Dementia Risk After Glaucoma Surgery and Medical Therapy: A Multicentre Real‐World Cohort Study

ABSTRACT

Background and Purpose

Glaucoma, a neurodegenerative optic neuropathy and leading cause of irreversible vision loss, shares vascular and neurodegenerative mechanisms with dementia. The 2024 Lancet Commission newly identified visual impairment as a modifiable dementia risk factor. As glaucoma management ranges from long‐term medical therapy to laser/surgical intervention, it remains uncertain whether treatment modality influences long‐term systemic outcomes such as dementia. This study compares dementia risk in patients receiving glaucoma surgery versus medical therapy.

Methods

Glaucoma patients aged ≥ 40 years and diagnosed and managed before July 2021 were identified in the TriNetX database and followed from the first definitive treatment—surgery or initiation of medical therapy—until dementia diagnosis, death, or July 2026. The primary outcome was all‐cause dementia; secondary outcomes included Alzheimer's disease, vascular dementia, and visual loss. Subgroup analyses examined surgical timing, modality, age, sex, and comorbidities.

Results

After 1:1 propensity score matching, 51,923 patients were included in each cohort (mean age 64 years; 55.3% female). Glaucoma surgery was associated with reduced risks of all‐cause dementia (HR 0.81) and vascular dementia (HR 0.53), a non‐significant reduction in Alzheimer's disease (HR 0.87), and a higher risk of visual loss (HR 1.52), consistent with more advanced disease among surgical candidates. Associations were stronger with late surgery, traditional incisional procedures, younger age, and absence of comorbidities.

Conclusions

Patients who underwent glaucoma surgery were associated with a lower subsequent risk of all‐cause dementia and vascular dementia, with a non‐significant trend for Alzheimer's disease, compared with those receiving medical therapy. These findings suggest that surgical management is associated with more favourable long‐term cognitive outcomes beyond ocular benefits, which may provide additional context for treatment‐escalation decisions in suitable patients.

Highlights

• Compared with medical therapy, glaucoma surgery was associated with a lower risk of all‐cause dementia and vascular dementia, an association that extended beyond its expected effects on vision preservation.

• Traditional incisional surgery was associated with a greater reduction in dementia risk than minimally invasive glaucoma surgery.

• The association between surgery and lower dementia risk was more pronounced in younger patients and in men.

• The association between surgery and lower dementia risk was present across patients with and without common comorbidities, though it was generally more pronounced among those without these conditions.

Abbreviations

CIs

Confidence intervals

EHRs

Electronic health records

HRs

Hazard ratios

IOP

Intraocular pressure

MIGS

Minimally invasive glaucoma surgeries

PSM

Propensity score matching

1. Introduction

Glaucoma, a progressive optic neuropathy, is the leading cause of irreversible blindness. The condition currently affects an estimated 70 to 80 million people worldwide and is expected to affect more than 110 million by 2040 [1]. Beyond visual loss, glaucoma reduces quality of life, increases fall risk, contributes to depression, and poses substantial economic burden [2, 3]. The cornerstone of treatment is intraocular pressure (IOP) reduction. First‐line therapy is generally medical; however, ocular surface toxicity, systemic adverse effects, and poor adherence may limit its long‐term effectiveness [4]. Laser trabeculoplasty provides an alternative first‐line or adjunctive option but is unsuitable for certain eyes and may not provide durable control [5]. Surgical intervention—including minimally invasive glaucoma surgery (MIGS), glaucoma drainage devices, and traditional incisional procedures such as trabeculectomy—is considered when disease progresses despite maximal therapy, when adherence or intolerance is problematic, or based on patient preference [6]. Given that real‐world treatment escalation is increasingly individualized, with some patients moving to laser or surgery early while others remain on medical therapy for years, understanding how these major therapeutic pathways differ not only in IOP control but also in long‐term systemic outcomes provides clinically meaningful context for shared decision‐making [7].

