Higher glycated hemoglobin amplifies the effect of apolipoprotein E epsilon 4-related cognition and olfaction impairments in type 2 diabetes

Abstract

BACKGROUND

Apolipoprotein E epsilon 4 (APOE4) is recognized as a genetic risk factor for cognitive decline and neurodegeneration in both type 2 diabetes mellitus (T2DM) and Alzheimer’s disease, while glycated hemoglobin (HbA1c) reflects persistent hyperglycemia and serves as a key indicator of long-term glycemic control in T2DM. Although both factors have been individually linked to neurobehavioral deficits, it remains uncertain whether HbA1c contributes to APOE4-related cognitive and olfactory impairment in individuals with T2DM.

AIM

To investigate the role of HbA1c in APOE4-associated cognitive and olfactory dysfunction in patients with T2DM.

METHODS

Of 636 T2DM patients were recruited from five medical centers in Wuhan, Hubei Province, China. APOE genotyping was evaluated by polymerase chain reaction using Gerard’s method. Cognitive and olfactory functions were assessed by mini-mental state examination and Connecticut chemosensory clinical research center test, respectively. Regression analysis was employed to assess the independent and interactive effects of HbA1c on APOE4-associated cognitive and olfactory function.

RESULTS

APOE4 was associated with increased risks of cognitive impairment [odds ratios (OR) = 1.815, P = 0.021] and olfactory dysfunction (OR = 2.588, P < 0.001). Higher HbA1c levels were also related to worse cognitive (OR = 1.189, P < 0.001) and olfactory performance (OR = 1.149, P = 0.011). HbA1c exerted a moderating effect, yet not a mediating effect, between APOE4 and its impacts on cognition and olfaction. Specifically, a higher level of HbA1c exacerbated the damaging effect of APOE4, as shown by significant interaction effects on both cognitive impairment (OR = 2.687, P < 0.001) and olfactory dysfunction (OR = 1.440, P = 0.027).

CONCLUSION

Elevated HbA1c levels are associated with increased risks of cognitive and olfactory impairments in patients with T2DM and may exacerbate the detrimental effects of APOE4. These findings underscore the need for early preventive strategies targeting individuals with both poor glycemic control and APOE4 carriage to mitigate neurodegenerative risk.

Key Words: Glycated hemoglobin; Apolipoprotein E epsilon 4; Type 2 diabetes mellitus; Cognitive impairment; Olfactory function

Core Tip: This study explores how the apolipoprotein E epsilon 4 (APOE4) genotype and glycated hemoglobin (HbA1c) levels influence cognitive and olfactory performance in individuals with type 2 diabetes mellitus. The findings suggest that APOE4 is linked to both cognitive and olfactory deficits, and that elevated HbA1c levels further intensify these effects. Rather than serving as a mediator, HbA1c modifies the relationship, strengthening the adverse impact of APOE4. These results indicate that both APOE4 and poor glycemic control contribute to neurofunctional decline, underscoring the importance of early intervention in at-risk diabetic populations.

INTRODUCTION

Type 2 diabetes mellitus (T2DM) is a growing global health crisis, particularly in aging populations, with projections estimating a prevalence of 642 million by 2040[1]. Characterized by chronic hyperglycemia due to insufficient insulin secretion and insulin resistance[2]. Over time, these changes may lead to complications, including impaired cognitive and olfactory function[3]. It is increasingly recognized as a risk factor for neurodegenerative diseases such as Alzheimer’s disease (AD) and Parkinson’s disease[4-6]. AD, the most common form, often begins with subtle symptoms, among which olfactory dysfunction (OD) frequently precedes cognitive decline[7,8]. Notably, individuals with T2DM often experience early deficits in both cognition and olfaction[9], suggesting a potential link between metabolic dysregulation and neurodegeneration. T2DM and AD share several overlapping pathological mechanisms, including inflammation, insulin resistance, oxidative stress, and disturbances in lipid metabolism, all of which contribute to disease development and progression[10]. Hyperglycemia in T2DM triggers chronic inflammation, promotes neuronal injury, and exacerbates the neuroinflammatory responses characteristic of AD[11]. Similarly, insulin resistance, a hallmark of T2DM, impairs brain insulin signaling, disrupts synaptic function and cognitive performance, thereby speeds up AD progression[12]. Moreover, disrupted lipid metabolism is a common feature of both diseases, aggravating AD and promoting cardiovascular risk in T2DM[13]. Apolipoprotein E epsilon 4 (APOE4) plays an additional role by affecting cerebral lipid handling, particularly cholesterol transport and clearance. Additionally, APOE4 is a prominent genetic risk factor for AD, recent studies suggest that there is a 2-3 fold increase in risk when carrying one copy of ε4 and an 8-12 fold increase when carrying two copies of ε4[14]. These converging pathological mechanisms highlight the critical importance of early intervention in both conditions.

