A meta-analysis including over 12 million individuals showed that non-affective psychotic disorders more than double dementia risk (RR ≈ 2.5) (8). Large population-based and registry studies across multiple countries report a two- to threefold increased dementia risk in schizophrenia, particularly before age 65 and among individuals diagnosed in midlife (9–11). In a Medicare cohort, dementia prevalence rose from 28% at age 66 to 70% by age 80 in schizophrenia, compared with 1.3 and 11.3% in matched controls (12). A meta-analysis including more than 200,000 dementia cases in schizophrenia populations confirmed a significantly elevated risk of dementia, though with substantial heterogeneity (13). Propensity score–matched analyses further demonstrate elevated risk across dementia subtypes, including AD (aHR ≈ 2.1), with cardiometabolic, neurological, and substance-related comorbidities substantially amplifying vulnerability (9). Stronger associations have been observed in women and in older age groups, underscoring the growing public health relevance of schizophrenia–dementia comorbidity (14, 15). Overall, large registry and longitudinal studies consistently indicate that schizophrenia is associated with increased risk of subsequent dementia, although some cohorts of very old patients show attenuated or non-significant differences compared with age-matched controls (16–22). Furthermore, the elevated prevalence of dementia diagnoses in individuals with schizophrenia may partly reflect diagnostic misclassification and ascertainment bias rather than solely true neurodegenerative processes. Long-standing baseline cognitive deficits reduced cognitive reserve, cumulative antipsychotic exposure, and histories of institutionalization may lower the threshold for meeting clinical dementia criteria during normative aging, even in the absence of primary neurodegenerative pathology (2, 23, 24). Moreover, functional impairment—an essential component of dementia diagnosis—is frequently present in schizophrenia across the lifespan, which complicates the differentiation between chronic psychiatric-related disability and progressive neurodegeneration (24). The risk of dementia in schizophrenia patients may be partially overestimated due to diagnostic and ascertainment biases, though a genuine elevated risk likely remains. Stafford et al. (25) demonstrated that in very late-onset schizophrenia (LOS), the association was substantially attenuated (HR: 2.22) after accounting for potential misdiagnosis and ascertainment bias, compared to the unadjusted estimate (HR: 4.22). Moraiti and Porfyri (26) noted that studies have significant limitations including small sample sizes and selection of chronic older patients, Lyketsos et al. (27) emphasized that dementia development in schizophrenia is “not inevitable.” However, the risk of dementia in patients with schizophrenia is significantly higher compared to those without serious mental illness, as evidenced by a large retrospective cohort study involving over 8 million participants (28).

Emerging evidence suggests that antipsychotic exposure may be associated with increased dementia risk in schizophrenia, with one study reporting a 92% higher risk in individuals under 65 (HR = 1.92) (29), and others linking specific agents (olanzapine, risperidone, clozapine) to elevated AD risk, potentially via impaired microglial amyloid-β clearance (30); however, postmortem data have not confirmed increased Alzheimer-type neuropathology in treated patients (31). Treatment-resistant schizophrenia has been associated with selective working memory deficits and reduced hippocampal volume rather than global impairment (32). Importantly, dementia risk estimates in schizophrenia may be confounded by healthcare access and diagnostic disparities, as individuals with schizophrenia spectrum disorders are less likely to receive prior mild cognitive impairment (MCI) diagnoses or anti-dementia treatment and may experience delayed recognition of dementia (33–35). Accordingly, while epidemiological data robustly support an association between schizophrenia and dementia, they do not consistently indicate a disorder-specific causal link with AD, highlighting the need for longitudinal, biomarker-informed studies and refined diagnostic frameworks. Selected studies are presented in Table 1.