Dementia, on the other hand, particularly Alzheimer's disease and vascular dementia, represents another major global health challenge. It currently affects more than 55 million individuals worldwide and is projected to more than triple by 2050 [8]. The 2024 Lancet Commission on Dementia Prevention identified visual impairment as a newly recognized modifiable risk factor for dementia, highlighting the importance of sensory health in cognitive ageing [9]. Glaucoma is of particular relevance because it contributes substantially to vision loss and shares overlapping vascular and neurodegenerative mechanisms with dementia; both conditions involve β‐amyloid deposition, tauopathy, and microvascular dysfunction [10]. Furthermore, cognitive impairment may worsen glaucoma outcomes by reducing treatment adherence and health‐care engagement, demonstrating a bidirectional association between the two diseases [11].

Despite these shared pathways and the growing recognition of visual health as a contributor to dementia prevention, it remains unknown whether glaucoma treatment strategy influences subsequent dementia risk. Previous research examining dementia risk in glaucoma has focussed primarily on the presence of glaucoma or the degree of visual impairment, rather than on the specific treatments patients received. Consequently, whether medical therapy and surgical management differ in their associations with subsequent dementia has not been investigated. To address this gap, the present study compares surgical treatment—including laser, minimally invasive, and traditional surgical interventions—and medical therapy in relation to incident dementia using data from a large, multi‐institutional electronic health records (EHR) network.

2. Materials and Methods

2.1. Data Source

This retrospective cohort study (Taipei Medical University‐Joint Institutional Review Board TMU‐JIRB N202510072) was conducted using deidentified data from EHRs maintained by TriNetX. TriNetX has patient data covering more than 200 million individuals across more than 20 countries [12]. The study adhered to the Strengthening the Reporting of Observational Studies in Epidemiology guidelines for cohort studies (Table S1) [13].

2.2. Study Cohort

This study enroled individuals aged ≥ 40 years who had received a confirmed diagnosis and management (either surgery or medication) of glaucoma (ICD‐10‐CM codes H40‐H42) before July 2021. Patients were required to have at least 2 recorded instances of a glaucoma diagnosis documented in their EHR. Patients were categorized on the basis of glaucoma treatment modality into surgery and medication cohorts. The surgery cohort comprised patients who had undergone a surgical intervention for glaucoma (e.g., trabeculectomy, tube shunt implantation, canaloplasty, laser trabeculoplasty, and MIGS), with the date of the first qualifying surgery considered the index date. The medication cohort comprised patients who had received at least 2 prescriptions of an antiglaucoma agent (Anatomical Therapeutic Chemical code S01E) and had not undergone glaucoma‐related surgery throughout the study period, with the date of the first prescription serving as the index date. Patients were followed up until July 2026. Patients with a history of dementia, Alzheimer's disease, Parkinson's disease, stroke, brain injury, or malignant brain tumour prior to the diagnosis of glaucoma were excluded. The diagnostic, prescription, and procedural codes used for cohort identification are provided in Table S2.

2.3. Outcome Measures

The primary outcome was the incidence of all‐cause dementia. Patients were followed from the index date until the first occurrence of an outcome event, death, or the end of the study period (July 2026), whichever occurred first. Secondary outcomes included Alzheimer's disease, vascular dementia, and visual loss. Subgroup analyses were performed on the basis of the time between glaucoma diagnosis and surgery (≥ 3 years vs. < 3 years), age, sex, hyperlipidaemia, type 2 diabetes mellitus, hypertension, depression, obesity, and surgery subtype (traditional surgery or MIGS). To minimize the likelihood of reverse causation, dementia events occurring within the first 12 months after the index date were excluded from the primary analysis. Sensitivity analyses were additionally performed using alternative lag windows of 6 and 24 months to assess the robustness of the findings. E‐values were calculated for the primary and key secondary outcomes to quantify the degree of unmeasured confounding that would be required to explain away the observed associations [14]. An all‐cause mortality analysis was also performed as a negative control to assess residual confounding related to general health status and healthcare engagement.