Cognitive and OD are increasingly recognized in individuals with T2DM, yet their underlying mechanisms remain poorly understood[15,16]. The olfactory bulb projects directly to the entorhinal cortex and hippocampus, regions essential for memory formation and among the earliest affected in Alzheimer’s pathology. Glycated hemoglobin (HbA1c), a marker of chronic hyperglycemia, plays a central role in T2DM progression, but its link to neurofunctional outcomes, particularly in cognitive and olfactory domains, has been largely overlooked[17].

Although APOE4 is a well-established genetic risk factor for AD[18], its interaction with HbA1c in T2DM has not been clearly defined. Cognitive and olfactory impairments are frequently observed in individuals with T2DM, yet they are seldom examined in combination. Given their shared neuroanatomical substrates and the convergence of metabolic and genetic risk factors, investigating the joint impact of APOE4 and glycemic control on these domains is of clinical relevance. Clarifying these associations may enhance early risk detection and contribute to more precise strategies for neurodegeneration prevention. Despite the growing interest in metabolic-cognitive links, the combined effects of APOE4 and HbA1c on brain function remain insufficiently studied. This study is the first to evaluate these interactions simultaneously in patients with T2DM, offering new insights into integrated risk assessment and personalized intervention approaches.

MATERIALS AND METHODS

Data sources and recruitment

A total of 1041 patients with T2DM were recruited from five medical centers in Wuhan, Hubei Province, China, between January 2012 and November 2018. Specifically, the sample included 364 patients from the Central Hospital of Wuhan, 177 from Wuhan No. 1 Hospital, 146 from Wuhan General Hospital of Guangzhou Military Region, 31 from Liyuan Hospital of Tongji Medical College of HUST, and 323 from Jianghan Road Community Hospital. A consecutive sampling approach was employed.

The diagnosis of T2DM was based on the World Health Organization National Diabetes Group Criteria (2006)[19]. Inclusion and exclusion criteria were applied as detailed in prior publications[20-22]. Briefly, the inclusion criteria were as follows: (1) Age ≥ 45 years; (2) Residency in the area for at least five years (≥ 5 years); and (3) The ability to complete neuropsychological assessments and provide written informed consent. The exclusion criteria were as follows: (1) A prior diagnosis of dementia before T2DM (n = 28); (2) History of major neurological disorders, including prior head injury, stroke, brain tumor, coma, transient ischemic attack, or epilepsy (n = 81); (3) Auditory or visual impairment, thyroid disease, use of medications affecting cognitive function, alcohol dependence, depression, schizophrenia, or other severe psychiatric conditions, including acute stress reactions, post-traumatic stress disorder, or brief psychotic episodes (n = 61); (4) Comorbidities known to affect metabolic status or red blood cell turnover, including thalassemia, sickle cell disease, advanced chronic kidney disease (stage 4 or higher), severe anemia, hyperthyroidism, Cushing’s syndrome, acromegaly, or a history of glucocorticoid, erythropoietin, or high-dose antioxidant use (n = 53); (5) Self-reported or clinically documented olfactory loss, active rhinitis, or recent upper respiratory tract infection (n = 45); (6) Inability to complete cognitive or olfactory tests due to language barriers, illiteracy, or lack of cooperation (n = 36); and (7) Incomplete or missing data on any core variables, including HbA1c levels, APOE genotype, mini-mental state examination (MMSE) scores, or olfactory threshold scores (OTS) (n = 101). A total of 405 patients were excluded, yielding a final study cohort of 636 participants for subsequent analyses.

The Medical Ethics Committee approved the study protocol (No. 2013-S071), Tongji Medical College, Huazhong University of Science and Technology. All procedures followed the ethical principles outlined in the Declaration of Helsinki, and written informed consent was obtained from all participants.

