Prevalence of motoric cognitive risk syndrome among older adults in Brazil and evaluation of effect modification by race

Abstract

Background

Motoric cognitive risk syndrome (MCRS) is a pre-dementia syndrome of growing interest, yet it remains understudied in Latin America with a significant lack of information on the interaction between its risk factors and race.

Objective

To estimate the prevalence of MCRS among older adults in Brazil, investigate its association with various clinical and sociodemographic variables, and explore the potential of effect modification by race.

Methods

This cross-sectional, population-based study was conducted among community-dwelling older adults in Brazil, with data collected between 2015 and 2016. The diagnosis of MCRS was established following the standard recommended by the original study that first described it. We used Poisson regression models to analyze the association between MCRS and a list of 21 variables identified from a systematic review.

Results

A total of 4677 participants aged 60 years and older were included. The prevalence of MCRS in the Brazilian population of older adults was 4.34% (95% CI: 3.20%–5.48%). Higher levels of education and physical activity showed protective associations with MCRS, while depression and stroke demonstrated risk associations. A significant cross-over interaction between race and depression regarding MCRS was observed, such that the association of depression with MCRS was approximately three times higher among White individuals than Black individuals.

Conclusions

Our results challenge previous estimates that Latin America is the region with the highest prevalence of MCRS among older adults and signal the need for further studies to better investigate the modification of effect of the association between depression and MCRS by race.

Keywords

Alzheimer's disease Brazil dementia depression interaction motoric cognitive risk syndrome prevalence race risk factors

Introduction

The motoric cognitive risk syndrome (MCRS) is a pre-dementia syndromic construct, developed in 2012 by Verghese and colleagues,¹ characterized by slow gait, cognitive complaints in the absence of functional dependency for basic activities of daily living, and the absence of a diagnosis of dementia. A recent meta-analysis of 6 large epidemiological studies, including a total of 9156 participants, identified that the presence of MCRS is associated with a 1.5-fold increase in the risk of developing dementia over time.² Since the diagnosis of MCRS is easily performed, it represents a tool of great interest for clinical practice, providing an opportunity to identify individuals at increased risk for dementia development and implement preventive interventions.³

A recent systematic review of 62 studies estimated, through meta-analysis, a prevalence of 9.0% for MCRS worldwide.⁴ That review estimated a prevalence of 19.1% for MCRS in Latin America, classifying it as the region with the highest prevalence of this condition. Such estimation was based on only two studies, one of which corresponds to a Brazilian cross-sectional study that analyzed a convenience sample of 346 individuals aged 60 or older attending a geriatrics and gerontology reference service in the city of Brasilia, finding a prevalence of 24% for MCRS.⁵ Beyond the limitation related to the lack of representativeness of the Brazilian population in that study, it assessed the association of MCRS with a list of 10 risk factors through simple statistical models, i.e., without adjustment for any confounding variables.

Several studies have investigated risk factors for MCRS. A recent systematic review on this topic included 31 studies.⁶ Interestingly, none of the studies included in that review assessed whether the race/ethnicity of their participants modified the association between the investigated risk factors and the occurrence of MCRS. The growing recognition of the importance of better understanding the relationships between ethnic-racial issues and dementias lends particular relevance to this literature gap.⁷ Brazil, due to its diversity and population admixture, provides a valuable setting for studying these issues.

Given the absence of Brazilian population-based data on the prevalence of MCRS among older adults, the limited information available on its risk factors in the country, and the lack of international data on the interaction between race and known risk factors for this syndrome, the present research aimed to estimate the prevalence of MCRS among older individuals in Brazil, investigate its association with various clinical and sociodemographic variables in an exploratory manner, and evaluate the potential role of race as a modifier of these associations.

Methods

The present study corresponds to a cross-sectional analysis of baseline data from the ELSI-Brazil study, collected between 2015 and 2016. This database contains population-based information from 9412 individuals aged 50 years and older residing in the community across 70 municipalities in the five Brazilian regions. The ELSI-Brazil study was approved by the Fundação Oswaldo Cruz (FIOCRUZ) ethics committee, Minas Gerais, Brazil (protocol number 34649814.3.0000.5091). Further details about the methods of the ELSI-Brazil study are available elsewhere.⁸

For the present research, individuals aged 60 years and older recruited at the baseline of the ELSI-Brazil study were selected. Following the same procedures adopted in the original validation study of MCRS, participants who were confined to bed, self-reported having severe hearing loss, severe visual impairment, or a diagnosis of dementia made by a physician were excluded.¹ Additionally, participants with missing data on gait speed, the presence of cognitive complaints, or the presence of disability for basic activities of daily living were excluded.

This report followed the recommendations of the STROBE guidelines for the reporting of observational studies in epidemiology (Supplemental Material).⁹

Diagnosis of motoric cognitive risk syndrome

In this study, the original criteria proposed by Verghese in 2012 were used to determine the presence of MCRS, namely: the presence of slow gait and cognitive complaints, in the absence of disabilities for basic activities of daily living, or a previous diagnosis of dementia.¹

Participants’ gait speed was measured following procedures detailed elsewhere.⁸ Slow gait was defined as a standard deviation below the mean of gait speed adjusted for age and sex.

The presence of cognitive complaints was obtained through responses to the following questions in the questionnaire of the ELSI-Brazil study: “How would you rate your memory?”, with response options including “excellent; very good; good; fair; poor; don't know; no response”; and “Comparing your current memory with that of two years ago, do you think your current memory is: better; the same; worse; don't know; no response”. Participants who reported fair or poor memory, or who reported worsening memory compared to two years ago, were classified as having cognitive complaints.^10,11

The independence regarding activities of daily living was assessed based on responses to specific items in the ELSI-Brazil questionnaire related to these activities, following the criteria used by Verghese et al. in the original article validating MCRS.^1,12 Individuals who expressed the need for assistance in performing at least one of the following basic activities of daily living were classified as dependent: bathing, dressing, walking indoors, and transferring from bed to chair.