2.4. Demographic Characteristics and Propensity Score Matching

To reduce baseline differences between the surgery and medication cohorts, we applied 1:1 propensity score matching (PSM) using a greedy nearest‐neighbour algorithm with a calliper of 0.25 times the standard deviation. A standardized mean difference of < 0.1 was considered indicative of well‐balanced covariates. Patients were matched by age at index date, sex, and race. Clinical comorbidities and ophthalmic conditions, specifically, age‐related cataract (H25), other cataract (H26), visual disturbances (H53), degeneration of the macula and posterior pole (H35.3), blindness and low vision (H54), type 2 diabetes mellitus (E11) with ophthalmic complications (E11.3), essential hypertension (I10), hyperlipidaemia (E78.5), and overweight or obesity (E66), were also incorporated into the PSM process. Patients were also matched by a range of laboratory and biometric parameters (i.e., systolic and diastolic blood pressure, body mass index, serum low‐density lipoprotein cholesterol, total cholesterol, high‐density lipoprotein cholesterol, triglycerides, and haemoglobin A1c), as well as dementia medication use (Donepezil, Memantine, Lecanemab, and Rivastigmine).

2.5. Statistical Analysis

Data were analysed using the TriNetX platform in July 2026. The data were visualized using R (version 4.4.2; released October 31, 2024) and relevant packages. Descriptive statistics are presented as means ± SDs for continuous variables and frequencies (%) for categorical variables. Kaplan‐Meier survival analyses were performed, and a Cox proportional hazards model was used to estimate hazard ratios with 95% confidence intervals (95% CIs) and to evaluate outcome‐free survival. All analyses followed an intention‐to‐treat approach. p values of < 0.05 indicated significance.

3. Results

3.1. Baseline Characteristics

Before PSM, the surgery and medication groups comprised 51,955 and 241,945 patients, respectively. After PSM, each cohort comprised 51,923 matched patients. The patient enrolment process is illustrated in Figure 1. The mean age of the matched cohorts was 64 years, and no significant intergroup differences in baseline variables were observed, with standardized mean differences being less than 0.1 (Table 1). The mean follow‐up durations in the surgery and medication cohorts were 2488 and 2733 days, respectively.

FIGURE 1.

Study flow diagram.

TABLE 1.

Baseline characteristics of study cohort before and after propensity score matching.

