1. Introduction

The size of the elderly population is increasing worldwide. The United Nations project that this increase will intensify in the coming decades, mostly due to the rise in average life expectancy. The number of elderly people in the world (more than 60 years old) will increase by 56% in the next 15 years and the “oldest old” (more than 80 years old) will triple in number by 2050. This rapid demographic ageing will increase the prevalence of disease and disability, with a particular emphasis expected for the impairment of cognitive functions.

Loss of memory, learning difficulties and a decrease in the ability to concentrate on a task characterizes cognitive impairment in the elderly. This ranges from mild deficits, which are not clinically detectable, to dementia. There are many different etiologies of cognitive impairment, ranging from vascular conditions to neuronal degeneration and stroke. Cognitive impairment leads to a decrease in the life quality of elders and increases the risk of dementia and mortality. Additionally, it has significant social consequences, resulting in the loss of autonomy and independence and leading to an increased need for permanent caregivers and assistance by health services.

There is a scarcity of studies reporting the prevalence of cognitive impairment at a given time point, as well as of the incidence of newly diagnosed cases. Both of these measures help to identify disease trends within a population, giving information not only on how common the condition is but also at what speed new cases are emerging. This information is essential to assess the overall burden of disease and to develop hypotheses regarding the causes and factors that increase the risk of disease. Good quality scientific data on cognitive impairment are needed, both to identify groups at risk of developing cognitive changes at an early stage and to identify the optimum time at which to implement preventive and corrective measures. A better understanding of cognitive impairment and its lifetime course is needed to define and implement strategies to both prevent initial cognitive impairment and either stop or delay its progression towards dementia once established. In 2015, the COSMIC studies (Cohort Studies of Memory in an International Consortium) was published, which used data from cohort studies in several countries around the world, applied uniform criteria to harmonize data, and reported the prevalence of cognitive impairment. Our systematic review complements the COSMID study as it includes information on the prevalence as well as incidence of cognitive impairment by considering the latest studies published after 2015, and it includes information from Portugal.

The free-form research question we used to drive this research was “What is the worldwide cognitive impairment prevalence and incidence in older adults, as reported by observational studies?” The PICO structure to our research question was as follows: Population—older adults; Intervention—observational studies; Comparison—worldwide; Outcome—prevalence and incidence of cognitive impairment. Our objective was to review the global epidemiological data on cognitive impairment and to derive prevalence and incidence estimates for this nosological entity.

2. Materials and Methods

We conducted a systematic search of the PubMed electronic database on 4 January 2019. We did not seek unpublished data. We considered all studies published until 4th January 2019 for the analysis. The search details were “cognitive impairment”[All Fields] AND ((“epidemiology”[Subheading] OR “epidemiology”[All Fields] OR “prevalence”[All Fields] OR “prevalence”[MeSH Terms]) OR (“epidemiology”[Subheading] OR “epidemiology”[All Fields] OR “incidence”[All Fields] OR “incidence”[MeSH Terms])) AND (elders[All Fields] OR older[All Fields]). For the first evaluation, we imported a total of 3645 references to Endnote. In order to increase the information for Portugal, we conducted a second related search on the same day. The search details were “cognitive impairment”[All Fields] AND ((“epidemiology”[Subheading] OR “epidemiology”[All Fields] OR “prevalence”[All Fields] OR “prevalence”[MeSH Terms]) OR (“epidemiology”[Subheading] OR “epidemiology”[All Fields] OR “incidence”[All Fields] OR “incidence”[MeSH Terms])) AND (elders[All Fields] OR older[All Fields] OR (“aged”[MeSH Terms] OR “aged”[All Fields])) AND (“portugal”[MeSH Terms] OR “portugal”[All Fields]). A total of 53 references were imported and added to the database. We did not limit the search results by the language of publication. We eliminated duplicates (8 references).

References were verified using a two-step process. For the first step, articles were selected based on information available in the title and/or abstract. The full text of the selected articles was read in the second step to determine the agreement of each article with the adopted criteria. We included reports with epidemiological data on cognitive impairment (CI), mild cognitive impairment (MCI) and cognitive impairment not dementia (CIND). These terms are used differentially but overlap to some extent and there was no standard rule that would allow us to draw a clear distinction between them, therefore they were assumed to refer broadly to the same entity and treated as such.

The exclusion criteria were as follows: non-original full-length articles (e.g., a systematic review, guidelines, meta-analysis, review, comment, editorial, note, meeting abstract); case-reports; non-human/in vitro; non-elderly population (studies conducted in populations described as consisting exclusively or partially of children, adolescents or adults); language (papers not written in English, Spanish, French or Portuguese were excluded); treatment/intervention/diagnostic studies; no data on cognitive impairment (studies that did not report prevalence or incidence of cognitive impairment); cognitive impairment in specific subgroups, such as patients with dementia, depression, HIV and Parkinson’s disease; studies including the oldest old only (over 85 years old); institutionalized participants in hospitals, clinics or nursing homes (to obtain data for older people present in the general population and not report on a special population).

We collected data regarding the participants’ age, sample size, diagnostic methods used, world region, and estimates of prevalence and/or incidence of cognitive impairment. We provide Supplementary Material with the characteristics of the cohort studies.

We assessed the quality of the studies included using the National Heart, Lung, and Blood Institute (NHLBI) Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies [10], categorized as >80% yes = “Good”, 60–80% yes = “Fair”, and <60% yes = “Poor” to assess the internal validity and risk of bias for each study and the overall quality. We took into account the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) 2009 Checklist (Table S4) and the Quality Assessment tool from the NHLBI to verify methodological quality and the quality of the included studies (Table S3).

Tables with the results of the Quality Assessment Tool and the score from the PRISMA 2009 Checklist are available as Supplementary Material.

2.1. Statistical Analysis

The included studies differed in several parameters such as participant inclusion age, sample size, diagnostic methods used and world region (Europe, Asia, North America, South America and Australia). Due to the large variance of reported prevalence and incidence estimates, we divided data into more homogeneous groups, and within-group comparisons were made with Kruskal–Wallis for independent samples test and the median quartiles using Tukey’s Hinges method. We used non-parametric statistical techniques due to the asymmetrical distribution of the sample. The statistical analyses were performed using SPSS^® version 21. Prevalence was reported as a percentage and incidence is reported in cases per 1000 person-years, while the median (25–75 percentile) are reported for both.

2.2. Data Analysis

For prevalence data, we subdivided papers into three groups according to inclusion age: (1) participants aged from 50 to 59 years (mean = 52.93 years; SD = 2.50 years)—14 papers; (2) participants aged from 60 to 69 years old (mean = 63.28 years; SD = 2.49 years)—57 papers; and (3) participants aged 70 years or older (mean = 75.11 years; SD = 3.02 years)—9 papers. Regarding sample size, 26 studies had fewer than 1000 participants (mean = 504.19 participants; SD = 222.12 participants); 22 had between 1001 and 2500 participants (mean = 1695.82 participants; SD = 408.78 participants); 18 had between 2501 and 5000 participants (mean = 3614.33 participants; SD = 519.19 participants) and 14 studies had more than 5000 participants (mean = 7314.50 participants; SD = 1682.56 participants). According to the diagnostic method used to identify cognitive impairment, 9 studies accounted for the presence of cognitive complaints by either the patient or family, the absence of dementia and a neurological evaluation; 62 used only standard neurological tests to determine cognitive impairment (including but not restricted to Mini Mental State Examination (MMSE), Montreal Cognitive Assessment (MOCA), and the Short Portable Mental Status Questionnaire); and 8 simultaneously used both of the previously described methods. We analyzed data by world region: 25 studies in Europe; 13 studies in North America; 3 studies in South America; 35 studies in Asia; 2 studies in Africa; and 2 studies in Australia (Table S1).