Exposure variables

In this research, the systematic review by Iqbal et al.⁶ was adopted as a starting point for the selection of variables of interest to assess their association with the occurrence of MCRS. That systematic review scrutinized risk factors that had been analyzed by at least three studies. Overall, the review evaluated 17 risk factors, namely: age, sex, education level, diabetes, hypertension, stroke, obesity, depression, falls, smoking, alcohol consumption, sedentary lifestyle, arthritis, heart disease, coronary artery disease, polypharmacy, and cancer. In addition to these variables, we included dyslipidemia as an exposure variable of interest due to its role as an additional cardiovascular risk factor, and the race variable, as it is a specific focus of this study. A detailed description of the exposure variables utilized in this study is available in the Supplemental Material.

Statistical analyses

Continuous variables were described using mean and standard deviation (SD) when they had approximately normal distribution or, otherwise, using median and interquartile range (IQR). The assessment of the statistical distribution of continuous variables for normality was performed through the inspection of quantile-quantile plots and histograms.¹³ Categorical variables were described using absolute numbers and proportions. Additionally, simple comparisons of the distribution of variables of interest regarding the presence or absence of MCRS were conducted using Welch's t-test for continuous variables and Rao & Scott's Chi-square test for categorical data.¹⁴

The evaluation of the association between exposure variables of interest and the occurrence of MCRS was assessed through simple and multivariable quasi-Poisson regressions with robust variance, and the results were presented in the form of prevalence ratios (PR).^14,15 We tested the linearity assumption of the relationship between the logarithm of the expected values for the outcome and the predictors, as well as the assumption that the variance is a linear function of the mean of those models by analyzing residual plots.^16,17 We also evaluated multicollinearity by means of variance inflation factors, probed for the existence of influential values through plots of Cook's Distance, and confirmed that there were no fitted values greater than 1.¹⁴ To allow comparability of the results of multivariable analyses with international studies, adjustment for confounding followed the standard used most commonly in other large population-based studies included in Iqbal's review.^1,6,18,19 Thus, three distinct statistical models were produced. Model 1 corresponds to simple models, where crude associations between each exposure variable and the occurrence of MCRS were tested without adjustment for confounding. In Model 2, the relationship between each of the exposure variables of interest and MCRS was adjusted for the following covariates: age, sex, and education level. Finally, in Model 3, in addition to the three sociodemographic variables inserted in Model 2, all analyses were also adjusted for the total number of medications used by participants.

Additionally, effect modification analyses were conducted by inserting an interaction term between each variable of interest and the race variable in each of the models adjusted for sex, age, education level, and number of medications (Model 3) tested. In the presence of statistically significant interaction, recommendations by Knol and Vanderweele²⁰ were followed for presenting the results of this type of analysis. Since the number of participants with MCRS in the “other” race category was limited, initial effect modification analyses conducted violated the positivity assumption,²¹ requiring these analyses to be performed in a subset including only White and Black participants.

Finally, we calculated E-values as a form of sensitivity analysis to evaluate the strength of unmeasured confounding that would be needed to negate any statistically significant finding.^22–24 When calculating the E-values for the effect modification contrasts in both the multiplicative and additive scales, we did not assume unidirectional bias confounding the associations between the exposure of interest and the outcome at both levels of the effect modifier because this is considered the most conservative approach.²⁵

All analyses were assessed with a statistical significance level of 0.05 and adjusted according to the complex sampling approach adopted in the design of the ELSI-Brazil study, computed using the R software (version 4.1.2).^14,26–29 Missing data were described as such and we did not use any imputation procedures. Given the exploratory nature of our analyses, no adjustment for multiple comparisons were performed.³⁰

Results

After applying all eligibility criteria, the sample of this research included 4677 individuals aged 60 years and older. Figure 1 describes the flowchart of participant selection. The median age of the overall sample was 68 years (IQR 64 to 74), and 2758 (59%) participants were female. Among the participants included in the study, 196 (4.19%) met the criteria for MCRS, 112 (57.1%) of whom were women.

Figure 1.

Flowchart of participant selection.

The overall prevalence of MCRS among older individuals in Brazil was estimated at 4.34% (95% CI 3.20% to 5.48%). Supplemental Table 1 describes the prevalence of the syndrome according to sex, age group, and race. There were no statistically significant differences regarding the prevalence of MCRS based on sex, race, or age group.

Table 1 describes the general characteristics of the sample included in this study and its distribution according to the presence of MCRS. In the simple analyses shown in this table, only education level, level of physical activity, presence of depressive symptoms, as well as a history of stroke, demonstrated a significant association with the occurrence of MCRS.

Table 1.

Sample characteristics according to the presence of the diagnosis of Motoric Cognitive Risk Syndrome (MCRS) among older adults participating in the Brazilian Longitudinal Study of Ageing (2015–2016).