Study (year)Sample size/cohortFollow-upFindingsKey conclusionHedges et al. (12)Medicare registry; age 66–80Cross-sectional registryDementia prevalence 28% → 70% vs. 1.3% → 11.3% in controlsDramatically elevated dementia risk in schizophreniaRibe et al. (10)Danish registry (>2.8 M; ≥50 years)Up to 18 yearsIRR ≈ 2.1 for dementia; strongest effect <65 yearsSchizophrenia doubles dementia risk, especially in earlier-onset casesJonas et al. (42)428 psychosis patients30 years~16-point IQ decline; deterioration began ~14y before onsetSupports dual neurodevelopmental + neurodegenerative modelHendrie et al. (17)31,588 older adults (1,635 with schizophrenia; mean age 70)10 yearsDementia prevalence 64.5% vs. 32.1% in non-schizophrenia groupMarkedly higher dementia prevalence and mortality in schizophreniaMentzel et al. (20)168,780 individuals ≥65y (2,103 with schizophrenia; New Zealand)7 years (registry)Dementia 23% vs. 25% in controlsNo significant increase in dementia prevalence in very old schizophrenia cohortTeipel et al. (22)1,686,759 dementia cases; 3,373,518 controls (Germany)Mean 6.6 yearsSchizophrenia associated with OR >2.5 for dementiaAmong strongest clinical predictors of dementia diagnosisAlmeida et al. (16)37,770 men aged 65–85 (dementia-free at baseline)Up to 17.7 yearsSub-HR 2.67 (95% CI 2.30–3.09) for psychotic disordersNearly threefold increased dementia riskKershenbaum et al. (18)28,340 adults ≥65 years in UK psychiatry services6 yearsStandardized dementia incidence rate 2.9 (1.5–4.7) for schizophreniaDementia incidence elevated vs. general populationStevenson-Hoare et al. (21)~1 M Welsh EHR + 228,937 UK Biobank participantsLongitudinal EHRHR 2.87 (Wales) and 4.46 (UKB) for schizophreniaStrong association between schizophrenia and later dementiaKørner et al. (19)12,616 late-onset + 7,712 very-late-onset schizophrenia (Denmark)3–4.6 yearsRR 2.21–3.47 vs. controls2–3 × higher dementia risk in late/very-late schizophrenia

Selected studies of dementia risk and cognitive decline in schizophrenia.

EHR, electronic health records; HR, hazard ratio; IRR, incidence rate ratio; IQ, intelligence quotient; M, million; RR, relative risk.

Cognitive trajectories in schizophrenia

Despite the increased risk of dementia, cognitive decline in schizophrenia is highly heterogeneous. Cross-sectional meta-analyses consistently show large, generalized cognitive deficits in older individuals with schizophrenia, whereas longitudinal studies and meta-analyses largely support relative short- to medium-term stability (1–5 years), even in chronically ill and functionally impaired patients (36–40). Longitudinal research identifies distinct cognitive trajectories, including stable, gradually declining, and more rapidly deteriorating courses. Approximately half of patients remain cognitively stable, while smaller subgroups show modest or accelerated decline, particularly those with severe negative symptoms, recurrent psychotic episodes, treatment resistance, or institutionalization (2, 41–43). Fluctuations across cognitive domains may occur in relation to relapse–remission patterns. A systematic review of 23 longitudinal studies found mixed evidence for late-life cognitive decline. Of the three studies applying DSM/ICD dementia criteria, two reported increased dementia prevalence. Among studies assessing specific cognitive domains, results were heterogeneous, and many did not control key risk factors such as education, vascular burden, or apolipoprotein E (APOE) genotype (44). Furthermore, systematic reviews further support the existence of multiple cognitive phenotypes within schizophrenia spectrum disorders, including cognitively preserved, intermediate, and globally impaired subgroups (45). When decline occurs, it is typically modest and domain-specific, with fluid abilities more vulnerable than crystallized functions such as verbal knowledge, which often remain stable or improve. However, long-term studies exceeding 10 years remain limited and yield inconsistent findings (46).

Overall, evidence slightly favors the presence of late cognitive decline, but it remains unclear whether this reflects schizophrenia’s pathophysiology or superimposed risk factors.

Cognitive screening in schizophrenia: validity considerations

Cognitive impairment is a core feature of schizophrenia, consistently demonstrated across multiple studies using screening tools such as the Mini-Mental State Examination (MMSE) (47) and the Montreal Cognitive Assessment (MoCA) (48). Patients with schizophrenia typically score in the range of 17.91–22.7 on the MoCA, indicating mild to moderate impairment. For instance, Amdur et al. (49) reported a mean of 22.7 ± 4.3 in 112 stable patients, while Rademeyer et al. (50) found 22.53 ± 3.91 in 30 outpatients, and Doğu et al. (51) reported 17.91 ± 3.83 in 135 patients. Across approximately 631 participants, Dedovic et al. (52) observed that 83.7% scored below 26 (53). These results contrast with population norms (27.4 ± 2.2), confirming that cognitive deficits are pervasive in schizophrenia (48, 49, 50, 51–53).