To estimate the incidence of cognitive impairment, we divided the papers according to the same criteria: people aged: 50–59 years (mean = 55.33 years; SD = 0.58 years)—3 papers; 60–69 years old (mean = 64.50 years; SD = 2.38 years)—4 papers; ≥70 years (mean = 73.75 years; SD = 2.87 years)—4 papers. In terms of sample size, 2 studies had fewer than 1000 participants (mean = 608.50 participants; SD = 215.60 participants); 5 studies had 1001–2500 participants (mean = 1701.80 participants; SD = 479.51 participants); 3 studies had 2501–5000 participants (mean = 3102 participants; SD = 698.69 participants); and one study included 7166 participants. As for the diagnostic method used to identify cognitive impairment, three studies accounted for the presence of a patient or family report of cognitive complaints, the absence of dementia and a neurological evaluation; five studies used validated neurological tests; and three studies used both of the previously described methods. There were five studies carried out in both Europe and North America, and one was carried out in Asia (Table S2).

3. Results

Of the 3690 potentially relevant articles found, 296 were selected based on the information present in the title and/or abstract (step 1); after reading the full text, 85 were selected as relevant (step 2). Of these, 74 papers provided information only on cognitive impairment prevalence, 5 papers only provided information on cognitive impairment incidence and 6 papers provided information for both parameters (Figure 1). The quality assessment tool from the NHLBI was used to assess the methodological quality of the included studies; 77 studies had an overall rating of “good”, eight were rated “fair” and none were rated “poor”. Based on these findings, no papers were excluded from the analysis.

Figure 1. Flow chart summary of the literature search.

3.1. Prevalence of Cognitive Impairment

The prevalence of cognitive impairment (CI) reported in the 80 studies ranged from 5.1% to 41.0% (median = 19.0%; 25th percentile = 12.0%; 75th percentile = 24.90%) (Table 1 and Figure 2).

Figure 2. Prevalence of cognitive impairment reported by published papers, which are grouped by world region (95% confidence intervals were obtained from papers or calculated from the data presented).

Table 1. Summary of the prevalence of cognitive impairment reported by 80 studies included in analysis.

Grouping papers according to inclusion age (50–59 years old, 60–69 year old, and ≥70 years old), the reported prevalence ranges from 6.5% to 34% (median = 12%; 25th percentile = 9.6%; 75th percentile = 17.65%) in the first group, 5.1% to 37.5% (median = 20.1%; 25th percentile = 14.2%; 75th percentile = 24.7%) in the second group and from 11.6% to 41% (Med = 19%; 25th percentile = 15%; 75th percentile = 29.90%) in the last group.

When grouping and analyzing the effect of sample size, we divided the studies into four groups based on the number of participants they had (<1001, 1001–2500, 2501–5000 and >5000). The reported prevalence in the first group ranged from 5.3% to 37.5% (median = 22.75%; 25th percentile = 14.9%; 75th percentile = 31.4%), from 7.7% to 41% in the second group (median = 15.95%; 25th percentile = 11.60%; 75th percentile = 28.50%), from 6.5% to 32.7% in the third group (Med = 13.75%; 25th percentile = 9.60%; 75th percentile = 21.30%), and from 5.1% to 27% (median = 20.24%; 25th percentile = 18.8%; 75th percentile = 24.1%) in the last group.

Regarding cognitive impairment diagnostic methods, in the presence of cognitive complaints, the absence of dementia and with a neurological evaluation, the prevalence of cognitive impairment (CI) was from 9.6% to 33% (median = 15.4%; 25th percentile = 11.3%; 75th percentile = 23.4%). When only standardized neurological tests were used (MMSE, MOCA, Short Portable Mental Questionnaire, etc.), the prevalence of CI ranged from 5.1% to 41% (median = 18.9%; 25th percentile = 12.2%; 75th percentile = 24.7%). When both methods (neurologist evaluation, patient or family complaints and standardized neurologic tests) were used, the estimated prevalence of CI ranged from 10.7% to 34% (median = 21.30%; 25th percentile = 12.0%; 75th percentile = 28.90%).

With regard to the world region where data were collected, in Europe the prevalence of cognitive impairment ranges from 5.1% to 41% (median = 12.1%; 25th percentile = 9.94%; 75th percentile = 23.9%); in North America, it ranged from 7.1% to 28.3% (median = 20.1%; 25th percentile = 19%; 75th percentile = 24.70%); in South America, it ranged from 24.3% to 37.5% (median = 34%; 25th percentile = 29.15%; 75th percentile = 35.75%). In Asia the prevalence ranges from 6.5% to 37% (median = 19.44%; 25th percentile = 13.25%; 75th percentile = 25.55%). In Africa, CI prevalence ranged from 18.4% to 33% (median = 25.7%; 25th percentile = 18.4%; 75th percentile = 33%) and in Australia from 7.7% to 33.3% (median = 20.5%; 25th percentile = 7.7%; 75th percentile = 33.3%). No statistically significant differences within groups were found in the reported CI prevalence when grouping papers according to any of these variables.

3.2. Incidence of Cognitive Impairment

The incidence of cognitive impairment reported by the 11 included studies ranged from 22 to 215 per 1000 person-years, with a median incidence of 56.50 per 1000 person-years (25th percentile = 41.77; 75th percentile = 76.50) (Table 2 and Figure 3).

Figure 3. Incidence of cognitive impairment reported by the 11 included studies, which are grouped by world region (the 95% confidence intervals were obtained from papers or calculated with the data presented).

Table 2. Summary of cognitive impairment incidence as reported by the 11 studies included.

Grouping papers according to the age of the participants (50–59 years old, 60–69 years old, and ≥70 years old) yielded incidence estimates ranging from 22 to 41.77 per 1000 person-years (median = 30.7 per 1000 person-years (25th percentile = 26.35; 75th percentile = 36.24) in the first group, from 51.45 to 215 per 1000 person-years (median = 71.11 per 1000 person-years (25th percentile = 58.44; 75th percentile = 145.98) in the second group and from 47.19 to 76.50 per 1000 person-years (median = 58.45 per 1000 person-years (25th percentile = 51.84; 75th percentile = 68.45) in the last group. Statistically significant differences were found in the incidence of CI across age categories (p = 0.035).