	Total	MRCSAbsent	MRCSPresent	p
	(N = 4677)	(N = 4481)	(N = 196)	p
Age, Median (IQR)	68 (64, 74)	68 (64, 74)	68 (63, 74)	0.42^b
Age group
60 to 69	2633 (56.3%)	2518 (56.2%)	115 (58.7%)
70 to 79	1528 (32.7%)	1466 (32.7%)	62 (31.6%)	0.81^c
≥ 80	516 (11.0%)	497 (11.1%)	19 (9.7%)
Sex
Women	2758 (59.0%)	2646 (59.0%)	112 (57.1%)	0.45^c
Men	1919 (41.0%)	1835 (41.0%)	84 (42.9%)
Race
White	1873 (40.0%)	1802 (40.2%)	71 (36.2%)
Black	2457 (52.5%)	2350 (52.4%)	107 (54.6%)	0.24^c
Other^c	347 (7.4%)	329 (7.3%)	18 (9.2%)
Education level
Less than high school	3853 (82.4%)	3679 (82.1%)	174 (88.8%)	0.01^c
Completed high school or more	786 (16.8%)	765 (17.1%)	21 (10.7%)
Missing	38 (0.8%)	37 (0.8%)	1 (0.5%)
Body Mass Index (BMI) category
Low weight (BMI < 18.5)	78 (1.7%)	76 (1.7%)	2 (1.0%)
Normal weight (18.5≤ BMI <25)	1412 (30.2%)	1357 (30.3%)	55 (28.1%)	0.68^c
Overweight (25≤ BMI <30)	1865 (39.9%)	1789 (39.9%)	76 (38.8%)
Obesity (BMI ≥30)	1292 (27.6%)	1230 (27.4%)	62 (31.6%)
Missing	30 (0.6%)	29 (0.6%)	1 (0.5%)
Hypertension	2859 (61.1%)	2728 (60.9%)	131 (66.8%)	0.45^c
Missing	6 (0.1%)	6 (0.1%)	0 (0%)
Diabetes	863 (18.5%)	825 (18.4%)	38 (19.4%)	0.62^c
Missing	23 (0.5%)	23 (0.5%)	0 (0%)
Dyslipidemia	1513 (32.3%)	1448 (32.3%)	65 (33.2%)	0.91^c
Missing	37 (0.8%)	36 (0.8%)	1 (0.5%)
Angina	184 (3.9%)	171 (3.8%)	13 (6.6%)	0.07^c
Missing	25 (0.5%)	23 (0.5%)	2 (1.0%)
Previous myocardial infarction	321 (6.9%)	306 (6.8%)	15 (7.7%)	0.64^c
Missing	7 (0.1%)	7 (0.2%)	0 (0%)
Previous stroke	270 (5.8%)	252 (5.6%)	18 (9.2%)	<0.01^c
Missing	3 (0.1%)	3 (0.1%)	0 (0%)
Heart failure	397 (8.5%)	379 (8.5%)	18 (9.2%)	0.88^c
Missing	16 (0.3%)	15 (0.3%)	1 (0.5%)
Arthritis	1135 (24.3%)	1081 (24.1%)	54 (27.6%)	0.44^c
Missing	27 (0.6%)	27 (0.6%)	0 (0%)
Cancer	298 (6.4%)	286 (6.4%)	12 (6.1%)	0.46^c
Missing	5 (0.1%)	4 (0.1%)	1 (0.5%)
Depression	1336 (28.6%)	1252 (27.9%)	84 (42.9%)	<0.001^c
Falls ≥1 in the last 12 months	1106 (23.6%)	1059 (23.6%)	47 (24.0%)	0.21^c
Missing	5 (0.1%)	4 (0.1%)	1 (0.5%)
Number of medications, median (IQR)	2 (1, 4)	2 (1, 4)	2(1, 4)	0.08^b
Missing	20 (0.4%)	20 (0.4%)	0 (0%)
Polypharmacy (≥5 medications)	783 (16.7%)	742 (16.6%)	41 (20.9%)	0.43^c
Level of physical activity
Low	1839 (39.3%)	1741 (38.9%)	98 (50.0%)
Moderate	1461 (31.2%)	1415 (31.6%)	46 (23.5%)	0.01^c
High	1278 (27.3%)	1233 (27.5%)	45 (23.0%)
Missing	99 (2.1%)	92 (2.1%)	7 (3.6%)
Smoking
Never Smoked	2127 (45.5%)	2037 (45.5%)	90 (45.9%)
Former smoker	1876 (40.1%)	1797 (40.1%)	79 (40.3%)	0.88^c
Current smoker	672 (14.4%)	645 (14.4%)	27 (13.8%)
Missing	2 (0.0%)	2 (0.0%)	0 (0%)
Alcohol consumption
Never drinks	3609 (77.2%)	3447 (76.9%)	162 (82.7%)
Low or moderate consumption	695 (14.9%)	668 (14.9%)	27 (13.8%)	0.12^c
High consumption	365 (7.8%)	358 (8.0%)	7 (3.6%)
Missing	16 (0.3%)	16 (0.4%)	0 (0%)

Note: Bold type was used to highlight statistically significant findings.

Welch's t test

Rao-Scott Chi square test

Other: Racial group formed by Asians, native Brazilians, and individuals who did not report any race.

Table 2 describes the results of simple and multivariable Poisson regression analyses examining the association between exposure variables of interest and the occurrence of MCRS. Overall, the three types of statistical models tested yielded quite similar results. Specifically, only four variables—education level, depression, history of stroke, and level of physical activity—showed statistically significant associations with the occurrence of MCRS. Race and all other variables did not demonstrate a significant association with the occurrence of MCRS in any of the models tested.

Table 2.

Analysis of the relationship between the exposure variables of interest and the occurrence of Motoric Cognitive Risk Syndrome (MCRS) among older adults participating in the Brazilian Longitudinal Study of Ageing (2015–2016).