A meta-regression of 56 studies (n = 5,588) found that MMSE scores in schizophrenia declined by ~1 point every 4 years—five times faster than in the general population—primarily driven by institutionalized patients, whereas community-dwelling patients remained relatively stable (54). In 272 partially remitted outpatients, scores were 2–3 points below population norms across all ages, but gaps did not widen with age; performance was associated with education and race, with difficulties most common in memory, attention, and construction tasks (55). Middle-aged and older outpatients (n = 161) scored significantly lower than controls, with 23% scoring ≤24 on the MMSE. Impairments were most prominent in delayed recall and attention, and poorer scores were associated with lower education, structured living, unmarried status, and severe negative symptoms (56).

While MMSE identifies clinically relevant deficits, it may not capture subtle cognitive impairments that MoCA can detect. In small clinical samples (n = 30), MoCA demonstrated moderate sensitivity (41.7%) and specificity (66.7%) relative to MMSE (57). In a larger outpatient cohort (n = 64), MoCA effectively detected both mild and severe cognitive impairments (AUC ≈ 0.81–0.82), also explaining additional variance in functional outcomes (58). MoCA performance is also correlated with education, illness severity, and negative symptoms, and predicted length of hospitalization (59). Long-term follow-up studies following first-episode psychosis further confirmed MoCA’s concurrent validity with the MATRICS Consensus Cognitive Battery (MCCB), suggesting a < 25 cut-off for impairment in chronic psychotic populations (60).

We were unable to identify data reporting MoCA or specific MMSE scores, nor longitudinal changes in these measures, in patients with a prior diagnosis of schizophrenia who were subsequently diagnosed with dementia. In 73 hospitalized older patients with schizophrenia, 66% were classified as “demented” using the MMSE, indicating a high prevalence of cognitive impairment (61). In screening assessments, the usefulness of the Clock Drawing Test (CDT) has also been highlighted (62, 63), as well as the INECO Frontal Screening (IFS) (64).

Taken together, these findings emphasize that baseline cognitive deficits in schizophrenia can confound the interpretation of MMSE and MoCA scores, potentially leading to over- or underestimation of dementia risk. Therefore, clinicians should consider the limitations of these screening tools, as well as individual patient factors such as education, illness severity, and living environment, when assessing cognitive decline in schizophrenia.

Sex differences and late-onset schizophrenia

Evidence consistently demonstrates sex differences in schizophrenia in age at onset, premorbid functioning, symptom profiles, and substance use. Men typically show earlier onset, poorer premorbid functioning, greater negative symptom severity, lower affective symptoms, and higher rates of alcohol or substance misuse, whereas findings on broader psychopathology, neurocognition, social cognition, and personal resources remain inconsistent (65). These sex- and age-dependent differences extend to dementia risk. In general schizophrenia populations, men show modestly elevated risk (incidence rate ratio 2.38; 95% CI 2.13–2.66) and higher dementia-related mortality (relative risk 5.19 vs. 2.40 for females) (10, 66). However, women with LOS (after age 40) exhibit the highest dementia risk, being 3–4 times more likely to develop dementia than earlier-onset cases, with age-specific patterns showing the greatest relative risk under 65 and peak absolute risk after 65 (10, 26). In women, psychotic symptoms fluctuate with hormonal changes: exacerbations occur during low-estrogen phases of the menstrual cycle, postpartum, or following estrogen-lowering treatments, and menopause is associated with increased hallucinations, delusions, reduced cognitive functions, and decreased antipsychotic efficacy (67, 68). Perimenopause represents a sensitive neurobiological period during which declining estrogen signaling disrupts cerebral bioenergetic regulation, resulting in a hypometabolic brain state that may heighten vulnerability to neurodegenerative processes later in life (69, 70).

