A Systematic Review and Meta-Analysis on the Prevalence of Mild Cognitive Impairment and Dementia in Mexico

Aging is a process that involves different aspects such as biological, psychological, and social features, which together lead to the development of age-related diseases [1, 2]. Among these, mild cognitive impairment (MCI) is a neurocognitive state that has an impact on some cognitive domains, such as memory and learning, complex attention, executive function, language, perceptual motor function, and social cognition [3, 4]. In contrast, dementia represents a set of major disorders of a progressive nature, characterized by a significant decrease in one or more cognitive domains, together with changes in personality and/or habitual behavior, which together compromise the functional abilities of the subject to carry out daily activities [5].

Age still remains one of the major risk factors for the onset of MCI and dementia [1], and it is predicted that by 2050, there will be 152.8 million instances of dementia worldwide [6]. Alzheimer’s disease is the most common type of dementia [7]. In Mexico, it has been reported that there is a prevalence of 16% for MCI and 8.6% for dementia [8‒10].

On a global scale, Alzheimer’s disease and other dementias were the seventh largest cause of mortality in 2019 [11]. Additionally, these conditions are the leading contributors to dependency among older people in Mexico [12]. Taking into account that the number of older adults is expected to grow, this could result in a significant increase in the burden on the Mexican healthcare system and family caregivers [13]. Some of the studies on the prevalence of these conditions in Mexico used different methods, tools, and populations with different health statuses. All these heterogeneous results may be a problem in identifying the true prevalence of MCI and dementia in Mexico. Besides, the estimated prevalence of these conditions has an enormous variety among countries and within regional studies [6, 14]. To our knowledge, in Mexico, there is not a systematic review or meta-analysis available that presents essential figures on the prevalence of these conditions. In this systematic review, we intend to access the maximum number of reports published on the topic and determine the prevalence of MCI and dementia in Mexican older adults.

This systematic review was performed according to Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines [15] (Fig. 1) and registered in the International Prospective Register of Systematic Reviews (PROSPERO) under reference number CRD42023425005. A systematic electronic search of the literature was conducted by four independent researchers (A.L.G.-S., V.M.-Z., D.G.G.-C., and E.S.-G.) from February to May 2023 using PubMed, Cochrane, Lilacs, Scielo, and ResearchGate databases, and the gray literature was performed on Google Scholar. This systematic review, which was reported in accordance with the PEO strategy, included population: older population from Mexico, exposure: subjects with MCI and dementia, and outcome: prevalence of MCI and dementia.

The search terms used were “dementia” or “Mild Cognitive Impairment” and “Mexico” and “prevalence.” The search strategy was developed on PubMed, and then the keywords were adapted to the different databases. These terms were used in English and Spanish.

After identification and exclusion of duplicates, three authors (V.M.-Z., E.S.-G., and D.G.G.-C.) screened the titles and abstracts. The full texts of selected studies were examined by all authors (A.L.G.-S., V.M.-Z., E.S.-G., D.G.G.-C., and A.A.-N.) regarding the eligibility criteria. The remaining articles that were not eliminated were assessed in full text, and then the final inclusion and exclusion criteria were applied. The eligibility criteria for the studies were (1) original studies (cross-sectional and epidemiological studies), (2) studies performed on subjects with dementia or MCI, (3) patients over 55 years of age, and (4) the Mexican population that lives in Mexico. The exclusion criteria were (1) the editor’s letter and congress abstracts, (2) studies that included the “MCI” or “dementia” terms, but no data on the prevalence were presented, (3) studies that identified dementia derived from neuropsychiatric diseases, (4) the estimation of the prevalence of dementia was obtained using computational learning methods, and (5) those studies that reported the prevalence of dementia using the same database (10/66 dementia research group). A total of 6 studies published dementia prevalence using the same database from 10/66 dementia research group; therefore, in order to avoid duplication of data on the prevalence, the last published manuscript of this research group was used for the meta-analysis. Publications in English and Spanish were included.

Data extraction from the selected articles was performed independently by three investigators (V.M.-Z., E.S.-G., and D.G.G.-C.) into a predefined database. Data were regarding the year of publication, the name of the first author, separating diagnoses of IMC or dementia, the total population, and the number of total cases according to sex, conditions, and instruments used. In the event of a discrepancy between V.M.-Z., E.S.-G., and D.G.G.-C., the final consensus was reached by another coauthor (A.L.G.-S. and A.A.-N.). The quality of the methodology used in each study was evaluated using the Newcastle-Ottawa Scale (NOS) [16] (online suppl. Table S1; for all online suppl. material, see https://doi.org/10.1159/000539560).