	Glaucoma surgery versus medical therapy
	Before PSM			After PSM
	Surgery (n = 51,955)	Medical therapy (n = 241,945)	SMD	Surgery (n = 51,923)	Medical therapy (n = 51,923)	SMD
Age, mean ± SD	63.9 ± 10.3	63.3 ± 10.7	0.06	63.9 ± 10.3	64.1 ± 10.2	0.01
Sex, number (%)
Male	23,239 (44.7%)	110,772 (45.8%)	0.02	23,223 (44.7%)	22,966 (44.2%)	0.01
Female	28,713 (55.3%)	130,879 (54.1%)	0.02	28,697 (55.3%)	28,931 (55.7%)	0.01
BMI, mean ± SD	29.4 ± 6.8	30.0 ± 7.3	0.09	29.4 ± 6.8	29.9 ± 7.2	0.07
Race, number (%)
White	28,472 (54.8%)	123,219 (50.9%)	0.08	28,456 (54.8%)	29,093 (56.0%)	0.02
African American	12,298 (23.7%)	54,471 (22.5%)	0.03	12,285 (23.7%)	11,853 (22.8%)	0.02
Asian	2845 (5.5%)	21,613 (8.9%)	0.13	2845 (5.5%)	2663 (5.1%)	0.02
Underlying disease, number (%)
Age‐related cataract	24,049 (46.3%)	45,329 (18.7%)	0.62	24,017 (46.3%)	24,319 (46.8%)	0.01
Other cataract	10,938 (21.1%)	27,236 (11.3%)	0.27	10,915 (21.0%)	10,625 (20.5%)	0.01
Degeneration of macula and posterior pole	5768 (11.1%)	15,185 (6.3%)	0.17	5756 (11.1%)	5615 (10.8%)	0.01
T2DM with ophthalmic complications	5896 (11.3%)	13,735 (5.7%)	0.20	5872 (11.3%)	5718 (11.0%)	0.01
Visual disturbances	10,012 (19.3%)	20,859 (8.6%)	0.31	9987 (19.2%)	9676 (18.6%)	0.02
Blindness and low vision	3667 (7.1%)	7531 (3.1%)	0.18	3645 (7.0%)	3341 (6.4%)	0.02
Essential hypertension	22,737 (43.8%)	98,836 (40.9%)	0.06	22,710 (43.7%)	22,400 (43.1%)	0.01
T2DM	14,890 (28.7%)	57,489 (23.8%)	0.11	14,861 (28.6%)	14,708 (28.3%)	0.01
Hyperlipidaemia	14,254 (27.4%)	62,431 (25.8%)	0.04	14,243 (27.4%)	14,307 (27.6%)	0.00
Overweight and obesity	7125 (13.7%)	30,759 (12.7%)	0.03	7119 (13.7%)	6984 (13.5%)	0.01
Lab data, mean ± SD
Total cholesterol (mg/dL)	179.0 ± 45.3	179.6 ± 45.6	0.01	179.0 ± 45.3	179.8 ± 45.6	0.02
LDL‐C (mg/dL)	101.1 ± 37.0	101.7 ± 37.0	0.02	101.1 ± 37.0	101.6 ± 37.1	0.02
HDL‐C (mg/dL)	51.3 ± 18.9	51.0 ± 19.7	0.01	51.3 ± 18.9	50.7 ± 20.0	0.03
Triglyceride (mg/dL)	132.2 ± 93.1	132.3 ± 98.0	0.00	132.2 ± 93.1	133.1 ± 104.6	0.01
Systolic blood pressure (mmHg)	132.1 ± 19.0	132.2 ± 19.6	0.01	132.1 ± 19.0	132.7 ± 19.5	0.03
Diastolic blood pressure (mmHg)	75.4 ± 11.3	75.5 ± 11.8	0.01	75.4 ± 11.3	75.3 ± 11.6	0.02
HbA1c (%)	6.7 ± 1.8	6.7 ± 1.7	0.02	6.7 ± 1.8	6.8 ± 1.8	0.03

Abbreviations: BMI, body mass index; HbA1c, haemoglobin A1c; HDL‐C, high‐density lipoprotein cholesterol; LDL‐C, low‐density lipoprotein cholesterol; PSM, propensity score matching; SD, standard deviation; SMD, standardized mean difference; T2DM, type 2 diabetes mellitus.

3.2. Glaucoma Surgery Is Associated With Lower Dementia Risk but Higher Visual Loss Risk Than Medical Therapy Is

Compared with the medical group, the surgery group had significantly lower risks of all‐cause dementia (HR 0.81, 95% CI 0.74–0.88) and vascular dementia (HR 0.53, 95% CI 0.43–0.67), and a numerically lower but not statistically significant risk of Alzheimer's disease (HR 0.87, 95% CI 0.74–1.03; Figure 2). The number of cases of frontotemporal dementia and Lewy body dementia was too low to allow for meaningful Kaplan‐Meier plots to be generated. The surgery group had a higher risk of blindness (HR 1.52, 95% CI 1.45–1.60) than the medical group did, likely reflecting the greater disease severity typically seen in patients selected for surgical intervention.

FIGURE 2.

Risk of dementia and visual loss within 10 years after glaucoma surgery or medical treatment. (A) Forest plot. (B) Kaplan‐Meier curves.

