Effect of poor glycemic control in cognitive performance in the elderly with type 2 diabetes mellitus: The Mexican Health and Aging Study

Type 2 Diabetes Mellitus (DM) rates and its complications are increasing faster than expected. Prevalence increased from 4.3 to 9.0% in men and 5.0 to 7.9% in women from 1980 to 2014. By year 2025, diabetes would exceed 700 million cases. Thus, 37.9% of the rising prevalence is due to population growth and ageing [1]. In our country, Mexican National Health and Nutrition Survey (ENSANUT) reported a DM prevalence of 27.4% in the geriatric population [2]. The increasing prevalence of diabetes will lead to a higher number of people with diabetes-related cognitive impairment. For this reason, efforts to control not only type 2 DM, but hypertension, depression and other risk factors are essential for dementia prevention around the world [3, 4].

Numerous reports have shown that patients with DM are at increased risk of Alzheimer’s disease (AD) dementia and vascular dementia [5]. It is currently accepted that worldwide type 2 DM increases 1.5 to 3 times the risk of dementia [6]. The Rotterdam study was one of the first population-based studies to examine the relationship between type 2 DM and cognitive impairment. It reported an increased risk of 1.9 (95% CI: 1.3–2.8) for this population [7]. In Mexico, a longitudinal study which included data from the Mexican Health and Aging Study (MHAS) reported a relative risk for cognitive impairment of 2.08 (95% CI: 1.59–2.73) among type 2 DM in elderly [8].

An optimal glycemic control is needed to prevent or reduce type 2 DM complications such as nephropathy, retinopathy, neuropathy and cognitive disorder. Although, epidemiologic studies have associated poor type 2 DM control with cognitive decline, clinical evidence about glycemic control goals for type 2 DM in elderly is contradictory and lacking. The American Diabetes Association (ADA) 2019 guidelines suggest a reasonable glycated hemoglobin (Hb_A1c) goal of < 7% for most patients and a less stringent < 8% for those with limited life expectancy or multiple comorbid conditions [9, 10]. The American Geriatric Society (AGS) recommends a Hb_A1c goal < 7.5% for healthy population; while for the complex or comorbid and functional-dependent patient the aim should be a Hb_A1c of < 8% [11]. According to these guidelines, Hb_A1c goals are based on two variables, which must be considered before a decision: cognitive and functional status. Thus, in both guidelines Hb_A1c goal recommendation is < 8% for subjects with mild to moderate cognitive impairment [9‐11].

ADA 2019 guideline recommendations on a Hb_A1c goal below or greater than 8% are lacking on making a clear reference on the risk of adverse effects and cognitive impairment progression. Studies, such as the Memory in Diabetes (MIND) substudy of the Action to Control Cardiovascular Risk Diabetes (ACCORD), which has been one of the most influential when concerning glycemic goals found no difference in cognitive outcomes with either Hb_A1c < 6% or 7.0–7.9 [12]. This study could provide different insights towards the appraisal of cognitive impairment through knowledge of type 2 DM, one of its main risk factors. Although, cultural variances could be noted as environmental factors carry a great weight on the disease; glycemic control is a parameter that could be standardized across countries. The aim of the present study was to determine the association between glycemic control and cognitive performance among Mexican rural and urban community-dwelling older adults analyzed in the MHAS round 2012.

From a total of 216 participants with type 2 DM were included in the analysis, the mean age was 68.11 (± 6.4) years, 57.9% were female, and the mean education level was 4.72 (± 3.9) years. Type 2 DM participants also had a high prevalence of hypertension (69%) and IHD (7.9%). Obesity was reported in 40% of the participants. Smoking history was present in 9.7% and alcohol history in 18.1%. The mean score for depressive symptoms was 4.76 (± 1.9) points. Mean Hb_A1c blood level value was 8.34 ± 2.05%. When divided by glycemic control groups, no differences were found in any of the variables analyzed (Table 1).

As shown in Table 2, were classified as having low (35.1%), intermediate (31.4%) or high (33.3%) cognitive dysfunctions. Subjects in the lower cognitive performance group were older (69.7 ± 6.6 years vs 65.86 ± 5.18 p < .001) and had a lower educational level (2.5 ± 2.6 years vs 7.44 ± 4.15; p < .000) when compared to the high cognitive performance participants. No significant differences between the three groups were found for any of the other clinical variables, except for alcoholism, which was more frequent in the intermediate cognitive performance group (27.9%, p = .023).

When compared to the intensive glycemic control group those with poor glycemic control had a borderline association with worse cognitive performance in the unadjusted regression model. After adjusting for confounding variables, the multinomial regression analysis showed that the presence of Hb_A1c ≥ 8% (poor glycemic control) was associated with low (Odds Ratio (OR) 3.17, 95% CI = 1.17–8.60, p = .024), and intermediate (OR 3.23, 95% CI = 1.27–8.20, p = .014) cognitive performance. This trend was not found for the standard glycemic control (Hb_A1c 7–7.9%) (Table 3).

In our study of community-dwelling older Mexican adults with type 2 DM, a poor glycemic control (Hb_A1c ≥ 8%) was associated with worse cognitive performance when compared to intensive control group. The uncontrolled type 2 DM group had a positive association with overall low cognitive performance, while the standard controlled population (Hb_A1c 7–7.9%) did not show an association.