Taking into account the number of participants included in the studies (<1001, 1001–2500, 2501–5000 and >5000), the reported CI incidence ranged from 41.77 to 60.4 per 1000 person-years (median = 51.09 per 1000 person-years (25th percentile = 41.77; 75th percentile = 60.40) for group 1, from 22 to 76.8 (median = 56.50 per 1000 person-years (25th percentile = 47.19; 75th percentile = 76.50) for group 2, from 30.70 to 65.42 (median = 51.45 per 1000 person-years (25th percentile = 41.08; 75th percentile = 58.44) for group 3 and the only study with more than 5000 participants reported an incidence of 215 cases per 1000 person-years. No statistically significant differences were found between the groups.

According to the cognitive impairment diagnostic methodology used, studies that evaluated the presence of cognitive complaints and the absence of dementia, and included a neurological evaluation, reported a CI incidence ranging from 41.77 to 215 per 1000 person-years (median = 76.5 per 1000 person-years (25th percentile = 59.14; 75th percentile = 145.75). The studies that used neurological tests (MMSE, MOCA, Short Portable Mental Questionnaire) reported an incidence from 22 to 76.80 per 1000 person-years (median = 51.45 per 1000 person-years (25th percentile = 30.7; 75th percentile = 60.4). The studies that used both methods reported a CI incidence ranging from 47.9 to 65.42 per 1000 person-years (median = 56.50 per 1000 person-years (25th percentile = 51.82; 75th percentile = 60.96). There were no statistically significant differences among these groups.

In Europe, the incidence of cognitive impairment ranges from 30.70 to 76.50 per 1000 person-years (median = 56.5 per 1000 person-years (25th percentile = 51.45; 75th percentile = 76.5). In North America, this ranged from 41.8 to 215 per 1000 person-years (median = 60.4 per 1000 person-years (25th percentile = 47.19; 75th percentile = 65.42) and in Singapore the incidence was reported as 22 per 1000 person-years. We did not find statistically significant differences among groups.

One study reported an incidence of 215 per 1000 person-years, which is 11.85 standard deviations over the mean of the other ten studies. Excluding that study from the data analysis changes the reported median incidence to 53.97 per 1000 person-years (25th percentile = 39.0; 75th percentile = 68.19). In the group of participants with the minimum inclusion age (60–69 years old), the median incidence was 65.42 per 1000 person-years (25th percentile = 58.44; 75th percentile = 71.11), statistically significant differences were found within the group (p = 0.05). In the group of studies that evaluated the presence of cognitive complaints, the absence of dementia included a neurological evaluation, the median incidence was 59.14 per 1000 person-years (25th percentile = 41.77; 75th percentile =76.50), and we did not find statistically significant differences within the group. In the group of studies from North America, the median incidence was 53.80 per 1000 person-years (25th percentile = 44.48; 75th percentile = 62.91), and we did not find statistically significant differences within the group.

4. Discussion

4.1. Methodological Considerations

Our objective was to review the global epidemiological data to derive prevalence and incidence estimates for cognitive impairment. We included reports with three different constructs: cognitive impairment (CI), mild cognitive impairment (MCI) and cognitive impairment not dementia (CIND). Besides the different names, we could not distinguish consistently between them, so all were assumed to refer broadly to the same entity and were treated as such.

We expected a significant degree of heterogeneity among studies, so data were aggregated by age group, study sample size, diagnostic methods used, and world region.

Despite all the studies having elderly people as the study focus, the minimum inclusion age for participants diverged greatly between studies and could bias the results. For example, higher estimates of cognitive impairment could be a result of a more elderly sample, as several different studies reported an increase in cognitive impairment prevalence with increasing age. In our study, we found that in terms of the incidence of cognitive impairment, studies that had an inclusion age starting at 60 years had a median incidence higher than those with an inclusion age over 70 years old, and the difference was statistically significant. The lower incidence at higher ages might imply that the rate of conversion from healthy cognition to cognitive impairment might reach a plateau at some point after 60 but before 70 years of age, considering that cognitive impairment is a milder form of decline that, at older ages, can progress to dementia.

Regarding the sample size of the studies, the main objective was to compare the results of studies with hundreds of participants to others with thousands of participants, and we found no significant quantitative differences among these.

While there were no statistically significant differences regarding the method used to identify cases of cognitive impairment, for studies of the prevalence of cognitive impairment, the median prevalence of cognitive impairment was higher for the method that used a neurological evaluation paired with neurological tests. With regard to studies on the incidence of cognitive impairment, the median was higher for neurological evaluations. In the future, we aim to further explore the optimum methods to identify cognitive impairment and develop recommendations that will lead to a better and more accurate diagnosis.

We aggregated data by world region to examine the geographic differences that may influence the epidemiology. In terms of the prevalence of cognitive impairment, there were no statistically significant differences. However, in terms of incidence of cognitive impairment, in Europe it was lower than in North America; this could be due to cultural or genetic effects, or a combination of both, with an impact on cognitive impairment severity and progression.

Regarding cognitive impairment incidence studies, the Mejia-Arango study from Mexico reports an incidence that is 11.85 standard deviations above the mean of the other ten studies (Figure 3). Although the median is not greatly affected by outliers, this study alone increases the median reported incidence from 53.97 per 1000 person-years (without) to 56.50 per 1000 person-years (with). Several procedural characteristics set this study apart. Briefly, it was the only study that used a version of the Cross-Cultural Cognitive Examination (CCCE) as a cognitive decline screening approach, which might point to a culturally diverse background of the participants. To those unable to complete the questionnaire due to limitations of language or health, a brief version of the informant questionnaire of Cognitive Decline in the Elderly (IQCODE) was applied. Additionally, 32.70% of the participants in the study were illiterate, and education years have a meaningful impact on cognitive impairment frequency. The high cultural heterogeneity implicit in this choice of instruments and reported illiteracy prevalence raises doubts over whether this incidence estimate is valid for Mexicans in general or highly influenced by a specific sub-population within Mexico. Due to its methodological particularities and high incidence estimate, we analyzed all of the incidence variables excluding the Mejia-Arango paper; however, there were no statistically significant differences within groups with or without it.

4.2. Future Directions

The ability to rely on tests to identify the incipient cases of cognitive impairment with high accuracy (sensitivity and specificity) is crucial for population studies and population interventions, as it is both impractical and cost-prohibitive to have a specialist neurological assessment of large numbers of unaffected individuals. We believe that our results highlight the need for the development of a consensus regarding the best initial markers of cognitive impairment, the development of more reliable tests to detect incipient cognitive impairment cases, both with better reliability and more universally applied cut-off points that account for the factors known to influence cognitive declines such as age and education. Equally, to examine age-related cognitive decline, studies should restrict the inclusion age to 60 years old, as this is the threshold for older people as defined by the WHO. We expect that the implementation of these measures would lead to a more reliable and valid diagnosis of cognitive impairment and a more accurate global view of the prevalence and incidence of cognitive impairment in older people, which is fundamental to the delineation of public health measures aimed at this risk group. Results from this systematic review may inform public health decisions through accurate regional estimates of cognitive impairment for the definition of adequate measures regarding modifiable risk factors, particularly in people over 60 years old. Detection and treatment of diabetes and hypertension, reduction in levels of obesity, smoking cessation, increased physical activity, and better education should be public health priorities. We also provide some suggestions for methodologies on further cognitive impairment studies, as there are significantly different social and economic structures in different world regions, and even in different countries within the same world region, it would be essential to conduct studies aimed explicitly at understanding cognitive impairment in the specific region.