All statistical analyses were undertaken using the software RStudio version 1.0.153. The prevalence of MCI and dementia reported in the selected studies among healthy Mexican older adult populations was evaluated in cross-sectional, epidemiological, and pooled studies. Then, the prevalence of MCI and dementia among men and women was pooled using a random-effects model and afterward presented in a forest plot. Statistical heterogeneity was assessed using the I² index. The statistical program was also used to determine the responsible factors, such as diagnostic criteria and comorbidities, for the observed heterogeneity through meta-regression.

In total, 233 articles were identified. Additionally, 11 articles were found in the gray literature. Altogether, 46 articles were eliminated due to duplications in the search and 156 manuscripts were removed by title and abstract before screening. Altogether, 42 articles were reviewed in full text; only 33 were assessed for eligibility, and 17 were excluded. Finally, 16 studies were selected for the present review and meta-analysis (Fig. 1).

The years of publication of the studies ranged from 2004 to 2022. Among studies, 7 solely reported the prevalence of MCI [17‒23], 4 reported the prevalence of dementia [24‒27], and 5 identified the prevalence of MCI and dementia using the same study sample [24, 28‒31]. Two studies used population base data from surveys performed in different low- and middle-income countries [18, 32]; another study used data collected from some Latin American countries (LACs) [25]; and five studies used national databases from Mexico [17, 18, 22, 23, 30]. The rest of the studies were carried out in states in Mexico: Mexico City (n = 4), Nuevo Leon (n = 2), Tamaulipas (n = 1), Puebla (n = 1), Guanajuato (n = 1), Durango (n = 1), Morelos (n = 1), and Jalisco (n = 1). Thirteen of the 16 studies were carried out in urban areas of Mexico. The study population included subjects with different diagnoses, such as type 2 diabetes (T2D) (n = 2), chronic kidney disease (CKD) (n = 1), Sjögren’s syndrome (n = 1), and stroke diagnosis (n = 1). Cognitive assessment in subjects with MCI reported in the studies includes the Mini-Mental State Examination (MMSE) (n = 7), the National Institute on Aging-Alzheimer’s Association (n = 1), the Montreal Cognitive Assessment (MoCA) (n = 1), the Canadian criteria for vascular cognitive impairment (VCI) without dementia (n = 1), Algorithm (verbal fluency, Mini-Cog, and functionality) (n = 1), and the Cross-Cultural Cognitive Examination (n = 1). Furthermore, the cognitive assessments in older adults with dementia reported were MMSE (n = 4), Community Screening Instrument for Dementia (CSI-D) (n = 1), MoCA (n = 1), Cognitive Abilities Screening Instrument (CASI), Clinical Dementia Rating Scale (CDR) (n = 1), and Algorithm (verbal fluency, Mini-Cog, and functionality) (n = 1). A total of 20,978 subjects with MCI and 13,763 with dementia were analyzed in this review. Among all studies, subjects with MCI aged over 55 years and dementia subjects aged over 60 years (Table 1, 2).

A total of 12 studies on MCI prevalence were analyzed and plotted. Torres-Perez et al. [28] found the highest prevalence (53%) among patients with CKD in Puebla, followed by Arauz et al. [29] with 41% among older adults with a diagnosis of stroke, and Zavala-Rubio et al. [20] with a 31% among subjects with T2D. The lowest prevalence (4.3%) was reported among older adults from the metropolitan area of Guadalajara by Velazquez-Brizuela et al. [24]. Other studies reported a prevalence of MCI ranging from 6% to 31%. The overall prevalence of MCI was estimated from pooled information from 12 selected studies, and it was found to be 18% (95% CI 0.10–0.27). Analyses by sex showed a higher prevalence in women with 21% (95% CI 0.08–0.38) than in men with 18% (95% CI 0.06–0.33). In all analyses, a heterogeneity of >90% was found (Fig. 2).