Variables	Model 1^a	Model 2^a	Model 3^a
Variables	PR (95%CI)	PR (95%CI)	PR (95%CI)
Age	0.99 (0.97 to 1.01)	0.99 (0.96 to 1.01)	0.99 (0.96 to 1.01)
Age group
60 to 69	REF	REF	REF
70 to 79	0.80 (0.57 to 1.12)	0.77 (0.55 to 1.08)	0.76 (0.54 to 1.06)
≥80	0.78 (0.44 to 1.39)	0.75 (0.42 to 1.34)	0.74 (0.41 to 1.31)
Male sex^a	1.13 (0.83 to 1.53)	1.13 (0.82 to 1.55)	1.15 (0.83 to 1.58)
Race
White	REF	REF	REF
Black	1.53 (0.90 to 2.61)	1.43 (0.84 to 2.44)	1.45 (0.86 to 2.45)
Other	1.65 (0.82 to 3.32)	1.51 (0.77 to 2.97)	1.53 (0.78 to 3.00)
Education ≥ completed high school^a	0.50 (0.28 to 0.86)	0.48 (0.28 to 0.84)	0.48 (0.28 to 0.84)
Obesity	1.21 (0.81 to 1.80)	1.27 (0.87 to 1.85)	1.27 (0.87 to 1.85)
Low weight (BMI < 18.5)	0.59 (0.11 to 3.08)	0.59 (0.11 to 3.07)	0.59 (0.11 to 3.07)
Normal weight (18.5≤ BMI <25)	REF	REF	REF
Overweight (25≤ BMI <30)	0.88 (0.54 to 1.44)	0.88 (0.54 to 1.45)	0.88 (0.54 to 1.45)
Obesity (BMI ≥30)	1.12 (0.70 to 1.77)	1.18 (0.75 to 1.84)	1.18 (0.76 to 1.85)
Hypertension^a	1.16 (0.79 to 1.69)	1.14 (0.78 to 1.66)	1.12 (0.72 to 1.73)
Diabetes^a	0.90 (0.59 to 1.37)	0.91 (0.60 to 1.39)	0.88 (0.56 to 1.38)
Dyslipidemia^a	1.03 (0.66 to 1.59)	1.02 (0.65 to 1.59)	1.02 (0.66 to 1.59)
Angina^a	1.82 (0.95 to 3.49)	1.83 (0.95 to 3.51)	1.83 (0.95 to 3.53)
Previous myocardial infarction^a	1.17 (0.61 to 2.23)	1.18 (0.61 to 2.25)	1.13 (0.55 to 2.32)
Previous stroke^a	1.97 (1.20 to 3.24)	1.88 (1.14 to 3.09)	1.86 (1.12 to 3.10)
Heart failure^a	1.05 (0.57 to 1.92)	1.06 (0.58 to 1.95)	1.01 (0.56 to 1.79)
Arthritis^a	0.86 (0.60 to 1.25)	0.86 (0.59 to 1.24)	0.84 (0.59 to 1.21)
Cancer^a	0.78 (0.40 to 1.51)	0.80 (0.42 to 1.53)	0.79 (0.41 to 1.49)
Depression^a	2.02 (1.42 to 2.88)	2.03 (1.43 to 2.88)	2.03 (1.44 to 2.87)
Falls ≥1 in the last 12 months	1.32 (0.85 to 2.03)	1.36 (0.89 to 2.08)	1.37 (0.89 to 2.11)
No., medications used	1.01 (0.94 to 1.09)	1.02 (0.95 to 1.10)	1.02 (0.95 to 1.10)
Polypharmacy (≥5 medications)^a	1.23 (0.73 to 2.08)	1.26 (0.75 to 2.13)	1.40 (0.69 to 2.85)
Level of physical activity
Low	REF	REF	REF
Moderate	0.57 (0.38 to 0.85)	0.58 (0.38 to 0.87)	0.57 (0.38 to 0.87)
High	0.63 (0.40 to 1.00)	0.62 (0.39 to 0.98)	0.62 (0.39 to 0.99)
Smoking
Never	REF	REF	REF
Previous Smoking	0.93 (0.62 to 1.39)	0.85 (0.56 to 1.31)	0.85 (0.55 to 1.30)
Current Smoker	0.89 (0.56 to 1.40)	0.79 (0.50 to 1.25)	0.80 (0.50 to 1.28)
Alcohol consumption
Never	REF	REF	REF
Low or moderate consumption	0.71 (0.47 to 1.08)	0.68 (0.45 to 1.03)	0.68 (0.45 to 1.04)
High consumption	0.50 (0.16 to 1.62)	0.46 (0.15 to 1.47)	0.47 (0.15 to 1.48)

Note: Bold type was used to highlight statistically significant findings.

PR: prevalence ratio; CI: confidence interval

Model 1: no adjustment for confounding variables; ^bModel 2: adjusted for age, sex, and education; ^cModel 3: adjusted for age, sex, education, and number of medications in use. ^dThe reference values of these variables are those of their opposite category (e.g., female vs. male sex) or the absence of the condition (e.g., absence of diabetes versus presence of diabetes).

Among all effect modification analyses of the associations between exposure variables of interest and the occurrence of MCRS by race, only in the model involving the depression variable was a significant interaction identified. The results of this effect modification analysis are described in Table 3, where a significant and consistent cross-over interaction is observed both in the multiplicative and additive scales, indicating that the association measure between depression and MCRS among Blacks is weaker than among Whites. On the multiplicative scale, this association is about 2.9 times weaker among Blacks than among Whites (i.e., 1/0.35). While among Whites, the presence of depression was associated with a 3.4 times higher probability of MCRS occurrence, among Blacks, this association did not even reach statistical significance. Furthermore, according to the data presented in this table, Black individuals with and without depression showed a prevalence of MCR about 1.78 and 2.17 times higher than that of White individuals without depression, respectively.

Table 3.

Results of the effect modification analysis of the association between depression and the occurrence of Motoric Cognitive Risk Syndrome according to the race of study participants.