APOE genotyping

All blood samples were collected and processed within two hours, undergoing centrifugation to separate platelets, white blood cells, red blood cells, and plasma, with aliquots subsequently stored at -80 °C. DNA was extracted from white blood cells, and APOE genotyping was carried out using polymerase chain reaction, following a modified version of Gerard’s method with the multiplex amplification refractory mutation system[23]. All genotyping procedures were conducted in certified laboratories under validated protocols to ensure quality and reliability.

Cognitive test

All participants completed a comprehensive medical history review and physical examination. The MMSE was administered by two trained inspectors with specialized backgrounds in neurology, with repeat testing in a random 10% subset to confirm reliability. Cognitive impairment was diagnosed using Petersen’s criteria[24]: (1) Memory complaint; (2) Normal daily living activities; (3) Normal general cognitive function; (4) Memory ability lower than expected for age; and (5) MMSE scores at < 27 was diagnosed as cognition impairment (CI), while MMSE scores at ≥ 27 was diagnosed as cognition normal (CN).

Olfactory test

All participants underwent olfactory testing using the Connecticut chemosensory clinical research center (CCCRC) protocol. Assessments were administered with standardized instructions by blinded examiners across all sites. The olfactory detection threshold test employed a series of 12 graded concentrations of butanol (ranging from 2.3 × 10^-5% to 4%), beginning from the lowest concentration and increasing to a maximum of 4%. Deionized water served as the control. The detection threshold was defined as the lowest concentration at which participants correctly identified butanol four consecutive times at the same dilution level. If identification was incorrect in any of the four attempts, the concentration was gradually increased until four correct identifications were achieved consecutively. Olfactory test, presented in Figure 1, reflect olfactory function, with higher scores indicating better function. Based on previous CCCRC-based studies, scores below 2 are typically indicative of anosmia[25]. In this study, OD was defined as an OTS ≤ 2[26].

Figure 1 Study profile. Apolipoprotein E genotype and glycated hemoglobin level in peripheral blood samples were assessed from 636 patients with type 2 diabetes mellitus. Cognitive and olfactory functions, primarily mediated by the temporal lobe (dashed-line box), were evaluated using the mini-mental state examination and the Connecticut chemosensory clinical research center test, respectively. T2DM: Type 2 diabetes mellitus; HbA1c: Glycated hemoglobin; APOE: Apolipoprotein E.

HbA1c test

Fasting blood samples were collected at 8:00 in the hospital. Plasma was separated by centrifugation, and HbA1c was measured using high-performance liquid chromatography (HPLC)[27]. After protein precipitation with acetonitrile, the supernatant was filtered and injected into an HPLC system with a reversed-phase C18 column. The mobile phase consisted of acetonitrile and water with 0.1% trifluoroacetic acid, and detection was performed at 210 nm. HbA1c levels were quantified using calibration curves, and results were expressed as the percentage of HbA1c relative to total hemoglobin. All HbA1c assays were conducted using standardized protocols with internal quality controls to ensure accuracy and reproducibility.

Covariates of the study population

Demographic and clinical characteristics, such as age and sex, were collected using a semi-structured questionnaire. Additionally, medical history, particularly hypertension, hyperlipidemia, coronary artery disease, and the duration of diabetes, was recorded.

Statistical analysis

All data were analyzed using IBM SPSS 27.0 and GraphPad Prism 10.0. Continuous variables were presented as mean ± SD while categorical variables were presented as frequency (percentage). Group differences in MMSE scores, OTS, and other baseline characteristics were compared based on APOE4 carrier status using the student’s t-test for normally distributed variables, the Mann-Whitney U test for skewed distributions. Binary logistic regression analyses were conducted to evaluate the associations of APOE4 and HbA1c with cognitive and OD separately. Spearman’s rank correlation was employed to analyze the relationship between MMSE scores and OTS.

The interaction of APOE4 and HbA1c on CI and OD were analyzed using binary logistic regression models, with adjustments for covariates including age, gender, diabetes duration, history of hypertension, hyperlipidemia and coronary artery disease.