Lag‐window sensitivity analyses with 6‐month and 24‐month washout periods yielded consistent findings (Figure S1). The stability of hazard ratios after excluding outcome events occurring shortly after treatment initiation indicates that the observed protective association is unlikely to be driven by early immortal person‐time.

E‐values for all‐cause dementia were 1.77 for the point estimate and 1.53 for the confidence interval limit; the corresponding point‐estimate E‐value for Alzheimer's disease was 1.56 (the confidence‐interval‐limit E‐value was not applicable, as the 95% CI already included the null value), and 3.18/2.35 for vascular dementia. In negative‐control analysis, glaucoma surgery was associated with a smaller reduction in all‐cause mortality (HR 0.96, 95% CI 0.93–0.99) than in all‐cause dementia.

3.3. Subgroup Analysis by Timing of Intervention and Surgical Modality

With stratification by surgery timing (early surgery, defined as performed within 3 years of glaucoma diagnosis; late surgery, defined as performed 3 years or more after diagnosis), both early and late intervention were associated with significant reductions in dementia risk and with higher blindness rates (Figure 3A). Early surgery was associated with lower risks of all‐cause dementia (HR 0.80, 95% CI 0.74–0.88), Alzheimer's disease (HR 0.80, 95% CI 0.68–0.93), and vascular dementia (HR 0.62, 95% CI 0.50–0.77). Late surgery exhibited even stronger protective effects, with greater reductions in all‐cause dementia (HR 0.46, 95% CI 0.40–0.54), Alzheimer's disease (HR 0.48, 95% CI 0.36–0.64), and vascular dementia (HR 0.25, 95% CI 0.17–0.37). Blindness risk remained higher in both early and late surgery groups, consistent with more advanced disease among surgical candidates. The stronger dementia associations observed in the late‐surgery group may reflect greater absolute benefit among patients with more severe baseline disease.

FIGURE 3.

Subgroup analysis by timing of intervention and surgical modality. (A) < 3 versus ≥ 3 years after glaucoma diagnosis. (B) Traditional surgery versus minimally invasive surgery.

We next performed subgroup analysis by surgical modality (traditional vs. minimally invasive surgeries; Figure 3B). No significant associations were observed between minimally invasive surgeries and dementia outcomes. Minimally invasive surgeries was not associated with a reduced risk of all‐cause dementia (HR 1.11, 95% CI 0.88–1.39), Alzheimer's disease (HR 1.03, 95% CI 0.66–1.61), vascular dementia (HR 0.96, 95% CI 0.56–1.66), but was associated with a lower risk of blindness (HR 0.69, 95% CI 0.55–0.85). By contrast, traditional surgeries, including trabeculectomy and tube shunt implantation, were consistently associated with reduced risks of all‐cause dementia (HR 0.82, 95% CI 0.75–0.90), Alzheimer's disease (HR 0.82, 95% CI 0.69–0.98), and vascular dementia (HR 0.54, 95% CI 0.42–0.69), accompanied by a higher risk of blindness (HR 1.53, 95% CI 1.45–1.60).

3.4. Subgroup Analysis by Demographic and Comorbid Factors

The protective association of glaucoma surgery was consistent across demographic groups (Figure 4). Patients aged 40–65 years were less likely to experience all‐cause dementia (HR 0.63, 95% CI 0.54–0.75), Alzheimer's disease (HR 0.57, 95% CI 0.41–0.81), or vascular dementia (HR 0.59, 95% CI 0.40–0.88) than were patients aged ≥ 65 years (all‐cause dementia HR 0.77, 95% CI 0.71–0.84; Alzheimer's HR 0.86, 95% CI 0.72–1.01; vascular dementia HR 0.50, 95% CI 0.40–0.64). The protective association of surgery was observed in both males and females. Across all comorbidity subgroups, blindness risk remained consistently elevated in surgical patients.

FIGURE 4.

Subgroup analysis by demographic and comorbid factors.