A high Hb_A1c level (> 10%) is associated with an increased risk of all type dementia (HR 1.20, 95% CI 1.07–1.35) [21]. However, studies that analyze glycemic control, specifically, Hb_A1c levels ≥8% and their association with cognitive performance, are scarce. After a sub-analysis, a US prospective study of 5099 participants showed an association between Hb_A1c levels (≥8% and 7–7.9%) and mild cognitive impairment (MCI) measured by proxy (Hazzard Risk (HR) 1.89, CI 95% 1.14–3.14, p < 0.05 and HR 1.65, CI 95% 1.13–2.42, p < 0.01, respectively) in an older adult population [22]. Our results support these association, as an Hb_A1c ≥ 8% was associated with worse cognitive performance.

Multiple studies have identified that an intensive vs a standard glycemic treatment had no beneficial or detrimental effects on cognition [12, 23, 24]. A meta-analysis involving five studies, that included 24,297 participants, found that neither intensive (Hb_A1c 6.0–7.0%) nor standard (Hb_A1c 7.1–8.0%) glycemic control, when compared to each other, had significant cognitive decline rates (SMD = 0.02; 95% CI = − 0.03 to 0.08) [25]. One of the studies mentioned above included, where guideline recommendations are based mostly is the ACCORD-MIND study. Two thousand nine hundred seventy-seven participants (aged 55–80 years) with higher Hb_A1c levels (> 7.5%) were randomly assigned to an intensive treatment goal (Hb_A1c < 6.0%) or a standard strategy (Hb_A1c: 7.0–7.9). In this North American trial, as previously, authors found no difference when comparing cognitive outcomes between groups after 40-week treatment, establishing a greater amplitude of therapy goal for patients [12]. In our study, we were not able to find an association between standard glycemic control (Hb_A1c 7–7.9%) and low cognitive performance, supporting the previously reported analysis.

Several studies have evaluated the impact of type 2 DM on cognition; however, methodological differences are noted [25]. Data from the English Longitudinal Study of Ageing (ELSA) showed in 5189 participants, a longitudinal association between Hb_A1c levels and a rate of change in cognitive scores, where 1 mmol/mol increment in Hb_A1c was significantly associated with increased rate of decline in global cognitive z scores (− 0.0009 SD/ year, 95% CI -0.0014, − 0.003, p 0.002) [26]. Our study provides an association of glycemic control levels and global cognitive performance supporting the data mentioned above; unlike other studies, additional specific Hb_A1c goals are analyzed and type 2 DM is not taken as a single entity.

After several studies of diabetes and cognitive impairment, authors have gone as far as proposing the possibility of type 3 of DM. Chronic hyperglycemia contributes to conditions such as inflammation, accumulation of advanced glycation end products, and oxidative stress, which in turn lead to cognitive impairment [27]. Studies have shown that persons with DM experience a progressive cognitive decline, particularly characterized by a lower psychomotor speed and alterations in cognitive domains such as attention and executive function [28]. Disruption in glucose metabolism leads to lower cognitive dysfunctions through different mechanisms; a) GLUT transporter altered sensitivity, b) insulin resistance, and c) vascular dysfunction. Chronic hyperglycemia is a phenomenon that inhibits brain autoregulation since GLUT transporters diminish their function in order to protect neurons from an increased glucose influx. When glucose is restored to a normal level, GLUT transporters fail to recover, causing an absence of intraneuronal glucose in a process called neuroglycopenia [29]. Insulin resistance could lead to apoptosis by disruption of a secondary pathway; insulin receptor phosphorylation which disrupts long term potentiation, thus increasing inflammation and generating oxidative stress [30]. Vascular homeostasis is also affected by type 2 DM. The presence of atherosclerotic plaques, endothelial dysfunction, increased shear stress, inflammation, impaired vasodilation, and increased vasoconstriction, are some of the mechanisms that lead to vascular injury. The theories presented above, often converge into a type 3 DM diagnosis [31, 32].

The combination of factors seen in the type 2 DM population (hypertense, obese, IHD) mirrors the population’s clinical characteristics that physicians are set to treat in the present and near future. As type 2 DM control is one of the strongest modifiable comorbidities that affect brain function, hypertension and obesity are equally relevant risk factors to target. Since currently there are no therapies to cure dementia, the treatment of modifiable risk factors should be emphasized [2, 3].

Some other factors may impact the glycemic control in older adults. In our study, type 2 DM participants with low cognitive performance were older and had a lower educational level. Studies have shown that age is the most important and non-reversible risk factor for the development of cognitive dysfunctions [33]. In elderly adults, pharmacologic management with multiple drugs leads to a low treatment adherence, given a higher number of side effects. Besides, non-pharmacologic treatments such as a diet and lifestyle interventions are generally less effective, since the modification of eating and physical activity habits is usually a difficult task [34, 35]. Education has previously been described as a protective factor for cognitive impairment, as higher education allows the development of “cognitive reserve” and a lower educational level is associated with a 5.6 greater risk of dementia [36]. Also, a lower educational level could influence glycemic control. Adherence to treatment and lifestyle recommendations, disease complications, and awareness are some of the variables in which a lower educational level, over time, has a negative impact [37, 38].

Our study has several limitations. The cross-sectional nature of this study is a major limitation for making cause-effect statements. Also, since the MHAS data was gathered through a survey, many subjects were excluded because there was a lack of biomarker availability. In addition, only a small sample of the cohort was analyzed as we sought to study fulfill the diagnosis of diabetes in community dwelling residents; this could lead to low statistical power and limiting the external validity of the results. Regarding type 2 DM, diagnosis was limited to one Hb_A1c measurement. A history of duration in years and information on previous control strategies was not obtained. Nevertheless, our study has several strengths. The MHAS is a large representative sample of community older adults; considering that control of cognitive impairment risk factors is a primary prevention strategy that should be prioritized.

Our study shows a Hb_A1c value ≥8% in older adults with diabetes is associated with a worse cognitive performance. Adequate glucose control should also be promoted in older adults.