A subanalysis was conducted, distinguishing studies that were carried out in particular groups or communities from studies that accessed prevalence from national surveys, in order to ascertain whether the prevalence of MCI was overstated among the chosen studies. Studies that employed national surveys had a prevalence of 10% (95% CI 0.06–0.14), but cross-sectional studies had a prevalence of 25% (95% CI 0.13–0.39), indicating that those national surveys might be underestimating the range of MCI prevalence, but cross-sectional studies included various health status backgrounds, which suggests an increase in the prevalence (Fig. 3).

A total of 9 studies reported the prevalence of dementia. Vega-Quintana Alejandro et al. [26] found the highest prevalence (30%) in the population that visited the geriatric clinic in Leon, Guanajuato. Prevalence of dementia ranged from 4 to 30%. The lowest prevalence (4%) was reported by Alanis-Niño et al. [31] among older adults from four different clinics. The overall prevalence of dementia was estimated from pooled information from 9 selected studies, and it was found to be 10% (95% CI 0.06–0.14). Analyses by sex showed a higher prevalence in women with a 14% (95% CI 0.05–0.25) than in men with a 10% (95% CI 0.04–0.17). In all analyses, a heterogeneity of >90% was found (Fig. 4).

Again, a subanalysis was conducted among dementia studies. Studies that employed national databases had a prevalence of 9% (95% CI 0.07–0.010); among the cross-sectional studies, a similar prevalence of 10% (95% CI 0.05–0.18) was found. In all analyses, a heterogeneity of >80% was found (Fig. 5).

Finally, a meta-regression analysis was performed to take into account some risk factors, such as comorbidities (T2D, CDK, CVD, hypertension, depression) for the examined population and the cognitive assessment used for diagnoses. In the univariate meta-regression analysis, having a CVD (p = 0.003), depression (p < 0.001), and hypertension (p = 0.008) were found to be significantly associated with heterogeneity in older adults with MCI. For subjects with dementia, there was statistical significance among the cognitive assessments used for diagnoses of MMSE (p < 0.001), depression (p < 0.001), and T2D (p = 0.005) (Table 3).

To the best of our knowledge, this is the first meta-analysis that estimates the pooled prevalence of MCI and dementia among older adults in Mexico. This meta-analysis included 16 studies performed among older adults living in Mexico and found an overall prevalence of 18% (95% CI 0.10–0.27) for MCI and 10% (95% CI 0.06–0.14) for dementia. This is similar to recent findings from a systematic review and meta-analysis performed on seven developing countries, including Mexico, in which the prevalence of dementia ranged from 8.6 to 9.5% [33], but greater compared to other regions of the world such as North America (5.3%), Asia (7.1%), and Europe (4.7%) [34]. In 2022, another meta-analysis on the prevalence of dementia in LACs reported some high variations; the prevalence obtained in this study is higher (10%) compared to Chile (4.47%), Argentina (8.22%), Cuba (8.46%), and Venezuela (7.93%) but lower in contrast to Brazil (12.36%), Peru (12.10%), Trinidad (24.13%), Colombia (19.51%), and the Dominican Republic (11.67%) [35].

Some of the included studies in this manuscript were done on patients with different comorbidities, such as CKD, stroke, and T2D, among others. This is important because, due to the lack of descriptions of dementia subtypes, we were not able to estimate prevalence according to subtypes. It has been described as a strong association of these comorbidities with vascular dementia and AD, especially. In Mexico, there are high frequencies of T2D, hypertension, and obesity, among others; this could be linked to having a higher prevalence of dementia compared to other regions of the world [36, 37].

Regarding the sex differences, in a recent systematic review of the prevalence of dementia among Latin American and Caribbean countries, higher rates were reported among women (8.97%) than men (7.26%) [35]. Our data showed a similar scenario since women (14%) had higher rates than men (10%). This is also in line with the findings of Juarez-Cedillo et al. [38], who found that among 6,204 older adults from Mexico City, women had a greater prevalence of dementia than men (15.3 vs. 12.5%, respectively). Women are considered to have a greater risk of developing dementia; some reasons have been described, such as that women live longer than men and are also more likely to suffer stress and depression [39, 40]. The role of the number of pregnancies, age of menopause, and levels of hormones is still under research [41, 42].