With stratification by comorbid conditions, glaucoma surgery was associated with reduced dementia risk across the comorbidity subgroups. Patients without depression (HR 0.78, 95% CI 0.69–0.86), diabetes (HR 0.80, 95% CI 0.71–0.90), hypertension (HR 0.74, 95% CI 0.58–0.94), hyperlipidaemia (HR 0.71, 95% CI 0.60–0.83), or obesity (HR 0.75, 95% CI 0.66–0.84) who received surgery were less likely to experience all‐cause dementia than were such patients who received only medical therapy, and patients with depression (HR 0.80, 95% CI 0.68–0.93), diabetes (HR 0.72, 95% CI 0.63–0.81), hypertension (HR 0.80, 95% CI 0.73–0.88), hyperlipidaemia (HR 0.87, 95% CI 0.78–0.97), or obesity (HR 0.85, 95% CI 0.77–0.95) who received surgery experienced attenuated but generally still favourable associations relative to those who received only medical therapy. Similar patterns were observed for Alzheimer's disease and vascular dementia. Across all comorbidities, the risk of blindness remained elevated, which is consistent with more advanced disease in surgical patients.

4. Discussion

In this large, propensity score–matched cohort study, glaucoma surgery was associated with reduced risks of all‐cause dementia and vascular dementia, a non‐significant trend towards reduced risk of Alzheimer's disease, and a greater risk of blindness compared with medical therapy. Traditional incisional procedures, which involve greater IOP control, had greater protective effects than those of minimally invasive techniques, with the traditional‐surgery‐only estimate essentially unchanged from the pooled surgery‐group estimate. Positive effects were consistent across age, sex, and comorbidity subgroups, with a particularly pronounced benefit in younger patients, males, and those undergoing later surgery. These findings suggest that in addition to preserving vision, surgical intervention is associated with a lower incidence of clinically diagnosed all‐cause and vascular dementia, with a similar but statistically non‐significant trend for Alzheimer's disease.

Patients selected for glaucoma surgery demonstrated a higher risk of blindness, which likely reflects a more advanced disease status at baseline. However, the surgery group consistently showed lower risks of dementia—particularly vascular dementia—compared with patients treated with medical therapy alone. This pattern suggests that although surgical candidates remain vulnerable to vision loss due to severe glaucoma, the more aggressive and sustained IOP reduction achieved through surgery was associated with a lower risk of subsequent dementia, potentially through improved ocular perfusion or reduced vascular stress [15]. Subgroup analyses further revealed that traditional incisional surgery conferred greater protection than minimally invasive techniques did, which is consistent with the superior and sustained pressure reduction achieved by more aggressive interventions [16].

Although glaucoma patients with some comorbidities were less likely to benefit from surgery than were those without these conditions, beneficial results remained robust across all groups, underscoring surgery's generalizability. Notably, we observed that the protective effect of glaucoma surgery on dementia risk was more pronounced among patients undergoing surgery at later disease stages and in younger individuals, partly reflecting confounding by indication and survivor bias: patients selected for surgery at advanced stages often receive more intensive ophthalmic and systemic care, and those who survive to undergo late intervention may represent a healthier subset [17]. By contrast, the greater relative benefit noted in younger patients may be explained by their higher neural plasticity and longer lifespan, which amplify the cognitive advantages of preserving visual function [18]. Additionally, because the baseline dementia incidence is lower in younger individuals, reductions in absolute risk translate into more substantial relative risk decreases [19]. Together, these findings suggest that surgery is associated with a reduced incidence of dementia beyond vision preservation for patients with glaucoma, highlighting the importance of considering long‐term cognitive outcomes in treatment planning.