4.3. Strengths and Limitations

The strengths of this study were its global view of the epidemiological data and the use of studies which reported on the general population, while excluding those that reported on people within the healthcare system or with a diagnosed underlying disease etiology. There are some methodological limitations and a risk of different types of bias associated with this study. Among these, we should mention publication bias, the selective reporting of data within studies and the incomplete retrieval of research. To try to reduce the risk of other biases, we aggregated papers into more methodologically homogeneous groups and compared the reported data within each group. By not restricting the initial search by publication language, we have an estimate of the size of our language bias. By only including reports written in Portuguese, English, Spanish or French, we excluded four studies. Reporting bias in published studies due to the selective reporting of subgroups of a population or the exclusion of non-significant outcomes measured by the study is a possibility that should be borne in mind. Additionally, we did not consider data on cognitive impairment etiology, as the main objective of this study was to review worldwide estimates of the prevalence and incidence of cognitive impairment in older adults regardless of etiology. Another limitation was that there was no pairwise review.

5. Conclusions

This systematic review reports that the global prevalence of cognitive impairment ranged from 5.1% to 41% with a median of 19.0%. The incidence of cognitive impairment ranged from 22 to 76.8 per 1000 person-years, with a median of 53.97 per 1000 person-years. We did not find statistically significant effects besides participant age in the studies sampled. For future studies, we propose the homogenization of the definition of cognitive impairment and the importance of the standardized cut-off scores of cognitive tests to compare different studies.

Introduction

The global population of older adults continues to grow. Projections from the United Nations indicate this trend will accelerate in the coming decades, primarily due to increased life expectancy. The number of individuals over 60 years old is expected to rise by 56% in the next 15 years, while the population over 80 years old is projected to triple by 2050. This rapid demographic shift is anticipated to increase the occurrence of diseases and disabilities, particularly affecting cognitive functions.

Cognitive impairment in older adults is characterized by memory loss, learning difficulties, and reduced ability to focus. This condition can range from mild, undetectable deficits to severe dementia. Various factors contribute to cognitive impairment, including vascular issues, neurodegeneration, and stroke. Such impairment diminishes the quality of life for older individuals and heightens their risk of developing dementia and increased mortality. Furthermore, it carries significant societal implications, leading to decreased autonomy and independence, and a greater demand for long-term caregivers and healthcare services.

Research on the prevalence (how common the condition is at a given time) and incidence (the rate at which new cases emerge) of cognitive impairment has been limited. Such data are crucial for understanding disease trends within a population, assessing the overall burden of the condition, and developing hypotheses about its causes and risk factors. High-quality scientific data on cognitive impairment are necessary to identify at-risk groups early and determine the optimal timing for preventive and corrective interventions. A deeper understanding of cognitive impairment and its progression throughout a person’s life is vital for defining and implementing strategies to prevent its onset or to halt or delay its progression to dementia once it has developed. In 2015, the COSMIC studies, which harmonized data from various international cohort studies, provided insights into cognitive impairment prevalence. The current systematic review complements the COSMID study by including more recent data published after 2015, covering both prevalence and incidence, and incorporating information from Portugal.

The guiding research question for this study was to determine the worldwide prevalence and incidence of cognitive impairment in older adults, as reported by observational studies. This question focused on older adults (Population), involved observational studies (Intervention), aimed for a global scope (Comparison), and measured the prevalence and incidence of cognitive impairment (Outcome). The objective was to review global epidemiological data on cognitive impairment and to derive estimates for its prevalence and incidence.

Materials and Methods

A systematic search was performed on the PubMed electronic database on January 4, 2019, considering all studies published up to that date. No unpublished data were sought. The primary search terms included "cognitive impairment," "epidemiology," "prevalence," "incidence," "elders," and "older." This initial search yielded 3645 references. To enhance information for Portugal, a secondary search was conducted on the same day, adding "Portugal" as a search term, which added 53 references to the database. The search was not limited by language of publication, and eight duplicate references were removed.

References were reviewed using a two-step process. First, articles were selected based on their titles and/or abstracts. In the second step, the full text of selected articles was read to confirm their alignment with the inclusion criteria. Reports providing epidemiological data on cognitive impairment (CI), mild cognitive impairment (MCI), and cognitive impairment not dementia (CIND) were included. These terms were considered to broadly refer to the same condition due to overlapping definitions and the absence of a standardized distinction. Exclusion criteria included non-original articles (e.g., reviews, meta-analyses), case reports, non-human or in vitro studies, studies on non-elderly populations, articles not written in English, Spanish, French, or Portuguese, treatment/intervention/diagnostic studies, studies lacking cognitive impairment data, studies on specific subgroups (e.g., dementia, depression, HIV, Parkinson’s), studies exclusively on the oldest old (over 85 years), and studies involving institutionalized participants.

Data collected included participant age, sample size, diagnostic methods, world region, and estimates of cognitive impairment prevalence and/or incidence. The quality of included studies was assessed using the National Heart, Lung, and Blood Institute (NHLBI) Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies. Studies were categorized as “Good” (over 80% adherence), “Fair” (60–80% adherence), or “Poor” (under 60% adherence) to evaluate internal validity and bias risk. The PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) 2009 Checklist was also considered to verify methodological quality. Overall, 77 studies were rated “good,” eight were rated “fair,” and none were rated “poor,” leading to no exclusions based on quality.

Statistical analyses utilized non-parametric techniques due to the asymmetrical distribution of the sample. Within-group comparisons were performed using Kruskal–Wallis for independent samples test and median quartiles via Tukey’s Hinges method. All analyses were conducted using SPSS® version 21. Prevalence was reported as a percentage, and incidence as cases per 1000 person-years, with both parameters reporting the median (25th–75th percentile).

For prevalence data analysis, papers were categorized into three age groups (50–59, 60–69, and 70+ years), four sample size groups (<1001, 1001–2500, 2501–5000, and >5000 participants), three diagnostic method groups (cognitive complaints with neurological evaluation, standard neurological tests only, and both methods), and six world regions (Europe, North America, South America, Asia, Africa, and Australia). For incidence data, papers were similarly divided by age (50–59, 60–69, and ≥70 years), sample size (<1001, 1001–2500, 2501–5000, and >5000 participants), diagnostic method (cognitive complaints with neurological evaluation, validated neurological tests, and both methods), and world region (Europe, North America, Asia).

Results

From 3690 potentially relevant articles, 296 were initially selected based on title and/or abstract. After full-text review, 85 articles were deemed relevant. Of these, 74 papers reported only cognitive impairment prevalence, five reported only incidence, and six reported both. All included studies exhibited good or fair methodological quality according to the NHLBI assessment tool.

The reported prevalence of cognitive impairment across 80 studies ranged from 5.1% to 41.0%, with a median of 19.0%. When papers were grouped by participant age, prevalence ranged from 6.5% to 34% in the 50–59 age group, 5.1% to 37.5% in the 60–69 age group, and 11.6% to 41% in the 70+ age group. Analysis by sample size, diagnostic method, and world region also showed variations in prevalence ranges, but no statistically significant differences were found in reported cognitive impairment prevalence when grouping papers according to any of these variables.