	Non-depressed		Depressed		Effect of depression by racial stratum
	N with/without MCRS	PR (95%CI)	N with/without MCRS	PR (95%CI)	PR (95%CI)
Whites	35/1369	1[Reference]	36/433	3.44 (1.99 to 5.96); p < 0.001	3.44 (1.99 to 5.96); p < 0.001
Blacks	68/1642	1.78 (1.15 to 2.75); p < 0.01	39/708	2.17 (1.19 to 3.94); p = 0.01	1.22 (0.75 to 1.96);p = 0.43
Effect Modification Measure on the Multiplicative Scale: PR ratio: 0.35 (0.19, 0.67); p < 0.01
Effect Modification Measure on the Additive Scale: RERI: −2.06 (−3.89, −0.23); p = 0.01

Note: Bold type was used to highlight statistically significant findings.

Models adjusted for sex, age, education, and number of medications used.

CI: confidence interval; MCRS: motoric cognitive risk syndrome; PR: prevalence ratio; RERI: relative excess risk due to interaction.

Figure 2 presents the estimated marginal prevalence of MCRS by multivariable model 3 as a function of the presence of depression stratified by race and provides a graphical perspective on the interaction between depression and race regarding the occurrence of MCRS.

Figure 2.

Estimated marginal prevalence of motoric cognitive risk syndrome (MCRS) according to the presence of depression and stratified by race.

Table 4 shows the E-values calculated as a form of sensitivity analysis to estimate the strength of the association between unmeasured confounders and both the outcome and the exposures evaluated that would be needed to nullify the observed exposure-outcome associations that were statistically significant for the regression models including sex, age, education, and number of medications as covariates. The E-values for the prevalence ratio estimate and confidence interval of the effect modification in the multiplicative scale are 2.77 and 1.74, respectively. The E-values for the prevalence ratio estimate and confidence interval of the effect modification in the additive scale are 2.65 and 1.61, respectively. Importantly, E-values must be interpreted in the context of the strength of association identified for other risk factors considering a similar set of adjustment for confounding.^23,24 The E-values presented above and in Table 4 provide evidence in favor of the robustness of our findings since the strength of associations between unmeasured confounders and both the exposures of interest and the outcome that would be required to nullify the statistical significance of our findings is stronger than most associations displayed in Table 2 for a large number of variables.

Table 4.

E-values for the prevalence ratio estimates and for the confidence interval limits closest to one for the four exposures that were statistically significant in the regression models including sex, age, education, and number of medications as covariates (i.e., regression model 3 described in the text).

Variables	PR and 95% CI for regression model 3	E-value for PR estimate^a	E-value for CI limit closest to 1^b
Education ≥ completed high school	0.48 (0.28 to 0.84)	3.59	1.67
Previous stroke	1.86 (1.12 to 3.10)	3.12	1.49
DepressionLevel of physical activity	2.03 (1.44 to 2.87)	3.48	2.24
Moderate	0.57 (0.38 to 0.87)	2.89	1.56
High	0.62 (0.39 to 0.99)	2.61	1.11

PR: prevalence ratio; CI: confidence interval.

The E-value for the Prevalence Ratio estimate represents the strength of association in terms of prevalence ratios between the unmeasured confounder and both the exposure and the outcome that would be needed to move the prevalence ratio estimate of the association between the exposure and the outcome to one, conditional on the covariates that the model was adjusted for.

The E-value for the Confidence Interval limit closest to one represents the strength of association in terms of prevalence ratios between the unmeasured confounder and both the exposure and the outcome that would be needed to move the confidence interval of the association between the exposure and the outcome encompass the value of one, conditional on the covariates that the model was adjusted for.

Discussion

In this population-based study, the prevalence of MCRS among older individuals in Brazil was estimated, and the association of this syndrome's occurrence with a series of potential risk factors, examined in other international studies, was investigated. The presence of depression and a history of stroke were positively associated with the occurrence of MCRS, while higher levels of education and moderate to high levels of physical activity showed negative associations. Additionally, it was identified that the race of the participants was a significant modifier of the association between depression and the presence of MCRS.

The prevalence of MCRS among older Brazilians of 4.3% found in this research is below the average of 9.0% estimated by the meta-analysis of 62 studies included in the most recent systematic review published on this topic,⁴ with individual prevalence measures ranging from 2.4% in Ireland to 33.3% in a study across multiple countries including the United States, Japan, France, and Australia. Unfortunately, the inclusion of a previous convenience-sample-based Brazilian study⁵ in that review significantly biased its results and led its authors to erroneously conclude that Latin America was the continent with the highest prevalence of MCRS.⁴

Regarding the factors associated with the occurrence of MCRS, the four variables that demonstrated statistically significant association with the occurrence of this syndrome—depression, history of stroke, level of education, and physical activity—also appear in other studies on this topic.^{10,18,31–33} The pathophysiological mechanisms underlying the association between a history of stroke and physical activity with the occurrence of MCRS are directly related to cardiovascular health and the consequent occurrence of cerebrovascular lesions in both gray and white matter. It is believed that the association between depression and MCRS reflects various mechanisms. On one hand, depression represents both a risk factor for the development of cerebrovascular disease,³⁴ and the presence of subclinical cerebrovascular lesions is known to be associated with the development of depressive symptoms.³⁵ Additionally, depression and depressive symptoms are associated with gait changes in elderly individuals,³⁶ cognitive complaints, and executive dysfunction.³⁷ The protective association between a higher level of education and the occurrence of MCRS likely stems from the fact that education contributes to a greater cognitive reserve in individuals and is therefore a protective factor for both MCRS and dementia.^38,39

In general, the association between race/ethnicity and the occurrence of MCRS has not been extensively studied in most studies focusing on this condition. To the best of our knowledge, only in the original study where MCRS was first described, was this issue investigated.¹ Interestingly, in that study, the raw data presented by the authors indicated a PR for MCRS approximately twice as high among Black individuals compared to Whites.