Late-onset schizophrenia (LOS; onset >45 years) appears to retain a genetic vulnerability similar to earlier-onset forms, while very-late-onset schizophrenia-like psychosis (VLOSLP; onset >60 years) is increasingly conceptualized as a heterogeneous and possibly multifactorial condition (71, 72). VLOSLP has been associated with elevated mortality—largely due to physical comorbidities and accidents—and with a higher subsequent risk of dementia compared to non-psychotic controls (73, 74). VLOSLP is more common in women and is characterized predominantly by positive symptoms, lower genetic loading, and higher mortality largely linked to physical comorbidity and accidents (2). Some longitudinal data suggest that a subset of late-onset cases later develop dementia, most commonly AD, supporting the hypothesis that in certain individuals VLOSLP may represent a prodromal stage of neurodegeneration (25, 75, 76). However, this interpretation remains debated. When age is controlled for, very-late-onset psychosis patients do not differ cognitively from earlier-onset groups, despite exhibiting more pronounced positive symptoms (77), and dementia rates are not higher than in aged early-onset schizophrenia populations (74). Cognitively, LOS is most consistently characterized by executive and visuospatial impairments, with some evidence of decline emerging around age 65, although existing data are limited by small and frequently institutionalized samples (23). Moreover, late-life schizophrenia presents a cognitive profile distinct from both AD and late-life depression, marked by greater impairment in learning than recall (78). Together, these findings suggest that while late-onset psychosis may, in some cases, signal evolving neurodegeneration, it does not uniformly represent a primary dementing process and should not be automatically equated with it.

Compared with MCI, schizophrenia patients may show relatively preserved recall and naming but similar deficits in attention, verbal fluency, and visuospatial functioning, further complicating differential diagnosis (79). Collectively, these findings suggest that cognitive decline in schizophrenia reflects heterogeneous trajectories shaped by illness timing, baseline cognitive reserve, and aging-related processes, rather than a uniform progression toward neurodegeneration.

Although the mechanisms underlying the elevated dementia prevalence in schizophrenia remain unclear, proposed explanations include a high burden of medical comorbidities, intrinsic cognitive and brain dysfunction, partially shared etiological pathways with dementia, and the impact of antipsychotic treatment and adverse health behaviors (13). The high prevalence of somatic comorbidities [e.g., metabolic syndrome in 39% of patients >50 years further contributes to accelerated cognitive aging and dementia risk (80)]. Schizophrenia is marked by high rates of cardiovascular and metabolic comorbidities, social isolation, smoking, and physical inactivity, all known dementia risk factors (81–84). A systematic review and meta-analysis of 27 studies (n = 10,174) found that metabolic syndrome, diabetes, and hypertension are significantly associated with greater global cognitive impairment in schizophrenia, particularly affecting executive functions, memory, and attention, whereas obesity and insulin resistance showed weaker or nonsignificant effects (85). Consistently, comorbid metabolic syndrome and diabetes have been linked to more severe cognitive deficits (86). Shared mechanisms may involve gut–brain axis dysregulation, while the high prevalence of obesity in schizophrenia likely reflects multifactorial influences, including antipsychotic treatment, lifestyle, and genetic and psychosocial factors (87).

Multiple studies consistently demonstrate that men with schizophrenia experience significantly higher cardiovascular mortality than women, with cardiovascular disease representing the leading cause of death and associated with markedly elevated standardized mortality ratios (88, 89). This excess male mortality introduces the possibility of survival bias in sex-specific dementia estimates. As highlighted in broader epidemiological research, higher premature mortality among men may artificially inflate the observed prevalence of dementia in women, as fewer men survive to the ages at which dementia typically manifests (90). Sex-related differences in selective survival may therefore partially account for the reported higher incidence of dementia among women (91). Copeli et al. (92) acknowledge that survival bias cannot be excluded in older adults with schizophrenia, but empirical quantification of its impact remains lacking. Within schizophrenia populations, cognitive impairment severity significantly impacts outcomes, with severely impaired patients experiencing more hospitalizations due to relapse and lower quality of life compared to those with mild impairment (93).

Schizophrenia, Alzheimer’s disease, and shared vulnerability: genetic and familial perspectives

In a large cohort of 486,297 UK Biobank participants, individuals with bipolar disorder (BD) and major depressive disorder (MDD) showed significantly increased AD risk (HR = 2.37 and 1.63, respectively), with schizophrenia demonstrating a borderline association, collectively suggesting that mental disorders may act as independent risk factors for AD through partially distinct pathogenic mechanisms (94). Genetic studies indicate a modest but positive association between schizophrenia and AD, alongside shared links of both conditions with lower general cognitive ability (95). A Mendelian randomization analysis based on large genome-wide association study (GWAS) demonstrated that genetic liability to schizophrenia is associated with an increased risk of all-cause dementia, AD, and VaD (96).