The incidence of cognitive impairment, reported by 11 studies, ranged from 22 to 215 per 1000 person-years, with a median of 56.50 per 1000 person-years. When grouped by participant age, incidence estimates varied, with statistically significant differences found across age categories (p = 0.035). Studies with an inclusion age starting at 60 years showed a higher median incidence compared to those starting over 70 years. Analysis by sample size and diagnostic method revealed varying ranges, but no statistically significant differences were found between these groups. Regionally, incidence ranged from 30.70 to 76.50 per 1000 person-years in Europe and 41.8 to 215 per 1000 person-years in North America, with one study from Singapore reporting 22 per 1000 person-years. No statistically significant differences were found among regional groups.

One study, from Mexico, reported an incidence of 215 per 1000 person-years, which was significantly higher than the other ten studies. Excluding this outlier changed the median incidence to 53.97 per 1000 person-years. Despite its impact on the median, re-analyzing all incidence variables without this paper did not yield statistically significant differences within the various demographic or methodological groups, except for the 60-69 age group where a statistically significant difference was found (p = 0.05). The outlier study’s distinct characteristics, such as using a version of the Cross-Cultural Cognitive Examination (CCCE) and a high percentage of illiterate participants, raised questions about its generalizability.

Discussion

The objective of this systematic review was to evaluate global epidemiological data and derive prevalence and incidence estimates for cognitive impairment. The study included reports using the terms cognitive impairment (CI), mild cognitive impairment (MCI), and cognitive impairment not dementia (CIND), treating them as broadly similar entities due to their overlapping definitions and lack of a universally accepted distinction. Anticipating significant heterogeneity among studies, data were aggregated and analyzed by participant age, study sample size, diagnostic methods employed, and world region.

While all studies focused on older adults, the minimum age for participant inclusion varied considerably. Higher cognitive impairment estimates were expected in older samples, as numerous studies report an increase in prevalence with advancing age. This study found that, for cognitive impairment incidence, studies with a minimum inclusion age of 60 years had a higher median incidence than those with an inclusion age over 70 years, and this difference was statistically significant. This finding might suggest that the rate of progression from healthy cognition to cognitive impairment could plateau between 60 and 70 years of age, given that cognitive impairment is a milder form of decline that can progress to dementia in later years. Regarding sample size, no significant quantitative differences were observed when comparing results from studies with hundreds versus thousands of participants. Although no statistically significant differences were found concerning diagnostic methods, studies using both neurological evaluations and tests reported a higher median prevalence for cognitive impairment, and neurological evaluations alone yielded a higher median incidence.

Geographic differences were examined by aggregating data by world region. While no statistically significant differences were found in cognitive impairment prevalence, the incidence in Europe appeared lower than in North America. This disparity could be influenced by cultural or genetic factors, or a combination thereof, affecting the severity and progression of cognitive impairment. The outlier incidence reported by the Mejia-Arango study in Mexico warranted particular attention due to its unusually high estimate and unique methodological characteristics, such as the use of the Cross-Cultural Cognitive Examination (CCCE) and the high prevalence of illiteracy among participants. Despite its substantial impact on the overall median incidence, its exclusion did not alter the lack of statistical significance within other analytical groups.

Future research should focus on developing a consensus regarding the best initial markers of cognitive impairment and more reliable tests for detecting incipient cases. These tests should have improved reliability and universally applied cut-off points that account for known influencing factors like age and education. To accurately examine age-related cognitive decline, studies should consider restricting participant inclusion to individuals aged 60 years and older, aligning with the World Health Organization’s definition of older people. Implementing these measures could lead to more reliable and valid diagnoses of cognitive impairment, providing a more accurate global understanding of its prevalence and incidence in older adults. This understanding is fundamental for shaping public health interventions targeting this at-risk demographic. The findings from this systematic review can inform public health decisions by providing accurate regional estimates of cognitive impairment, thereby facilitating the development of appropriate measures concerning modifiable risk factors. Priorities for public health should include the detection and management of diabetes and hypertension, reduction of obesity, smoking cessation, promotion of physical activity, and enhancement of educational attainment. Given the significant social and economic variability across regions and countries, region-specific studies are also crucial for a tailored understanding of cognitive impairment.

The strengths of this study include its global perspective on epidemiological data and its focus on studies from the general population, excluding those from healthcare systems or specific disease etiologies. Methodological limitations and potential biases, such as publication bias, selective data reporting within studies, and incomplete research retrieval, should be acknowledged. Efforts were made to mitigate the risk of other biases by grouping papers into more methodologically homogeneous categories for comparison. The decision not to restrict the initial search by publication language provided an estimate of potential language bias, with four studies excluded due to language barriers. Reporting bias in published studies, resulting from selective reporting of subgroups or non-significant outcomes, remains a possibility. Additionally, the study did not consider data on cognitive impairment etiology, as its primary goal was to review global prevalence and incidence estimates. A further limitation was the absence of a pairwise review.

Conclusions

This systematic review indicates that the global prevalence of cognitive impairment ranges from 5.1% to 41%, with a median of 19.0%. The incidence of cognitive impairment ranges from 22 to 76.8 per 1000 person-years, with a median of 53.97 per 1000 person-years (excluding a significant outlier). Beyond participant age, no statistically significant effects were found among the variables sampled in the included studies. For future research, a standardized definition of cognitive impairment and consistent, standardized cut-off scores for cognitive tests are recommended to facilitate meaningful comparisons across different studies.

Introduction

The global population of older adults continues to grow significantly, a trend projected to accelerate in the coming decades, largely due to increased life expectancy. The number of individuals over 60 years old is expected to rise by 56% in the next 15 years, and those over 80 years old are projected to triple by 2050. This rapid aging of the population is anticipated to increase the occurrence of diseases and disabilities, with a particular focus on the decline of cognitive functions.

Cognitive impairment in older adults is characterized by memory loss, difficulties with learning, and reduced ability to concentrate. This condition can range from very subtle deficits that are not easily detected clinically to more severe forms like dementia. Various factors can cause cognitive impairment, including vascular conditions, nerve degeneration, and stroke. Such impairment reduces an individual's quality of life, increases the risk of developing dementia and mortality, and has significant social consequences. These consequences include a loss of independence and a greater need for ongoing care and health services.

Current research lacks comprehensive studies that report both the existing number of cases (prevalence) and the rate of new cases (incidence) of cognitive impairment at specific times. Both measures are crucial for identifying disease trends within a population, providing information on how common a condition is and how quickly new cases appear. This data is essential for understanding the overall impact of cognitive impairment and for forming theories about its causes and risk factors. High-quality scientific data on cognitive impairment is necessary to identify individuals at risk early on and to determine the best times for preventive and corrective actions. A deeper understanding of cognitive impairment and its progression over a lifetime is needed to create and implement strategies that can prevent initial impairment or stop or slow its progression to dementia once it has begun. The COSMIC studies, published in 2015, used data from international cohort studies with uniform criteria to report cognitive impairment prevalence. The current systematic review builds upon the COSMIC study by including more recent data published after 2015, adding information on incidence, and incorporating data specifically from Portugal.