The present study did not show significant differences in the prevalence of MCRS among different racial groups. Nevertheless, we observed a significant role of race as a modifier of the association between depression and the occurrence of MCRS. Explanations for this finding are not clear, and we did not identify previous literature on the interaction between race/ethnicity, depression, and cognitive decline. However, literature on the relationships between race/ethnicity, depression, and dementia separately represents a highly complex field that requires further investigation. For instance, although a recent large cross-sectional study in the United States observed a higher frequency of depression among Black individuals compared to Whites,⁴⁰ several prior population-based epidemiological studies in that country identified a paradox characterized by a higher frequency of physical diseases such as hypertension and diabetes, and a lower frequency of depression among Black individuals than Whites.^41–43 Various hypotheses have been developed to explain this issue, including the possibility of selection bias, measurement errors, and different coping mechanisms for stress among individuals of different races/ethnicities. When investigating this latter hypothesis using longitudinal data from the National Epidemiologic Survey on Alcohol and Related Conditions with over 40,000 adult participants, Keyes et al. observed a positive association of alcohol consumption and smoking with the occurrence of depression among Whites but not Blacks.⁴² Similarly, those authors noted that unhealthy lifestyle habits among Black men had a protective association with the occurrence of depression compared to White men. Additionally, Breslau et al.⁴¹ demonstrated, using data from the National Comorbidity Survey in the United States, that for the same level of depression, Black individuals were less likely than Whites to respond positively to three questions from a depression questionnaire addressing suicidal thoughts, self-perception of worthlessness, and feelings of exhaustion. Nonetheless, these authors concluded that these differences were not sufficient to justify the difference in the prevalence of depression among racial/ethnic groups in that study.

A recent systematic review identified that two out of four international studies revealed a higher incidence of dementia among Black individuals compared to White individuals, calculating a pooled relative risk of 1.33 for Black individuals compared to Whites.⁴⁴ Additionally, a recent British study demonstrated that some risk factors for dementia, such as hypertension and diabetes, had different magnitudes depending on the race of the individuals.⁴⁵ Interestingly, in that study, no significant interactions between race and depression were observed in terms of the risk of developing dementia. On the other hand, a recent Brazilian study, also based on data from the ELSI-Brazil study, found no relevant differences between races in terms of the population-attributable fractions for 12 risk factors for dementia among White and Black individuals.⁴⁶

An additional possibility that should be considered as a potential explanation for the observed interaction between race and depression regarding the occurrence of MCRS is the healthy survivor bias. The existence of a shorter life expectancy among Black individuals compared to White individuals is a recognized fact in the country. Some research has identified a difference of 33 years in median survival favoring Whites over Blacks.⁴⁷ Therefore, Black individuals who have reached old age have overcome various risks for premature death and would be considered more resilient to various risk factors.⁴⁸

Our study has some limitations that should be considered. The main limitation pertains to its cross-sectional design, which precludes the establishment of causal relationships and creates the possibility of reverse causality with regards to the associations between certain exposures and the outcome. For example, the possibility of reverse causation could have led to overestimation of the association between depression and MCRS. Unfortunately, although the data from the second wave of the ELSI-Brazil study are already available, the necessary codes to match participants from the first and second waves have not yet been released. When these data become available, it will be possible to conduct more robust longitudinal analyses with incidence data to confirm the findings of this research. Another limitation concerns the use of self-reporting of a medical diagnosis to determine the presence of diseases such as hypertension, diabetes, and a history of stroke and heart attack. Although this is a common practice in population surveys, the possibility of information bias resulting from underdiagnosis of these conditions in groups with less access to healthcare services cannot be ruled out. That bias could have contributed to the underestimation of the association between those variables and MCRS. Due to the lack of more detailed information about participants’ self-classification in terms of ethnicity, it was only possible to analyze the complex ethnoracial issue based on the skin color self-reported by participants. Despite this, it is a common approach to deal with this issue in other studies^41,42,46,49 and it is hard to determine the direction of that kind of bias on our results given the degree miscegenation in the Brazilian population.⁵⁰ Finally, it is not possible to rule out the presence of uncontrolled confounders that might have changed the associations that we observed. However, the E-values that we calculated suggest some degree of robustness of our findings in that regard.

On the other hand, this research has several strengths, such as its national representativeness of the Brazilian population of community-dwelling older adults, to which our results are generalizable; the replication of the same selection and diagnostic criteria used in the original study that developed the construct of MCRS; the low number of missing values; the low probability of selection bias due to its population-based design; the use of a set of variables for confounding adjustment that facilitates comparison with major international studies; and the use of a validated scale to define the presence of depression episodes in contrast to self-reported diagnosis.

In conclusion, the results of this research provide data on MCRS prevalence representative of community-dwelling older people in Brazil and demonstrate that previous estimates based on a convenience sample study were exaggerated. Ultimately, the presented data should lead to a reassessment of previous meta-analyses that had concluded that Latin America was the region with the highest MCRS prevalence in the world. Additionally, the findings of this research reaffirm the importance of intersectoral actions involving the improvement of population education levels, promotion of physical activity, and prevention of cerebrovascular diseases for MCRS prevention and, consequently, future dementia development. Finally, the effect modification of the association between depression and MCRS by race deserves further exploration through future studies.