Twin studies consistently estimate substantially higher heritability for AD (71%) than GWAS, suggesting that a large proportion of genetic risk remains unexplained by currently identified common variants (97). Schizophrenia is a highly heritable disorder, although estimates vary substantially depending on study design and methodology. Large twin and population-based studies consistently report high heritability, typically ranging from ~70% to over 80%, including estimates of 79% in a nationwide twin cohort using methods accounting for censoring (98), 83% in a Finnish twin sample (99), 82–85% in the Maudsley Twin Register (100), ~80% in comprehensive review (3), and 0.67 (95% CI 0.64–0.71) in a Danish population cohort of over 2.6 million individuals (101). Genome-wide association studies similarly support substantial genetic liability, despite evidence for “missing heritability” not captured by common variants (102). In contrast, family-based designs yield lower estimates, with heritability of 31% in nuclear families and 44% in extended families (103), and analyses based on raw concordance data suggest more modest corrected estimates (~23%), highlighting the influence of shared environment and methodological assumptions (104). Meta-analytic evidence further supports substantial genetic contribution to schizophrenia liability and its overlap with heritable cognitive traits across large samples (>800,000 individuals) (105). Family and population studies robustly demonstrate high heritability of psychotic disorders, with risk increasing sharply with genetic proximity to affected relatives (106–108). Studies examining family history further show that familial loading for psychosis is associated with slightly earlier illness onset and greater severity of negative symptoms, with sex-dependent interactions supporting a diathesis–stress model rather than a deterministic genetic mechanism (65, 109).

Genomic analyses indicate no direct genetic correlation between schizophrenia and AD, but shared associations with social and environmental factors—such as loneliness, socioeconomic disadvantage, polygenic risk, and family psychiatric history—interact to modestly shape individual schizophrenia risk trajectories (110, 111). The finding that schizophrenia polygenic risk is inversely associated with psychosis in AD further argues against shared etiologic pathways (112). Within this framework, observed associations between schizophrenia, delusional disorder, LOS, and dementia may reflect a continuum of psychosis-related vulnerability rather than a direct inherited pathway specifically predisposing to AD (95). However, studies examined outcomes in offspring of parents with schizophrenia or cognitive features of schizophrenia, but none assessed dementia as an outcome (113). Existing studies focus on the risk of psychiatric disorders in children of parents with severe mental illness—showing a 32% probability of developing such disorders by adulthood (114). However, to our knowledge, we could not identify studies specifically examining dementia risk in offspring with schizophrenia in relation to family history. Environmental stressors may shape epigenetic patterns associated with schizophrenia risk; youth at familial high risk exhibit differences in epigenetic age, indicating altered biological aging trajectories, although the precise mechanisms underlying stress-related methylation changes remain unclear (115). Together, these findings suggest that any shared genetic vulnerability between schizophrenia and dementia is more likely mediated through indirect pathways and contextual factors than through overlapping, disorder-specific genetic architecture.