The additive interaction between APOE4 and high HbA1c on the risk of CI and OD was measured by calculating relative excess risk due to interaction (RERI) and attributable proportion (AP) using the method described by Hosmer and Lemeshow and calculating by Andersson et al’s calculator[28]. RERI was used to evaluate the joint effect of APOE4 and high HbA1c, while AP represented the proportion of risk attributable to their interaction. Calculations were conducted in Excel based on regression coefficients and covariance matrices. Confidence intervals (95%CI) were estimated using approximate variance methods, and P values and odds ratios (OR) were reported to three decimal places. P value < 0.05 was considered as statistically significant.

RESULTS

Participant characteristics

Table 1 shows the demographic characteristics of the participants in the T2DM cohort. The cohort consisted of 636 patients with T2DM, in which 234 (36.8%) were male, and mean (± SD) age was 64.17 (7.59) years. Overall, there were 551 APOE4- (non-APOE4 carriers) participants and 85 APOE4 + (APOE4 carriers) participants. HbA1c levels and fasting plasma glucose had no significant difference (P = 0.1454, P = 0.4243, Table 1) with APOE4. Similarly, diabetes duration, the prevalence of hypertension, hyperlipidemia and coronary artery disease had no significant differences observed (P = 0.2756, P = 0.5700 and P = 0.9940, P = 0.8100, respectively, Table 1).

Table 1 Clinical characteristics of the cohort grouped by apolipoprotein E genotypes, mean ± SD/n (%).

Characteristics	All patients	APOE genotype		P value
		Non-APOE4 carriers	APOE4 carriers
Number of subjects	636	551	85
Age (years)	64.17 ± 7.59	64.12 ± 7.6	64.54 ± 7.57	0.6328
Gender (male, %)	234 (36.8)	204 (37)	30 (35.3)	0.7580
Glycosylated hemoglobin (%)	7.83 ± 1.86	7.87 ± 1.89	7.55 ± 1.68	0.1454
Fasting plasma glucose (mmol/L)	9.08 ± 3.56	9.13 ± 3.66	8.8 ± 2.78	0.4243
Hypertension (yes)	356 (56)	306 (55.5)	50 (58.8)	0.5700
Hyperlipidemia (yes)	127 (20)	110 (20)	17 (20)	0.9940
Coronary artery disease (yes)	70 (11)	60 (10.9)	10 (11.8)	0.8100
Diabetes duration (years)	8.39 ± 6.8	8.27 (6.56)	9.14 (8.18)	0.2756
OTS	5.41 ± 2.29	5.48 (2.23)	4.93 (2.6)	0.0383
OTS, range (0-2)	118 (18.6)	91 (16.5)	27 (31.8)	< 0.0010
OTS, range (3-5)	89 (14)	81 (14.7)	8 (9.4)	0.1950
OTS, range (6-8)	429 (67.5)	379 (68.8)	50 (58.8)	0.0700
MMSE score	27.49 ± 2.4	27.57 ± 2.4	27 ± 2.37	0.0424
CN (MMSE ≥ 27)	471 (74.1)	416 (75.5)	55 (64.7)	0.0360
CI (MMSE < 27)	165 (25.9)	135 (24.5)	30 (35.3)	0.0360

APOE4: Apolipoprotein E epsilon 4; MMSE: Mini-mental state examination; OTS: Olfactory threshold score; CN: Cognition normal; CI: Cognition impairment.

APOE4 is associated with cognitive and olfactory impairments in T2DM patients

The APOE4 genotype has been identified as a significant genetic risk factor for neurodegenerative diseases, such as AD[29]. However, a growing number of research suggest that APOE4 not only affects cognitive function but also affects olfactory function. Harboring APOE4 alleles can impact olfactory function, even in healthy older adults[30]. To ensure analytical robustness, we assessed whether APOE4 status and HbA1c levels were independently distributed. No significant association was identified (P = 0.900, Supplementary Table 1). APOE4 carriers had significantly lower MMSE scores compared to non-APOE4 carriers (P = 0.0424, Figure 2A). Olfactory function, as measured by the OTS, is also significantly lower in the APOE4 carriers group compared to the non-APOE4 carriers (P = 0.0383, Figure 2B). These suggest that individuals carrying the APOE4 allele are more likely to have cognitive and OD. By single factor and multiple factor logistic regression analysis, we confirmed that APOE4 carriers have a 68.1% higher risk of cognitive impairment (P = 0.036, Table 2) and a 135.3% increased risk of olfactory impairment (P < 0.001, Table 2) compared to non-APOE4 carriers. These association was still significant after adjusted by age, gender, hypertension, hyperlipidemia, coronary artery disease, HbA1c and diabetes duration. (P = 0.021; P < 0.001, Table 2). Therefore, APOE4 is a significant genetic risk factor for both cognitive and olfactory impairments.