Compared to dementia, there is less evidence about the prevalence of MCI. The results obtained in this study showed a prevalence of MCI of 18% (95% CI 0.10–0.27) in older adults from Mexico. This is higher than the results reported in another systematic review that estimated a worldwide overall prevalence of MCI of 15.56% among community dwellers aged 50 years and older [43] and greater than 14.95% obtained from a meta-analysis of the overall prevalence rate for all-type MCI in LACs. In the same study, the evaluated countries, such as Brazil (5.76–17.4%) and Mexico (5.95–34.42%), showed a high range in the prevalence of MCI [44]. It was also found that MCI prevalence in women (21%) had a higher frequency than in men (18%) in accordance with international data that describe a higher prevalence of MCI in women than men and identified some risk factors associated with it, such as that women have less access to education and are prone to suffer more depression than men [45].

Among the included studies, the selected population was collected from urban and rural areas and different geriatric clinics; therefore, the population of older adults from Mexico is well represented in this study. Nevertheless, the majority of included studies were performed in the north and center of Mexico, and although some national studies were integrated into the analyses, the lack of cross-sectional studies in the south of Mexico is a limitation. The subanalyses performed in those studies based on national surveys (epidemiological) and cross-sectional designs showed no differences in the prevalence of dementia, but they were different for the MCI prevalence. The higher MCI prevalence in cross-sectional studies may be attributed to professional diagnosis because some older adults included were attending hospitals; therefore, they may receive more attention from clinicians, which could significantly raise the MCI detection rate.

Other reasons that could influence the estimation of the prevalence of MCI and dementia in the analyzed population from Mexico include having T2D, CVD, depression, and anxiety; all of these conditions have been described as associated factors with the risk of cognitive impairment and, later on, the development of dementia [46, 47]. The variables that were relevant to the estimation of the prevalence of dementia included MMSE (p < 0.001), depression (p < 0.001), and T2D (p = 0.005). MMSE was the instrument most applied in the vast majority of included studies, and according to our data, using MMSE did not affect the prevalence of MCI, but it was significant for the prevalence of dementia. Although the MoCA has been reported to have superior sensitivity and specificity for screening MCI, health professionals and researchers in Mexico are more familiar with the use of the MMSE, which is commonly used in dementia diagnosis [48]. Nevertheless, if this tool is used incorrectly, the percentage of the population with dementia may be overestimated. Moreover, education is an important factor in dementia; it has been reported that having less education is strongly associated with MMSE scores among Mexican Americans [49].

In conclusion, Mexican older individuals have a similar prevalence of dementia and MCI as reported by international data; nevertheless, the prevalence is higher than some LACs. In addition, Mexico has particular issues that must be resolved, such as the unreliability of epidemiologic data, the absence of dementia and MCI strategies, a lack of research in the southern regions of the country, and the high incidence of comorbidities (obesity, T2D, and CVD). Further research is necessary to determine strategies to reduce gaps related to the identification of these conditions in Mexico.

This study provides the first evidence and updated estimates of MCI and dementia prevalence in men and women in Mexico. First, Mexico has limited evidence on the epidemiology of MCI and dementia compared to other regions or countries, such as Europe and the USA. Further, in this study, the evidence from multiple studies and subanalyses was synthesized, although high heterogeneity is reported. Our findings provide some confidence in the robustness of the MCI and dementia prevalence estimates.

We are grateful to Julio Vega for helping us with the statistical analysis.

An ethics statement is not applicable because this study was based exclusively on published literature.

The authors declare no conflicts of interest with respect to the publication of this article.

The present study did not receive any specific research grant from public, commercial, or nonprofit agencies.

A.L.G.-S., V.M.-Z., E.S.-G., and D.G.G.-C.: conceptualization, methodology, and software. A.L.G.-S., V.M.-Z., E.S.-G., D.G.G.-C., G.C.-L., A.A.-N., and A.G.G.-D.: data curation, writing, and original draft preparation. V.M.-Z., E.S.-G., D.G.G.-C., and A.G.G.-D.: visualization and investigation. A.L.G.-S. and A.A.-N.: supervision. A.L.G.-S. and A.G.G.-B.: software and validation. A.L.G.-S., V.M.-Z., E.S.-G., D.G.G.-C., A.G.G.-D., and A.A.-N.: writing, reviewing, and editing. A.L.G.-S., V.M.-Z., E.S.-G., D.G.G.-C., A.G.G.-D., A.A.-N., and G.C.-L.: read and approved the final manuscript.

Valeria Magallón-Zertucheand Angel Gabriel Garrido-Dzib share first authorship.

The data used to support the findings of this study are included in the article. Further inquiries can be directed to the corresponding author.