Sex-stratified genome-wide and transcriptome-wide analyses demonstrate marked sex differences in AD genetic architecture, including female-specific loci and pathways, indicating that susceptibility to neurodegeneration is partially sex-dependent (116, 117). In both autosomal dominant and sporadic AD, genetic risk factors—most notably APOE-ε4—interact with sex to shape cognitive trajectories and neuropathological burden, with stronger associations observed in women, particularly in later life (118–121). APOE-ε4 exhibits strong sex-dependent effects, with women—particularly in mid-to-late older adulthood—showing greater vulnerability to AD–related neurodegeneration than men, including increased tau accumulation, more pronounced hippocampal and temporal cortical atrophy, and greater cognitive and neuropsychiatric burden (118, 121, 122). Notably, even in cognitively normal individuals, female APOE-ε4 carriers demonstrate stronger default mode network disruption, hypometabolism, and cortical thinning, indicating a female-specific susceptibility detectable at preclinical stages of AD (123–125). The available evidence does not support a consistent association between APOE-ε4 and schizophrenia risk independent of AD pathology. Several case–control studies found no significant differences in APOE-ε4 allele frequency between patients and controls (126–128), and a meta-analysis concluded that APOE-ε4 is at most modestly associated with schizophrenia in Caucasian populations, without playing a major etiological role (129). Although one earlier study reported a higher APOE-ε4 frequency in schizophrenia (130), these findings have not been consistently replicated. Conversely, some data suggest that APOE-ε4 may even be less prevalent in specific subgroups, such as individuals with later age at onset (>31 years) (131). However, emerging longitudinal evidence suggests that APOE-ε4 may influence the clinical course of schizophrenia. In a 20-year follow-up of 116 individuals from first hospitalization, ε4 carriage was associated with age-related worsening of psychotic symptoms, particularly hallucinations and delusions (132). Although ε4 frequency does not consistently differ between patients and controls, it has been linked to symptom severity, poorer outcomes, suggesting a role in shaping positive symptom expression. However, findings remain inconsistent across populations, including conflicting results regarding age at onset (133). Furthermore, a meta-analysis of 29 case–control studies demonstrated a significant moderating effect of age, with the APOE-ε4–schizophrenia association becoming more apparent in older individuals (132).

A meta-analysis by González-Castro et al. (134), including 28 association studies (4,703 controls; 3,452 patients), found no overall association between APOE variants and schizophrenia. Notably, a protective effect of the ε3 allele emerged in Asian populations, whereas no associations were observed in Caucasian samples, underscoring ethnic variability and possible sex-dependent effects. Beyond genetic association, APOE may be mechanistically relevant. The ε4 allele has been linked to myelin breakdown and accelerated cognitive decline in AD, and myelin abnormalities are also reported in schizophrenia. Supporting this overlap, post-mortem studies demonstrate increased APOE expression in the dorsolateral prefrontal cortex—particularly Brodmann areas 9 and 46—regions implicated in executive dysfunction in schizophrenia (133). Although global gene expression patterns differ between AD and schizophrenia, both disorders exhibit overlapping molecular dysregulation in the superior temporal gyrus (BA22), likely involving autophagy pathways, suggesting a partially shared biological substrate (135). Genetic evidence further supports convergence in psychotic symptom risk: higher schizophrenia polygenic risk scores are associated with increased likelihood of delusions in AD, pointing to shared liability for psychosis across the two conditions (136). Inflammatory mechanisms may also overlap, as triggering receptor expressed on myeloid cells 2 (TREM2) expression in leukocytes is elevated in both AD and schizophrenia, despite no corresponding increase in TYRO protein tyrosine kinase binding protein (TYROBP). differences or direct schizophrenia-associated genetic effects, highlighting its potential as a cross-disorder biomarker (137). Taken together, current evidence indicates that APOE-ε4 is unlikely to represent a robust shared genetic risk factor for schizophrenia but may act as an age-dependent modifier of symptom progression and late-life cognitive vulnerability, potentially reflecting partial overlap with neurodegenerative mechanisms.

In schizophrenia, well-established sex differences are observed in age at onset, premorbid functioning, symptom profiles, and substance use, while findings in neurocognition and broader psychopathology remain inconsistent (65). Hormonal fluctuations across the female lifespan appear to influence symptom expression, with periods of reduced estrogen associated with greater psychotic severity and cognitive vulnerability (67, 68). Differential associations between dementia severity and molecular markers—such as opposing correlations of DEK proto-oncogene protein (DEK) expression in the anterior cingulate cortex in men versus women—suggest that neurodegenerative pathways may diverge by sex (138).

Schizophrenia and Alzheimer’s disease: neuroimaging and biomarkers perspectives

Earlier neuropathological studies based on relatively small autopsy cohorts consistently failed to demonstrate an increased prevalence or severity of Alzheimer-type pathology—including amyloid β deposition and neurofibrillary tangles—in patients with schizophrenia, even among those with clinical dementia, suggesting that cognitive deficits in “pure” schizophrenia arise from biological mechanisms distinct from classical AD neurodegeneration, although these conclusions should be interpreted cautiously given the limited sample sizes (139–146).