Figure 2 Effects of apolipoprotein E epsilon 4 genotype on cognitive and olfactory function. A: Apolipoprotein E epsilon 4 (APOE4) carriers had significantly lower mini-mental state examination scores compared to non-APOE4 carriers (P = 0.0424); B: APOE4 carriers had significantly lower olfactory threshold scores compared to non-APOE4 carriers (P = 0.0383). APOE4 - represents non-APOE4 carriers, APOE4 + represents APOE4 carriers. APOE4: Apolipoprotein E epsilon 4; MMSE: Mini-mental state examination.

Table 2 Associations of apolipoprotein E epsilon 4 with cognitive and olfactory function analyzed by logistic regression.

Outcome	Variable	Model	β	SE	Wald	P value	OR	95%CI
CI	APOE4	Model 1	0.519	0.248	4.397	0.036	1.681	1.034-2.731
	APOE4	Model 2	0.514	0.253	4.133	0.042	1.673	1.019-2.747
	APOE4	Model 3	0.596	0.258	5.353	0.021	1.815	1.095-3.007
OD	APOE4	Model 1	0.856	0.260	10.859	< 0.001	2.353	1.414-3.915
	APOE4	Model 2	0.851	0.261	10.597	0.001	2.342	1.403-3.909
	APOE4	Model 3	0.951	0.268	12.621	< 0.001	2.588	1.532-4.374

CI: Cognitive impairment; OD: Olfactory dysfunction (olfactory threshold scores ≤ 2); OR: Odds ratio; CI: Confidence interval. Model 1 includes apolipoprotein E epsilon 4. Model 2 includes apolipoprotein E epsilon 4, age, and gender. Model 3 includes apolipoprotein E epsilon 4, age, gender, hypertension, hyperlipidemia, coronary artery disease, glycated hemoglobin and diabetes duration.

Higher HbA1c is associated with cognitive and olfactory impairments in T2DM patients

In patients with T2DM, HbA1c levels in peripheral blood are often elevated due to insulin resistance and impaired glucose regulation[31]. By single factor and multiple factor logistic regression analysis, we confirmed higher HbA1c levels significantly increase the risk of cognitive impairment (OR = 1.147, 95%CI: 1.046-1.258, P = 0.003, Table 3). After controlling for confounding factors, such as age, gender, hypertension, hyperlipidemia, coronary artery disease, APOE4 and diabetes duration, the association between HbA1c and cognitive impairment remained significant (OR = 1.189, 95%CI: 1.079-1.310, P < 0.001, Table 3). Meanwhile, HbA1c was significantly associated with OD (OR = 1.149, 95%CI: 1.032-1.280, P = 0.011, Table 3) adjusted for APOE4 genotype, age, gender, hypertension, hyperlipidemia, coronary artery disease and diabetes duration.

Table 3 Association of glycated hemoglobin with cognitive and olfactory function analyzed by logistic regression.

Outcome	Variable	Model	β	SE	Wald	P value	OR	95%CI
CI	HbA1c	Model 1	0.137	0.047	8.533	0.003	1.147	1.046-1.258
	HbA1c	Model 2	0.165	0.049	11.474	< 0.001	1.18	1.072-1.298
	HbA1c	Model 3	0.173	0.050	12.194	< 0.001	1.189	1.079-1.310
OD	HbA1c	Model 1	0.094	0.052	3.227	0.072	1.099	0.991-1.217
	HbA1c	Model 2	0.107	0.053	4.085	0.043	1.113	1.003-1.235
	HbA1c	Model 3	0.139	0.055	6.418	0.011	1.149	1.032-1.280

CI: Cognitive impairment; OD: Olfactory dysfunction (olfactory threshold scores ≤ 2); OR: Odds ratio; CI: Confidence interval. Model 1 includes glycated hemoglobin. Model 2 includes glycated hemoglobin, age, and gender. Model 3 includes glycated hemoglobin, apolipoprotein E epsilon 4, age, gender, hypertension, hyperlipidemia, coronary artery disease and diabetes duration.