Supplemental Material

sj-docx-1-alz-10.1177_13872877241300296 - Supplemental material for Prevalence of motoric cognitive risk syndrome among older adults in Brazil and evaluation of effect modification by race

Supplemental material, sj-docx-1-alz-10.1177_13872877241300296 for Prevalence of motoric cognitive risk syndrome among older adults in Brazil and evaluation of effect modification by race by João Paulo Martins, Fernanda Bono Fukushima, Leandra Navarro Benatti, Rodrigo Bazan, Katherine Di Santi Correa da Silva and Edison Iglesias de Oliveira Vidal in Journal of Alzheimer's Disease

Footnotes

Acknowledgments

The authors would like to thank Claudia M. Coeli, Rejane S. Pinheiro, Cristiane R.M. Nascimento, Marlon J. Aliberti, Ivana R. Goncalves, and Fernanda P.M. Razera for their comments on previous versions of this work and/or methodological advice. We are also indebted to the team responsible for the ELSI-Brazil study.

ORCID iDs

João Paulo Martins

Fernanda Bono Fukushima

Leandra Navarro Benatti

Rodrigo Bazan

Katherine Di Santi Correa da Silva

Edison Iglesias de Oliveira Vidal

Author contributions

João Paulo Martins (Conceptualization; Data curation; Formal analysis; Methodology; Visualization; Writing – original draft); Fernanda Bono Fukushima (Writing – review & editing); Leandra Navarro Benatti (Data curation; Writing – review & editing); Rodrigo Bazan (Writing – review & editing); Katherine Di Santi Correa da Silva (Data curation; Writing – review & editing); Edison Iglesias de Oliveira Vidal (Conceptualization; Data curation; Formal analysis; Methodology; Supervision; Writing – review & editing).

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This specific research did not receive any funding from agencies in the public, commercial, or not-for-profit sectors. The ELSI-Brazil baseline study was supported by the Brazilian Ministry of Health (DECIT/SCTIE—Department of Science and Technology, Secretariat of Science, Technology and Strategic Inputs (grant 404965/2012-1); The Older Adult Primary Healthcare Unit of the Lifecycle Department of the Primary Health Care Secretariat – Ministry of Health (COPID/DECIV/SAPS/MS) (grants 20836, 22566, and 23700); and the Brazilian Ministry of Science, Technology, Innovation and Communication. EIOV is a recipient of the Brazilian National Council for Scientific and Technological Development (CNPq) research productivity fellowship (grant 312499/2022-1) and is also supported by FAPESP (grant 2023/00823-9). The funders of the ELSI-Brazil study did not play any role in the conception, design, or conduction of the study, and were not involved neither in the analysis/interpretation of data nor in the preparation of the manuscript.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data availability

Data from the ELSI-Brazil study are openly available at

Supplemental material

Supplemental material for this article is available online.

References

Verghese

Wang

Lipton

, et al. Motoric cognitive risk syndrome and the risk of dementia. J Gerontol A Biol Sci Med Sci 2013; 68: 412–418.

Sekhon

Allali

Launay

, et al. Motoric cognitive risk syndrome, incident cognitive impairment and morphological brain abnormalities: systematic review and meta-analysis. Maturitas 2019; 123: 45–54.

Verghese

Ayers

Barzilai

, et al. Motoric cognitive risk syndrome: multicenter incidence study. Neurology 2014; 83: 2278–2284.

Wen

Peng

Wang

, et al. Prevalence of motoric cognitive risk syndrome among older adults: a systematic review and meta-analysis. Aging Ment Health 2023; 27: 1443–1455.

Moura

TGD

Pinheiro

. Síndrome de risco cognitivo motor em pessoas idosas de um serviço de saúde do Distrito Federal: estudo transversal. Rev Bras Geriatr E Gerontol 2021; 24: e200285.

Iqbal

Hasanain

Ahmed

, et al. Association of motoric cognitive risk syndrome with cardiovascular and noncardiovascular factors: a systematic review and meta-analysis. J Am Med Dir Assoc 2022; 23: 810–822.

Brijnath

Croy

Sabates

, et al. Including ethnic minorities in dementia research: recommendations from a scoping review. Alzheimers Dement (N Y) 2022; 8: e12222.

Lima-Costa

de Andrade

Souza de

PRB

, et al. The Brazilian Longitudinal Study of Aging (ELSI-Brazil): objectives and design. Am J Epidemiol 2018; 187: 1345–1353.

von Elm

Altman

Egger

, et al. Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. BMJ 2007; 335: 806–808.

10.

Sekhon

Allali

Beauchet

. The association of anxio-depressive disorders and depression with motoric cognitive risk syndrome: results from the baseline assessment of the Canadian longitudinal study on aging. Geroscience 2019; 41: 409–418.

11.

Ayers

Verghese

. Motoric cognitive risk syndrome and risk of mortality in older adults. Alzheimers Dement 2016; 12: 556–564.

12.

Gill

. Hospitalization, restricted activity, and the development of disability among older persons. JAMA 2004; 292: 2115.

13.

Mishra

Pandey

Singh

, et al. Descriptive statistics and normality tests for statistical data. Ann Card Anaesth 2019; 22: 67.

14.

Lumley

. Complex surveys: a guide to analysis using R. Hoboken, NJ: Wiley, 2010.

15.

Zou

. A modified poisson regression approach to prospective studies with binary data. Am J Epidemiol 2004; 159: 702–706.

16.

Ver Hoef

Boveng

. Quasi-Poisson vs. Negative binomial regression: how should we model overdispersed count data? Ecology 2007; 88: 2766–2772.

17.

Fox

. Applied regression analysis and generalized linear models, third Edition. Los Angeles: Sage, 2016.

18.

Doi

Verghese

Shimada

, et al. Motoric cognitive risk syndrome: prevalence and risk factors in Japanese seniors. J Am Med Dir Assoc 2015; 16: 1103.e21–1103.e25.

19.