When comparing glaucoma surgery with conservative medical therapy, clinicians must balance the greater efficacy of surgical IOP reduction against the higher procedural morbidity. Surgical approaches provide more substantial and durable IOP control than medications, particularly in moderate‐to‐severe or rapidly progressive disease, and may better preserve visual field in advanced cases [20]. Nevertheless, surgery carries increased risks of cataract formation, hypotony, infection, and ocular discomfort, which limits its role as a universal first‐line option [21]. In clinical practice, surgery is considered when glaucoma cannot be controlled with maximally tolerated medical or laser therapy, when adherence or intolerance limits the effectiveness of medications, or when patients present with advanced disease requiring aggressive pressure lowering [22]. MIGS and laser trabeculoplasty offer safer options earlier in the treatment course but typically achieve less robust IOP reduction, making them unsuitable for severe glaucoma [7]. Thus, real‐world decision‐making reflects a spectrum in which some patients escalate early to surgical treatment while others remain on medical therapy for years. In this context, understanding whether these major therapeutic pathways differ not only in ophthalmic outcomes but also in systemic endpoints—such as dementia risk—provides clinically meaningful information that may guide shared decision‐making. This added value becomes relevant when discussing treatment escalation with patients at high risk of progression or with limited tolerance for long‐term medications.

The observed association between glaucoma and an increased risk of Alzheimer's disease and vascular dementia can be explained by neurodegenerative, vascular, and molecular mechanisms [23]. Both conditions are characterized by progressive neuronal loss, with retinal ganglion cell degeneration in glaucoma paralleling cortical neurodegeneration in Alzheimer's disease [24]. Pathological hallmarks, including β‐amyloid deposition and hyperphosphorylated tau accumulation, have been detected in both glaucomatous retinas and Alzheimer's brains, suggesting a common proteinopathy that may drive neuronal dysfunction across visual and cognitive pathways [25]. Vascular dysregulation also plays key roles in both glaucoma and dementia pathophysiology. Impaired microcirculation and autoregulation at the optic nerve head reflect systemic vascular pathology, such as atherosclerosis and vasospasm, that closely mirrors the cerebral hypoperfusion implicated in both Alzheimer's and vascular dementia [26]. Genetic studies further support a shared susceptibility, identifying overlapping loci involved in mitochondrial metabolism, oxidative stress, and neuroinflammation [27]. Autoimmune processes, such as agonistic β2‐adrenergic receptor autoantibodies, have also been implicated in both glaucoma and Alzheimer's disease, potentially contributing to adrenergic dysregulation and progressive neuronal injury [28]. Finally, neuroimaging and biomarker studies have revealed structural and functional alterations in visual and associative cortical regions of the brain in patients with glaucoma, strengthening the theory that glaucoma is more than an ocular disease, involving central nervous system pathways also vulnerable in dementia [29, 30]. Collectively, these overlapping mechanisms highlight why glaucoma may not only represent an ophthalmic disease but also a systemic neurodegenerative condition, thereby explaining the heightened dementia risk.

A plausible biological explanation for the association between glaucoma surgery and reduced dementia risk is the surgical reduction of IOP, the central modifiable factor driving glaucomatous neurodegeneration. Excessive IOP induces both biomechanical stress and vascular compromise at the optic nerve head, accelerating retinal ganglion cell apoptosis and promoting trans‐synaptic degeneration along visual pathways to the brain [31]. These processes contribute to the neurodegenerative cascades characteristic of Alzheimer's disease and vascular dementia, including those that involve mitochondrial dysfunction and chronic neuroinflammation [32]. Experimental evidence indicates that chronic excessive IOP induces β‐amyloid deposition and tauopathy in the lateral geniculate nucleus of nonhuman primates, thereby linking ocular hypertension directly to Alzheimer's‐type pathology [33]. By surgically reducing IOP, operative intervention may reduce axonal strain, preserve retinal ganglion cell integrity, and improve ocular perfusion, thereby limiting the propagation of degenerative signals to cortical visual and associative areas. Such preservation of sensory input could mitigate the downstream synaptic dysfunction and protein aggregation that contribute to dementia pathophysiology [34]. Although the hypothesis is biologically plausible and supported by the presence of shared disease mechanisms, the clinical literature has yet to provide direct evidence that glaucoma surgery itself reduces incident dementia.