Association of cognitive performance and olfactory function in T2DM patients

A positive correlation was observed between MMSE scores and OTS (Spearman r = 0.13, P = 0.0008, Supplementary Figure 1A). Further we divided the population into three groups according to different OTS ranges. Participants with the OTS of 0-2 had lower MMSE scores compared to those with 3-5 OTS and 6-8 OTS (P < 0.0001; P < 0.0001, Supplementary Figure 1B). Moreover, we divided the population into CN group and CI group and revealed that the CI group had lower OTS than CN group (P < 0.0001, Supplementary Figure 1C).

The interaction between APOE4 and HbA1c on cognitive and olfactory function

To further explore the interaction, the population was divided into a low HbA1c group and a high HbA1c group according to the criterion of HbA1c < 7[32]. MMSE scores were significantly lower in the APOE4 (APOE4 +) with high HbA1c group compared to the APOE4 with low HbA1c group (P = 0.0277, Figure 3A). These results suggest that APOE4 modifies the relationship between high HbA1c and cognitive function. Meanwhile, Figure 3B shows that the OTS was significantly lower in the APOE4 (APOE4 +) with high HbA1c group compared to the non-APOE4 (APOE4 -) with high HbA1c group (P = 0.017, Figure 3B). These findings suggest that high HbA1c in APOE4 carriers is linked to poorer olfactory function.

Figure 3 Effects of apolipoprotein E epsilon 4 genotype and glycated hemoglobin level on cognitive and olfactory function. A: Mini-mental state examination scores were significantly lower in the apolipoprotein E epsilon 4 (APOE4) + with high glycated hemoglobin (HbA1c) group compared to the APOE4 - with high HbA1c group (P = 0.0277); B: Olfactory threshold score was significantly lower in the APOE4 + with high HbA1c group compared to the APOE4 - with high HbA1c group (P = 0.017). HbA1c: Glycated hemoglobin; APOE4: Apolipoprotein E epsilon 4; MMSE: Mini-mental state examination.

Then, using stratified regression analyses we found in the APOE4 carrier group, HbA1c was significantly associated with both cognitive impairment (OR = 2.830, 95%CI: 1.801-4.446, P < 0.001, Table 4) and OD (OR = 1.479, 95%CI: 1.085-2.015, P = 0.013, Table 4), but not in non-APOE4 carrier group (P = 0.150, P = 0.136, Table 4). We similarly found that APOE4 was significantly correlated with cognitive impairment (OR = 2.599, 95%CI: 1.374-4.914, P = 0.003, Table 5) and OD (OR = 2.992, 95%CI: 1.550-5.774, P = 0.001, Table 5) in the high HbA1c group, but not in the low HbA1c group (P = 0.544, P = 0.122, Table 5).

Table 4 Association of glycated hemoglobin with cognitive and olfactory function analyzed by stratified regression.

Subgroup	Patients	Variable	P value	CI OR (95%CI)	P value	OD OR (95%CI)
APOE4 carriers	85	HbA1c	< 0.001	2.830 (1.801-4.446)	0.013	1.479 (1.085-2.015)
Non-APOE4 carriers	551	HbA1c	0.150	1.081 (0.972-1.201)	0.136	1.095 (0.972-1.234)

APOE4: Apolipoprotein E epsilon 4; CI: Cognitive impairment; OD: Olfactory dysfunction; OR: Odds ratio; CI: Confidence interval.

Table 5 Association of apolipoprotein E epsilon 4 with cognitive and olfactory function analyzed by stratified regression.

Subgroup	Patients	Variable	P value	CI OR (95%CI)	P value	OD OR (95%CI)
High HbA1c group	386	APOE4	0.003	2.599 (1.374-4.914)	0.001	2.992 (1.550-5.774)
Low HbA1c group	250	APOE4	0.544	0.747 (0.291-1.916)	0.122	2.083 (0.821-5.281)

HbA1c: Glycated hemoglobin; CI: Cognitive impairment; OD: Olfactory dysfunction; OR: Odds ratio; CI: Confidence interval.