The main research question guiding this study was to determine the worldwide prevalence and incidence of cognitive impairment in older adults as reported by observational studies. This research specifically focused on older adults as the population of interest, observational studies as the intervention, a worldwide comparison, and the prevalence and incidence of cognitive impairment as the outcome. The primary goal was to review global data on cognitive impairment and to estimate its prevalence and incidence.

Materials and Methods

A systematic search of the PubMed electronic database was performed on January 4, 2019, including all studies published up to that date. No unpublished data was sought. The initial search used terms related to "cognitive impairment," "epidemiology," "prevalence," "incidence," and "elders" or "older." An additional search was conducted to specifically include data from Portugal. Duplicate entries were removed from the combined results. Articles were then screened in a two-step process: first by title and abstract, then by reading the full text of selected articles. Reports containing epidemiological data on cognitive impairment (CI), mild cognitive impairment (MCI), and cognitive impairment not dementia (CIND) were included. These terms were considered to broadly refer to the same condition due to their overlapping nature and lack of clear standard distinctions.

Studies were excluded if they were not original full-length articles (e.g., reviews, guidelines, meta-analyses), case reports, non-human or in vitro studies, or involved populations exclusively or partially composed of children, adolescents, or non-elderly adults. Papers not written in English, Spanish, French, or Portuguese were also excluded, as were treatment, intervention, or diagnostic studies. Studies that did not report prevalence or incidence data for cognitive impairment, those focusing on specific subgroups (e.g., patients with dementia, depression, HIV, or Parkinson’s disease), studies including only the oldest old (over 85 years), or those with participants from institutionalized settings (hospitals, clinics, nursing homes) were also excluded to ensure the data reflected the general older population. Data collected included participants' age, sample size, diagnostic methods, world region, and estimates of prevalence and/or incidence. The quality of included studies was assessed using the National Heart, Lung, and Blood Institute (NHLBI) Quality Assessment Tool, categorized as “Good,” “Fair,” or “Poor,” to evaluate internal validity and bias risk. Adherence to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) 2009 Checklist was also verified.

Due to the significant differences among the included studies, such as participant age, sample size, diagnostic methods, and geographic regions, the data was organized into more consistent groups for analysis. Non-parametric statistical techniques were employed because the data distribution was asymmetrical. Prevalence was reported as a percentage, and incidence as cases per 1000 person-years, with both measures also presented with their median and interquartile ranges. For prevalence analysis, papers were grouped by participant age (50–59, 60–69, and 70+ years), sample size (under 1000, 1001–2500, 2501–5000, and over 5000 participants), diagnostic method (cognitive complaints/neurological evaluation, neurological tests only, or both), and world region (Europe, North America, South America, Asia, Africa, and Australia). Incidence data was similarly grouped by age, sample size, diagnostic method, and region.

Results

From an initial pool of 3690 potentially relevant articles, 296 were selected based on title and abstract, and 85 were ultimately deemed relevant after full-text review. Of these, 74 studies reported only cognitive impairment prevalence, 5 reported only incidence, and 6 reported both. Methodological quality assessment using the NHLBI tool rated 77 studies as "good" and eight as "fair," with no studies rated "poor," leading to the inclusion of all selected papers in the analysis.

The reported prevalence of cognitive impairment across 80 studies ranged from 5.1% to 41.0%, with a median of 19.0%. When studies were grouped by the minimum participant age, prevalence estimates varied: 6.5% to 34% (median 12%) for those aged 50–59, 5.1% to 37.5% (median 20.1%) for those 60–69, and 11.6% to 41% (median 19%) for those 70 or older. Sample size groupings showed prevalence ranging from 5.3% to 37.5% for studies with fewer than 1001 participants and 5.1% to 27% for those with over 5000 participants, with median values varying between these groups. Diagnostic methods also showed variation, with prevalence ranging from 9.6% to 33% when based on cognitive complaints and neurological evaluation, 5.1% to 41% using only neurological tests, and 10.7% to 34% when both methods were combined. Regionally, prevalence varied from 5.1% to 41% in Europe (median 12.1%) to 24.3% to 37.5% in South America (median 34%). No statistically significant differences were found in reported cognitive impairment prevalence across any of these grouping variables.

The incidence of cognitive impairment, based on 11 studies, ranged from 22 to 215 cases per 1000 person-years, with a median of 56.50 cases per 1000 person-years. When grouped by participant age, incidence estimates varied significantly (p = 0.035). Studies with participants aged 50–59 years reported incidence from 22 to 41.77 per 1000 person-years (median 30.7), those aged 60–69 reported 51.45 to 215 (median 71.11), and those 70 or older reported 47.19 to 76.50 (median 58.45). Sample size and diagnostic methods did not show statistically significant differences in incidence rates. Regionally, incidence ranged from 30.70 to 76.50 in Europe and 41.8 to 215 in North America, with one study from Singapore reporting 22. No statistically significant differences were observed across regions. One study from Mexico reported an exceptionally high incidence of 215 per 1000 person-years, which was an outlier. Removing this study from the analysis changed the median incidence to 53.97 per 1000 person-years, although this exclusion did not lead to new statistically significant differences within groups previously without them. The Mexican study's high incidence may be influenced by its methodological particularities, such as the use of the Cross-Cultural Cognitive Examination and a high prevalence of illiteracy among participants, factors known to impact cognitive assessment.

Discussion

This systematic review aimed to provide global estimates of cognitive impairment prevalence and incidence. The study included various terms for cognitive impairment, such as CI, MCI, and CIND, as a clear distinction between them was not consistently possible in the available literature. Heterogeneity among studies was expected due to differences in participant characteristics, methodologies, and geographic locations, so data were analyzed by age group, sample size, diagnostic methods, and world region.

The minimum inclusion age of participants notably affected findings for cognitive impairment incidence, with studies starting at age 60 showing a higher median incidence than those starting at age 70. This may suggest that the rate of developing cognitive impairment might reach a plateau after age 60 but before age 70, possibly indicating a progression to dementia in older age groups. While sample size did not significantly affect quantitative differences in results, the choice of diagnostic method, while not statistically significant, did show a higher median prevalence when neurological evaluations were combined with neurological tests. Geographic differences in incidence were observed between Europe and North America, possibly due to cultural or genetic factors. An outlier study from Mexico reported an unusually high incidence, potentially influenced by its unique assessment tools and high participant illiteracy, raising questions about its general applicability.

Future research should focus on establishing a consensus for defining cognitive impairment and developing more reliable tests with standardized cut-off points that account for factors like age and education. To better understand age-related cognitive decline, studies should consistently define "older people" using thresholds such as the World Health Organization's guideline of 60 years old. Implementing these measures would lead to more accurate diagnoses and a clearer global understanding of cognitive impairment prevalence and incidence, which is essential for developing public health strategies. These strategies should address modifiable risk factors like diabetes, hypertension, obesity, smoking, physical inactivity, and education, particularly for individuals over 60. Region-specific studies are also needed to account for diverse social and economic structures.