This research offers valuable insights into the relationship between glaucoma surgery and dementia risk, revealing potential benefits beyond IOP control. The use of the TriNetX platform, which aggregates data from a large and diverse global patient population, enhances the generalizability of our findings across different demographic and geographic groups. However, this research has limitations. First, despite our adjustment for numerous confounders, unmeasured variables, including lifestyle, socioeconomic, and educational factors, may have influenced the results. Second, although immortal time bias remains an inherent concern, the long latency of dementia and the consistency of results across multiple lag‐window sensitivity analyses indicate that the observed associations are unlikely to be fully explained by this bias. Third, diseases were identified using ICD‐10‐CM diagnostic codes from EHRs rather than validated clinical parameters such as IOP, visual field defects, cognitive function scores, or measures of functional disability. Similarly, medication adherence in the medical therapy group could not be assessed, as TriNetX does not capture pharmacy fill or refill or patient‐reported adherence data, and differential adherence may have contributed to the observed between‐group differences. Established vascular risk factors such as smoking and obstructive sleep apnoea were likewise not included as covariates. Unlike the vital‐sign‐ and laboratory‐based measures used for matching, which are captured routinely at nearly every clinical encounter regardless of the reason for the visit, ascertainment of smoking status and obstructive sleep apnoea depends on active patient‐reported history‐taking or a dedicated diagnostic work‐up (e.g., referral for polysomnography), and both are known to be substantially underrecorded in structured EHR data—smoking status because of inconsistent documentation [35], and obstructive sleep apnoea because of a high rate of underdiagnosis in the general population [36]. Finally, selection bias may have occurred, given that patients undergoing surgery often have more severe disease than those managed medically. E‐value analyses indicated that an unmeasured confounder of a magnitude comparable to established dementia risk factors could account for part of this association, while a negative‐control analysis of all‐cause mortality showed a considerably smaller effect than that observed for dementia, suggesting that such selection is unlikely to fully explain the association, although it cannot be excluded as a contributing factor. Overall, the current findings can inform clinical decision‐making, highlighting the broader implications of aggressive IOP management, particularly in patients at high risk of dementia. Future large‐scale studies are warranted to confirm these associations and elucidate underlying mechanisms.

In conclusion, glaucoma surgery was associated with a lower subsequent risk of all‐cause dementia and vascular dementia, with a non‐significant trend for Alzheimer's disease, compared with medical therapy. These associations were consistent across clinical subgroups and stronger for traditional incisional procedures. This association may offer additional context when discussing treatment escalation with suitable patients, and supports further prospective investigation into the relationship between glaucoma treatment and dementia risk.

Author Contributions

L.‐Y.L., W.‐J.J., and C.‐H.T. conceptualized and designed the study. L.‐Y.L. and J.‐S.W. analysed, interpreted, and visualized the data. L.‐Y.L., S.‐Y.L., T.‐W.H. and F.S.Y. drafted the manuscript. All authors contributed to manuscript review and editing and approved the final version. S.‐H.L. agrees to serve as the guarantor.

Funding

The authors have nothing to report.

Ethics Statement

The study protocol was approved by the Joint Institutional Review Board of Taipei Medical University (approval number: N202510072). This study adhered to the ethical principles of the Declaration of Helsinki and the Strengthening the Reporting of Observational Studies in Epidemiology guidelines for cohort studies.

Consent

Given that all data were deidentified, the requirement for informed consent was waived by the institutional review board.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figure S1: Forest plot and Kaplan‐Meier curves for dementia‐free survival in patients with glaucoma undergoing surgery versus medical therapy, with sensitivity analyses using different lag windows (6, 12, and 24 months).

Table S1: STROBE Checklist for this Cohort Study.

Table S2: Diagnostic, Procedural, and Medication Codes Used for Cohort Definition.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.