To further support these observations, we analyzed the statistics through both single and multiple logistic regression (MLR) model. In the MLR model, the interaction between APOE4 status and HbA1c (APOE4 × HbA1c) on CI was statistically significant (OR = 2.667; P < 0.001, Table 6). This interaction suggests that the combination of APOE4 status and higher HbA1c is associated with a decline in cognition. This detrimental interaction remained significant even after controlling for covariates (OR = 2.714, P < 0.001; OR = 2.687, P < 0.001, Table 6). Similarly, in another MLR model, the interaction between APOE4 status and HbA1c on OD remains significant (OR = 1.489, P = 0.014, Table 6), indicates that the combination of APOE4 and elevated HbA1c levels is associated with a decline in olfactory function. Notably, this detrimental interaction remained significant even after controlling for covariates (OR = 1.490, P = 0.014; OR = 1.440, P = 0.027, Table 6).

Table 6 Interaction between apolipoprotein E epsilon 4 and glycated hemoglobin on cognitive and olfactory function.

Variable	Cognitive impairment		Olfactory dysfunction (OTS ≤ 2)
	OR (95%CI)	P value	OR (95%CI)	P value
Model 1 interaction (APOE4 HbA1c)	2.667 (1.725-4.122)	< 0.001	1.489 (1.084-2.047)	0.014
Model 2 interaction (APOE4 HbA1c)	2.714 (1.747-4.216)	< 0.001	1.490 (1.084-2.049)	0.014
Model 3 interaction (APOE4 HbA1c)	2.687 (1.725-4.187)	< 0.001	1.440 (1.042-1.990)	0.027

HbA1c: Glycated hemoglobin; APOE4: Apolipoprotein E epsilon 4; OR: Odds ratio; CI: Confidence interval; OTS: Olfactory threshold score. Model 1 includes apolipoprotein E epsilon 4, glycated hemoglobin and interaction. Model 2 includes apolipoprotein E epsilon 4, glycated hemoglobin, interaction, age, and gender. Model 3 includes apolipoprotein E epsilon 4, glycated hemoglobin, interaction, age, gender, hypertension, hyperlipidemia coronary artery disease and diabetes duration.

Joint associations and additive interactions of APOE4 with high HbA1c on the risk of cognitive impairment

The RERI and AP were calculated to examine the additive interaction between APOE4 and high HbA1c. The RERI for the APOE4 and high HbA1c group was 2.308 (95%CI: 0.178-4.439, Table 7), indicating a significant positive interaction. The AP for this interaction was 0.673 (95%CI: 0.371-0.975, Table 7), suggesting that the combined effect of APOE4 and high HbA1c contributed to 67% of the total risk for cognitive impairment. However, there was no additive interaction between APOE4 and high HbA1c on OD (Supplementary Table 2).

Table 7 Joint associations and additive interactions of apolipoprotein E epsilon 4 with high glycated hemoglobin on the risk of cognitive impairment.

Variables	Joint associations			Addictive interactions (APOE4 +)
	Number	APOE4 -	APOE4 +	RERI (95%CI)	AP (95%CI)
Overall CI	165
High HbA1c	111	1.291 (0.846-1.972)	3.364 (1.163-9.737)	2.308 (0.178-4.439)	0.673 (0.371-0.975)
Low HbA1c	54	0.774 (0.507-1.183)	0.297 (0.103-0.860)

HbA1c: Glycated hemoglobin; APOE4: Apolipoprotein E epsilon 4; CI: Cognitive impairment; 95%CI: 95% confidence interval; RERI: Relative excess risk due to interaction; AP: Attributable proportion.

DISCUSSION

APOE4 is strongly linked to an elevated risk of CI and OD. Interaction analysis further demonstrates that elevated HbA1c levels magnify the deleterious effects of APOE4, with poor glycemic control accelerating the deterioration of cognitive and olfactory functions among APOE4 carriers. These findings highlight the necessity of jointly considering genetic susceptibility and metabolic dysfunction when assessing neurobehavioral risks in T2DM populations.