The strengths of this study include its comprehensive global perspective and its focus on the general population, excluding individuals within healthcare systems or with specific diagnosed diseases. However, certain limitations exist, such as the potential for publication bias, selective reporting of data, and incomplete retrieval of research. Efforts were made to mitigate bias by grouping papers into methodologically homogeneous categories. The initial search included multiple languages, though only papers in Portuguese, English, Spanish, or French were ultimately included, leading to the exclusion of four studies. The study also did not investigate the causes of cognitive impairment. Additionally, a pairwise review process was not employed.

Conclusions

This systematic review determined that the global prevalence of cognitive impairment ranged from 5.1% to 41%, with a median of 19.0%. The incidence of cognitive impairment ranged from 22 to 76.8 per 1000 person-years, with a median of 53.97 per 1000 person-years (excluding one significant outlier). Participant age was the only variable that showed statistically significant effects on incidence across the studies analyzed. For future research, a standardized definition of cognitive impairment and consistent, universally applied cut-off scores for cognitive tests are crucial for enabling more meaningful comparisons across different studies.

Introduction

The global population of older adults is growing. The United Nations predicts this growth will speed up in the coming decades, mainly because people are living longer. The number of individuals over 60 years old is expected to increase by 56% in the next 15 years. Those over 80 years old are projected to triple in number by 2050. This rapid aging of the population is expected to lead to more diseases and disabilities, especially problems with thinking and memory skills, known as cognitive impairment.

Cognitive impairment in older adults includes memory loss, difficulty learning, and trouble concentrating. These issues can be very slight, making them hard to detect, or severe, like dementia. Various causes can lead to cognitive impairment, such as blood vessel problems, nerve damage, or strokes. This condition reduces an individual's quality of life and raises the risk of developing dementia or even death. It also has major societal impacts, as it can lead to a loss of independence and a greater need for caregivers and health services.

High-quality research on cognitive impairment is needed to identify individuals at risk early on and to determine the best time for prevention and treatment. A deeper understanding of cognitive impairment throughout a person's life is necessary to create strategies that prevent it from starting or slow its progress toward dementia once it has begun. This review builds on previous research, including a 2015 study, by adding more recent data on both how common cognitive impairment is (prevalence) and how often new cases appear (incidence), and specifically including information from Portugal. The main goal was to examine worldwide data on cognitive impairment to determine how common it is and how often new cases develop.

Materials and Methods

The study involved a systematic search of the PubMed database on January 4, 2019, looking for studies on cognitive impairment, epidemiology, prevalence, incidence, and older adults. A second search was performed to specifically gather more information related to Portugal. A total of 3645 initial references were found, with an additional 53 from the Portugal-focused search. Duplicate entries were removed.

References were reviewed in two steps: first, by title and abstract, then by reading the full text of selected articles. The study included reports with data on cognitive impairment (CI), mild cognitive impairment (MCI), and cognitive impairment not dementia (CIND). These terms were treated as broadly referring to the same condition due to their overlapping definitions.

Studies were excluded if they were not original research (e.g., reviews, case reports), involved non-human subjects, focused on non-elderly populations, were not written in English, Spanish, French, or Portuguese, or did not report on cognitive impairment prevalence or incidence. Studies of specific subgroups (like those with dementia, depression, or Parkinson's disease), or those involving only the very oldest adults (over 85) or institutionalized individuals, were also excluded to ensure the data reflected the general older population.

Information collected from each study included participants' age, sample size, diagnostic methods used, and the world region where the study took place, along with estimates of cognitive impairment prevalence and incidence. The quality of the included studies was evaluated using the National Heart, Lung, and Blood Institute (NHLBI) Quality Assessment Tool. Studies were rated as "Good," "Fair," or "Poor" to assess their reliability and potential for bias. No studies were excluded based on this quality assessment, as most were rated "Good."

Statistical analysis was performed using SPSS version 21. Due to the wide differences among studies in terms of age, sample size, diagnostic methods, and region, the data were grouped into more similar categories for comparison. Prevalence was reported as a percentage, and incidence was reported as cases per 1000 person-years. Medians and percentile ranges were used because the data were not evenly distributed. For prevalence, studies were grouped by participant age (50-59, 60-69, or 70+ years), sample size, diagnostic method, and world region. Incidence data were similarly divided for analysis.

Results

Out of 3690 articles initially identified, 85 were found to be relevant for the review after a two-step screening process. Seventy-four of these studies focused on the prevalence of cognitive impairment, five on incidence, and six reported on both. A quality assessment tool was used, and 77 studies were rated "good," eight "fair," and none "poor," meaning no studies were excluded based on quality.

The prevalence of cognitive impairment reported across 80 studies ranged from 5.1% to 41.0%, with a typical rate of 19.0%. When studies were grouped by participants' age, the typical prevalence varied: 12% for those aged 50-59, 20.1% for ages 60-69, and 19% for those 70 and older. Different sample sizes, diagnostic methods, and world regions (Europe, North America, South America, Asia, Africa, Australia) were also analyzed, but no statistically significant differences in prevalence were found between these groups.

For the incidence of cognitive impairment, 11 studies reported rates ranging from 22 to 215 cases per 1000 person-years, with a typical rate of 56.50 cases per 1000 person-years. When studies were grouped by the age of participants, statistically significant differences in incidence were observed. Studies focusing on individuals aged 60-69 showed a higher typical incidence (71.11 per 1000 person-years) compared to those aged 50-59 (30.7 per 1000 person-years) or 70 and older (58.45 per 1000 person-years).

No significant differences in incidence were found when studies were grouped by sample size, diagnostic methods used (e.g., cognitive complaints plus neurological evaluation versus neurological tests only), or world region (Europe, North America, Asia). One study from Mexico reported a very high incidence (215 per 1000 person-years), which notably affected the overall typical rate. However, even when this outlier study was excluded, the general findings and lack of other statistically significant differences remained consistent.

Discussion

The study aimed to provide overall estimates of cognitive impairment (CI) prevalence and incidence, treating related terms like mild cognitive impairment (MCI) and cognitive impairment not dementia (CIND) as the same condition due to their overlapping definitions. Significant differences were expected among studies, so data were analyzed based on participant age, sample size, diagnostic methods, and world region.

While studies primarily focused on older adults, variations in the minimum age of participants could affect results. For instance, the incidence of cognitive impairment was significantly higher in studies that began including participants at age 60 compared to those starting at age 70 or older. This might suggest that the rate of developing cognitive impairment slows down between ages 60 and 70, as milder forms of decline may progress to dementia in very old age. Sample size and diagnostic methods did not show statistically significant differences in prevalence or incidence across groups, although combining neurological evaluations with tests generally led to higher typical prevalence rates. Geographic differences in incidence were observed, with Europe showing lower rates than North America, potentially due to cultural or genetic factors.

One study from Mexico reported an unusually high incidence rate. This study had unique characteristics, including using a cross-cultural assessment and a high percentage of participants who were illiterate, which could have influenced its results. While this outlier increased the overall typical incidence, its exclusion did not change the general findings or the lack of other significant differences among groups.

For future research, there is a clear need for a shared definition of cognitive impairment and more consistent testing methods, including standardized cut-off scores that consider factors like age and education. Such improvements would lead to more reliable diagnoses and a more accurate global picture of cognitive impairment. This information is crucial for public health efforts, especially for adults over 60, focusing on risk factors like diabetes, high blood pressure, obesity, smoking, physical inactivity, and education.