Biomarker and neuropathological studies indicate that cognitive decline in schizophrenia cannot be simply equated with AD. Cerebrospinal fluid (CSF) profiles differ from the classical AD pattern: although reduced amyloid-β42 (Aβ42) levels are observed, tau concentrations are typically normal or even reduced, contrasting with the elevated tau characteristic of AD (147, 148). Peripheral biomarker data similarly suggest a distinct pattern, with lower serum total tau and p-Tau levels compared to controls (149). Proteomic analyses demonstrate a broad reduction across multiple Aβ isoforms rather than the selective Aβ42 decrease seen in AD, and amyloid-related markers appear associated with global cognition and treatment exposure (150).

Neuropathological findings are heterogeneous. While dementia severity in schizophrenia has been linked to neuritic plaques and hippocampal tangles—particularly among APOE-ε4 carriers, supporting a reduced cerebral reserve hypothesis (151)—only a minority of elderly patients meet full neuropathological criteria for AD (7.6% in one cohort), and many show either mixed pathologies or no definitive neurodegenerative substrate (152, 153). Structural imaging studies further demonstrate overlapping white matter microstructural abnormalities in schizophrenia and AD, associated with episodic memory and executive dysfunction, but these findings do not establish a uniform neurodegenerative process (154–156). Both schizophrenia and AD are associated with reduced hippocampal and amygdala volumes, although patterns of association vary by age of onset and negative symptom severity (157). Reduced gray matter volumes have been observed in the cingulate and orbitofrontal cortices, without clear correlations with cognitive or psychiatric symptom intensity (158). In older schizophrenia patients, comorbid dementia was associated with greater cognitive impairment and reduced hippocampal, amygdala, and thalamic volumes (159). In schizophrenia, hippocampal decline appears steeper than in healthy aging and correlates with cognitive and functional impairment, yet such reductions often emerge early and independently of dementia, limiting their specificity for AD and complicating differential diagnosis in later life (24, 160). Furthermore, neuroimaging studies reveal both overlapping and distinct brain network alterations in schizophrenia and AD. Resting-state functional connectivity analyses in 162 AD/MCI patients, 181 schizophrenia patients, and 315 controls showed shared and disorder-specific network abnormalities, driven mainly by default mode, visual, and subcortical networks (161). Structural MRI analyses similarly achieved 81% accuracy in differentiating the disorders and identified subcortical volumes—particularly the putamen—as key markers for classifying late-onset psychosis (162). Both conditions exhibited reduced thalamic nuclei volumes, with the right medial dorsal nucleus in schizophrenia associated with disorganized thought and auditory hallucinations (163). Altered fronto-parieto-limbic activation during emotional processing correlated with behavioural social dysfunction in both disorders, suggesting a shared neural substrate for social impairment independent of diagnosis (164). Additionally, lower plasma brain-derived neurotrophic factor (BDNF) levels have been associated with poorer cognitive performance, with reduced BDNF correlating with lower MMSE scores in chronic schizophrenia (165).

The discrepancy between increased dementia diagnoses and the absence of consistent neuropathological findings may reflect several mechanisms. These include neurobiological processes intrinsic to schizophrenia that are not detectable using standard postmortem markers, as well as an amplified impact of normal ageing due to reduced cognitive reserve. Additional risk factors—such as substance use, socioeconomic adversity, and long-term antipsychotic treatment—may impair cognition without producing macroscopic neuropathological changes (24).

Memory impairment—especially in delayed recall—remains a key overlapping feature in both conditions (166), although specific neuropsychological measures such as verbal memory and abstract reasoning may aid diagnostic differentiation (167). Within schizophrenia, cognitive heterogeneity is clinically meaningful. Better executive functions, working memory, and premorbid functioning have been associated with long-term remission (168), while nonremitted inpatients show greater impairments in processing speed, attention, and working memory (169).

Overall, these findings indicate that schizophrenia-associated dementia risk reflects a complex interplay of modest genetic overlap with AD, familial and sex-modulated vulnerability, environmental exposures, and neurobiological changes, rather than a straightforward convergence with classical Alzheimer-type pathology. Stratified analytic approaches considering sex, family history, and age at onset are required to disentangle inherited susceptibility from illness- and context-related effects.

Schizophrenia and vascular dementia

Vascular dementia affects approximately 2.4–30% of patients with schizophrenia, with significant variation across studies. Park et al. (170) specifically found a 2.4% prevalence in a nationwide South Korean study, while Moraiti and Porfyri (26