Canadian Gait Consortium, Beauchet

Sekhon

, et al. Association of motoric cognitive risk syndrome with cardiovascular disease and risk factors: results from an original study and meta-analysis. J Alzheimers Dis 2018; 64: 875–887.

20.

Knol

VanderWeele

. Recommendations for presenting analyses of effect modification and interaction. Int J Epidemiol 2012; 41: 514–520.

21.

Petersen

Porter

Gruber

, et al. Diagnosing and responding to violations in the positivity assumption. Stat Methods Med Res 2012; 21: 31–54.

22.

Chung

. The use of the E-value for sensitivity analysis. J Clin Epidemiol 2023; 163: 92–94.

23.

VanderWeele

Ding

. Sensitivity analysis in observational research: introducing the E-value. Ann Intern Med 2017; 167: 268.

24.

Mathur

Ding

Riddell

, et al. Web site and R package for computing E-values. Epidemiology 2018; 29: e45–e47.

25.

Mathur

Smith

Yoshida

, et al. E-values for effect heterogeneity and approximations for causal interaction. Int J Epidemiol 2022; 51: 1268–1275.

26.

R Core Team. R: A language and environment for statistical computing, https://www.R-project.org/ (2021).

27.

RStudio Team. RStudio: Integrated Development for R. RStudio, http://www.rstudio.com/ (2020).

28.

. Interactionr: an R package for full reporting of effect modification and interaction. Softw Impacts 2021; 10: 100147.

29.

Wickham

. ggplot2: elegant graphics for data analysis. second edition. Switzerland: Springer, 2016.

30.

Althouse

. Adjust for multiple comparisons? It’s not that simple. Ann Thorac Surg 2016; 101: 1644–1645.

31.

Verghese

Annweiler

Ayers

, et al. Motoric cognitive risk syndrome: multicountry prevalence and dementia risk. Neurology 2014; 83: 718–726.

32.

Aguilar-Navarro

Mimenza-Alvarado

Aguilar-Esquivel

, et al. Motoric cognitive risk syndrome: prevalence and risk of cognitive impairment in a population studied in the Mexican Health and Aging Study 2012–2015. J Nutr Health Aging 2019; 23: 227–231.

33.

Allali

Ayers

Verghese

. Motoric cognitive risk syndrome subtypes and cognitive profiles. J Gerontol A Biol Sci Med Sci 2016; 71: 378–384.

34.

Correll

Solmi

Veronese

, et al. Prevalence, incidence and mortality from cardiovascular disease in patients with pooled and specific severe mental illness: a large-scale meta-analysis of 3,211,768 patients and 113,383,368 controls. World Psychiatry 2017; 16: 163–180.

35.

Taylor

Aizenstein

Alexopoulos

. The vascular depression hypothesis: mechanisms linking vascular disease with depression. Mol Psychiatry 2013; 18: 963–974.

36.

Figgins

Pieruccini-Faria

Speechley

, et al. Potentially modifiable risk factors for slow gait in community-dwelling older adults: a systematic review. Ageing Res Rev 2021; 66: 101253.

37.

Snyder

. Major depressive disorder is associated with broad impairments on neuropsychological measures of executive function: a meta-analysis and review. Psychol Bull 2013; 139: 81–132.

38.

Meiner

Ayers

Verghese

. Motoric cognitive risk syndrome: a risk factor for cognitive impairment and dementia in different populations. Ann Geriatr Med Res 2020; 24: 3–14.

39.

Sharp

Gatz

. Relationship between education and dementia: an updated systematic review. Alzheimer Dis Assoc Disord 2011; 25: 289–304.

40.

Hooker

Phibbs

Irvin

, et al. Depression among older adults in the United States by disaggregated race and ethnicity. Gerontologist 2019; 59: 886–891.

41.

Breslau

Javaras

Blacker

, et al. Differential item functioning between ethnic groups in the epidemiological assessment of depression. J Nerv Ment Dis 2008; 196: 297–306.

42.

Keyes

Barnes

Bates

. Stress, coping, and depression: testing a new hypothesis in a prospectively studied general population sample of U.S.-born Whites and Blacks. Soc Sci Med 2011; 72: 650–659.

43.

Williams

Jackson

. Social sources of racial disparities in health. Health Aff (Millwood) 2005; 24: 325–334.

44.

Shiekh

Cadogan

Lin

L-Y

, et al. Ethnic differences in dementia risk: a systematic review and meta-analysis. J Alzheimers Dis 2021; 80: 337–355.

45.

Mukadam

Marston

Lewis

, et al. South Asian, Black and White ethnicity and the effect of potentially modifiable risk factors for dementia: a study in English electronic health records. PLoS One 2023; 18: e0289893.

46.

Suemoto

Mukadam

Brucki

SMD

, et al. Risk factors for dementia in Brazil: differences by region and race. Alzheimers Dement 2023; 19: 1849–1857.

47.

Fiorio

Flor

Padilha

, et al. Mortalidade por raça/cor: evidências de desigualdades sociais em Vitória (ES), Brasil. Rev Bras Epidemiol 2011; 14: 522–530.

48.

Jackson

Hudson

Kershaw

, et al. Discrimination, chronic stress, and mortality among Black Americans: a life course framework. In: Rogers

Crimmins

(eds) International handbook of adult mortality. Dordrecht: Springer, 2011, pp.311–328.

49.

Alves

Pereira

. Race, sex and depression-free life expectancy in Brazil, 1998–2013. Int J Popul Stud 2019; 4: 1–9.

50.

Manta FS de

Pereira

Vianna

, et al. Revisiting the genetic ancestry of Brazilians using autosomal AIM-indels. PLoS One 2013; 8: e75145.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.04 MB