The study's strengths include its global scope and focus on the general older population, excluding those already in healthcare systems or with specific diseases. However, it also has limitations, such as the potential for publication bias (where studies with positive results are more likely to be published), selective reporting within studies, and the exclusion of studies not published in specific languages (Portuguese, English, Spanish, or French). The review also did not explore the causes of cognitive impairment, focusing solely on its occurrence rates.

Conclusions

This systematic review found that the global prevalence of cognitive impairment ranged from 5.1% to 41%, with a typical rate of 19.0%. The incidence of new cognitive impairment cases ranged from 22 to 76.8 per 1000 person-years, with a typical rate of 53.97 per 1000 person-years. Aside from participant age, no other factors in the studies showed statistically significant effects on these rates. Future studies should aim for a consistent definition of cognitive impairment and use standardized scoring for cognitive tests to allow for better comparison of findings.

Introduction

The number of older people around the world is growing. Experts believe this increase will speed up in the coming years because people are living longer. The number of people over 60 will go up by more than half in the next 15 years. People over 80 will triple by 2050. This fast growth in the number of older people means there will be more sickness and health problems, especially with thinking skills.

Thinking problems in older adults mean trouble with memory, learning new things, and focusing. These problems can be slight, or they can be serious, like dementia. There are many reasons for thinking problems, such as issues with blood flow or changes in brain cells. These problems make life harder for older adults and raise the chance of getting dementia or other serious illnesses. They also mean older adults may need more help and care.

Not enough studies have looked at how common thinking problems are at one time, or how often new cases appear. Knowing these numbers helps doctors understand health trends and how big a problem thinking issues are. It also helps them guess what causes these problems. Good information is needed to find people at risk early and to start ways to prevent or fix these problems. We need to better understand thinking problems and how they change over time. This will help create plans to stop them from starting or slow them down if they do. A study in 2015 gathered information on how common thinking problems are around the world. This new study adds to that by including newer research since 2015, and also includes information from Portugal.

The main question for this study was: "How common are thinking problems and how often do new cases appear in older adults around the world, based on studies that just observe people?" The study looked at older adults across the world to find out how common these problems are and how often they begin.

Materials and Methods

Researchers looked for studies in a large online database called PubMed on January 4, 2019. They looked at all studies published up to that date. They searched for terms like "thinking problems," "how common," and "older people." They found many studies. To get more information about Portugal, they did a second search on the same day, adding "Portugal" to their search words. They did not limit their search to studies in English only.

They checked the studies in two steps. First, they looked at the titles and summaries of the articles. Then, they read the full articles of the ones that looked promising. They included studies that reported on thinking problems, mild thinking problems, and thinking problems that are not dementia. These terms often mean similar things, so they were all treated as the same type of thinking problem.

Studies were not included if they were not original research (like reviews or guides), were about animals, were about children or young adults, were not in English, Spanish, French, or Portuguese, were about treatments, did not have information on how common or new thinking problems were, were about thinking problems in specific groups (like people with dementia or Parkinson's), were only about the very oldest people (over 85), or were about people in hospitals or nursing homes. This was done to get information about older people living in the general public.

Information was gathered about the age of the people in the studies, how many people were in each study, how thinking problems were found, and what part of the world the study was done in. The quality of each study was checked using a special tool. This helped make sure the studies were well done and fair.

To make sense of the results, studies were put into groups. These groups were based on things like the age of the people in the study, how many people were in the study, how thinking problems were checked, and what part of the world the study came from (like Europe, Asia, North America). This was done because the studies were very different from each other. Simple math methods were used to compare these groups.

Results

Out of nearly 3,700 articles found, 296 were chosen after looking at their titles and summaries. After reading the full articles, 85 were selected as useful. Of these, 74 studies gave information on how common thinking problems are, 5 studies gave information on how often new cases appear, and 6 studies gave both. Most of the studies chosen were of good quality, and none were left out because of poor quality.

How common thinking problems are (called "prevalence") was looked at in 80 studies. The number ranged from about 5% to 41% of older adults, with an average of about 19%. When studies were grouped by the age of the people, their size, how thinking problems were found, or the part of the world they were from, there was no real difference in how common thinking problems were. For example, in Europe, thinking problems were found in 5% to 41% of people. In North America, the range was 7% to 28%.

How often new thinking problems appeared (called "incidence") was looked at in 11 studies. The number ranged from 22 to 215 new cases for every 1000 people each year, with an average of about 56 new cases. When studies were grouped by the age of the people, there was a real difference in how often new cases appeared. Studies that started with people aged 60-69 found more new cases than studies that started with people aged 70 or older. However, grouping studies by their size, how thinking problems were found, or the part of the world they were from, did not show real differences in how often new cases appeared.

Discussion

The main goal of this study was to find out how common thinking problems are and how often new cases start in older adults around the world. The study included different types of thinking problems because they often overlap and are hard to tell apart.

Studies were grouped by age, study size, how problems were found, and world region because there were many differences between them. Even though all studies focused on older people, the starting age of people in the studies varied a lot. The study found that when studies included people as young as 60, there were more new cases of thinking problems than when studies only included people 70 or older. This might mean that the rate of new thinking problems slows down after age 60, as these problems can then lead to dementia in very old age. The size of the study, whether it had hundreds or thousands of people, did not change the results much.

The way thinking problems were found also did not lead to big differences in how common they were. But for how often new cases appeared, using a doctor's checkup was linked to higher numbers than using only standard tests. Thinking problems were less common in Europe than in North America. This could be due to differences in culture or genes. One study from Mexico had a very high number of new cases. This might be because of how the study was done or the people it included (for example, many people in that study could not read). However, removing this study from the data did not change the overall findings much.

In the future, it is very important to have clear and standard ways to find early thinking problems. This is because it is too hard and costly to have many experts check large numbers of people. Tests are needed that work well and use the same cut-off points, taking into account things like age and education. Studies should also start with people who are 60 years old, as this is how the World Health Organization defines older people. Doing this will lead to a better understanding of thinking problems worldwide. This information is key for public health efforts to help older people at risk. These efforts should focus on treating health issues like diabetes and high blood pressure, reducing obesity, helping people stop smoking, encouraging more physical activity, and promoting better education. Studies should also be done to understand thinking problems in specific regions, as different parts of the world have different social and economic situations.

The strengths of this study include its global view and its focus on studies of the general public, not just people in hospitals or with specific diseases. However, there were some limits. Some studies may have been missed if they were not published or were not in one of the selected languages. Also, the study did not look at the causes of thinking problems, as its main goal was to report on how common and how often they occur.

Conclusions

This study found that thinking problems are common in older adults around the world. About 5% to 41% of older adults have these problems, with an average of 19%. New cases of thinking problems appear at a rate of 22 to 76.8 for every 1000 people each year, with an average of 53.97 new cases. The study found no big differences in these numbers when looking at different groups, except for the age of the people in the studies. For future studies, it is important to agree on how to define thinking problems and use standard test scores so that studies can be compared more easily.