Elevated HbA1c levels may impair cognitive function through multiple mechanisms. Chronic hyperglycemia induces dysfunction in endothelial cells, leading to mitochondrial oxidative stress, inflammatory activation, and disruption of the blood-brain barrier (BBB)[33,34]. BBB breakdown permits harmful plasma components to infiltrate the brain parenchyma, exacerbating neuronal injury and promoting neurodegeneration[34]. Furthermore, sustained hyperglycemia enhances neuronal glucotoxicity, impairs synaptic function, and accelerates cognitive decline[35,36]. HbA1c appears to further amplify APOE4-related neurotoxicity by acting as an effect modifier. Individuals carrying the APOE4 allele, already predisposed to heightened neuroinflammation and lipid dysregulation, may be particularly vulnerable to hyperglycemia-induced cerebral injury, resulting in more severe cognitive and olfactory deterioration[37-41]. Notably, elevated HbA1c may aggravate APOE4-related vulnerability through several converging mechanisms. Chronic hyperglycemia increases the accumulation of advanced glycation end products, which bind to receptor of advanced glycation end products and trigger oxidative stress and inflammation[42]. This process, together with reduced insulin signaling and lower insulin-degrading enzyme activity, promotes amyloid-β accumulation[43]. Additionally, hyperglycemia-induced activation of glycogen synthase kinase-3β enhances tau phosphorylation. In APOE4 carriers, these pathological changes may be amplified due to impaired lipid clearance and heightened neuroinflammatory responses[44], ultimately accelerating cognitive and olfactory decline. OD and mild cognitive impairment often precede the clinical onset of AD[45]. Early tau pathology typically affects structures such as the entorhinal cortex and hippocampus, regions essential for olfactory processing and memory formation, linking these early functional deficits to underlying neurodegenerative processes[46].

To our knowledge, this is the first study to investigate the amplifying effect of elevated HbA1c on APOE4-related cognitive and OD in patients with T2DM. While previous studies have separately linked poor glycemic control and APOE4 to neurodegenerative risk, their combined impact has been less well characterized[47]. Our results suggest that individuals carrying APOE4 and exhibiting poor glycemic control should be prioritized for early intervention. Strategies such as intensive glucose management, lifestyle modification, and regular cognitive and olfactory monitoring may provide cost-effective approaches to delay dementia progression in this high-risk population. Nevertheless, it should be noted that neurodegenerative changes in this study were assessed through functional evaluations rather than neuroimaging techniques such as positron emission tomography or magnetic resonance imaging, which constitutes a limitation.

Additionally, several limitations warrant consideration. First, the cross-sectional nature of this study limits causal inference regarding the relationship between HbA1c, APOE4, and neurobehavioral outcomes. Longitudinal studies are needed to clarify temporal and causal associations[48,49]. Second, the study population was primarily composed of Asian individuals, which may limit generalizability to other ethnic groups[50,51]. Third, although major confounding factors were adjusted in the regression models, including age, gender, hypertension, hyperlipidemia, coronary artery disease and diabetes duration, residual confounding from unmeasured factors like diet, education level, and physical activity cannot be fully excluded[52]. Future research should prioritize mechanistic investigations to clarify how hyperglycemia interacts with APOE4 in driving neuroinflammation, metabolic dysfunction, and cerebral injury[53-55]. In addition, randomized controlled trials are needed to determine whether intensive glycemic control can mitigate cognitive decline in APOE4 carriers[56]. These studies will be essential for informing precision-based strategies aimed at preventing or delaying neurodegeneration in high-risk diabetic populations.

CONCLUSION

We identify a significant association between the APOE4 allele and the incidence of CI and OD in individuals with T2DM. Stratified analyses indicated that cognitive and olfactory deficits were most pronounced in the high-HbA1c with APOE4 group. Multivariate regression confirmed a significant detrimental interaction between APOE4 and HbA1c on both cognitive and olfactory function. The observed interaction between APOE4 and HbA1c implies that poor glycemic control may intensify the impact of genetic vulnerability on cognitive and olfactory decline. These results support a more individualized strategy in T2DM management, where incorporating genetic risk factors into routine care could help identify patients at higher risk for neurodegenerative complications and guide earlier, targeted interventions. In this study, we propose that integrating glycemic management with conventional risk factor control may offer a more effective strategy for mitigating cognitive and olfactory impairments in T2DM populations, especially those with heightened genetic susceptibility, such as APOE4 carriers.

ACKNOWLEDGEMENTS

We are grateful to the patients in this study for their generous participation and their clinicians for including them in this study. We thank all the people who contributed to this program.