Cognition and its association with the factors of the EURO‐D: Suffering and Motivation. Findings from SHARE Wave 6

The aims of this study were (1) to analyse the relationship between cognition and clinical and sociodemographic variables, (2) to explore the relationship between cognitive tests and factors of EURO‐D depression scale (Suffering and Motivation), and (3) to determine the relevance of cognition with respect to clinical and sociodemographic variables in the scores of the EURO‐D factors.

aged 80 and over, in institutionalized individuals, and those suffering from some kind of dementia. [9][10][11][12] The relationship between cognition and depression has been subject to numerous studies, which have found that older adults with depressive disorders usually present cognitive complaints and that those with mild cognitive impairment usually suffer from depressive symptoms. [13][14][15][16][17] Moreover, it should be noted that perception of selfefficacy 18 and the concept of cognitive reserve 19 may act as protective factors against cognitive decline and as mediators of depressive symptoms. 20,21 Likewise, apathy can have a negative impact on the relationship between cognition and depression. 22 With respect to sociodemographic variables, older age, 1,23 female gender, 1,24 and not being married [25][26][27] are related to lower cognitive status. In addition, loneliness is associated with increased depressive symptoms and lower cognitive performance. 28, 29 Li et al 1 found that lower income and higher economic concerns were related to a lower cognitive status. In education, there is an association between lower educational level, a higher risk of depressive symptoms, 30 and lower cognitive performance, 31 and higher educational levels are related to greater cognitive reserve. 19 Finally, poorer self-perceived physical health, the presence of diseases, and activities of daily living (ADL) deficits contribute to depressive symptoms and lower cognition.
At the same time, lower cognitive status favours poorer self-perceived physical health. 15,32 The EURO-D 33 depression scale has been widely used to detect depressive symptoms and associated variables. A factor analysis of the EURO-D in the original scale generated 2 factors, Affective Suffering and Motivation, which have been validated by further research. [34][35][36][37] Several studies have analysed the variables associated with the 2 EURO-D factors and found that the female gender is more strongly associated with the Suffering factor, 34,[36][37][38] and lower verbal fluency with the Motivation factor. 36,38 Research has addressed the role of age and cognition, especially in the Motivation factor. The hypothesis of late-onset depression and executive function deficits 39 could explain the higher Motivation factor score in older age and poor cognition. 36 Given this background, it was considered useful to explore the relationship between general cognition and the Suffering and Motivation factors 34  the Suffering and Motivation factors, will allow us to assess their differential contribution to depression. Previous research indicates that a higher score of Motivation factor seems to be related to lower verbal fluency, higher vascular risk, and cognitive impairment. We consider it especially relevant to clarify the relevance of cognition, as an independent variable, with respect to other variables of possible confusion (age, gender, education, loneliness, or health) in the differences between the factors Suffering and Motivation of the EURO-D depression scale.
Thus, the aims of this study were (1) to analyse the relationship between cognition and clinical and sociodemographic variables, (2)

Key points
• Poor cognition was associated with older age, lower educational levels, higher economic difficulties, and higher depression levels.
• The depression scale (EURO-D) scores were associated with higher levels of loneliness, poorer self-perceived physical health, female gender, and poor cognition. In addition, the activities included in the variable "existence of ADL deficits" were dressing, bathing, eating, getting in or out of bed, using the toilet, preparing a hot meal, shopping for groceries, telephone calls, taking medications, and managing money. They were assessed by asking if the participants were able to do them by themselves or not.

| Loneliness scale
The 3-item loneliness scale was used (lack of company, isolation from others, and/or feeling of irrelevance), which indicates the frequency with which feelings of loneliness are experienced (almost never = 1, sometimes = 2, and/or almost always = 3), with a total score ranging from 3 to 9, as each item could range from 1 to 3. 42 A higher score indicates a higher presence of loneliness. In a study by Hughes et al, 42 the scale produced a moderate Cronbach's alpha (.72), as in the present study (.75). The 3-item questionnaire was adapted from the Revised UCLA Loneliness Scale (R-UCLA), after exploratory and confirmatory factor analysis. Both the 3-item loneliness scales and the R-UCLA have been validated and are able to be self-administered. We classified subjects with feelings of loneliness when the overall score was >3.

| Depressive symptoms
The instrument used was the EURO-D depression scale, 33 which consists of 12 dichotomous items to indicate the presence or absence of the following symptoms over the last month: depressive symptoms, pessimism, a desire for death, guilt, irritability, tearfulness, fatigue, sleep problems, disinterest, loss of appetite, lack of concentration, and lack of enjoyment. The cut-off point for depression is ≥4. The total score ranges from 0 to 12. Higher scores indicate a higher presence of depressive symptoms. In the EURODEP 35 study, the scale had a moderate Cronbach's alpha (.61-.75), as in the present study (.71).
The present study used the factorial solution from a previous study 34 that established the presence of 2 factors in the EURO-D scale: Suffering (composed of the items depressive symptoms, guilt, sleep problems, irritability, fatigue, and tearfulness) and Motivation (composed of the items pessimism, disinterest, loss of appetite, lack of concentration, and lack of enjoyment).

| General cognition
Specific items in the SHARE project were used to assess the cognitive status of participants based on immediate recall (presenting a list of 10 words that should be repeated immediately; range = 0-10), delayed recall (asking for the list of 10 words presented before; range = 0-10), orientation (asking about the day, month, and year and the day of the week; range = 0-4), subtraction (asking simple mathematical operations to mentally calculate; range = 0-5), and verbal fluency (asking about the maximum number of animals remembered in a given period of time; range = 1-10). A joint scale was created with all items with a total score range of 1 to 39. The higher the score, the higher the cognition.

| Statistical analysis
A descriptive study of the sample was performed using means and standard deviation for the continuous variables and frequencies for the categorical variables. The categorical variables were contrasted with the chi-squared test (χ 2 ), while the parametric tests of the t test and ANOVA F test were used for the continuous variables.

| Description of the sample
The study sample consisted of 63 755 participants with a mean age of 67.5 ± 9.7 years. The majority were women (56.2%), were married (69.0%), had more than 8 years of schooling (78.8%), and had no major economic difficulties (61.3%). The 56.0% did not feel lonely, presented favourable physical health (61.6%), had no ADL deficits (86.4%), and were suffering from one illness or none (51.2%).
With regard to depression, 27.4% of the participants scored above the cut-off point for depression (EURO-D: ≥ 4). The overall mean score for depression was 2.4 ± 2.2. The Suffering factor presented a higher average score than the Motivation factor (1.7 ± 1.5 vs 0.6 ± 0.9; P < .001).
The mean score of the general cognition was 22.4 ± 6.7. Full data are shown in Table 1.

| Variables associated with cognition
The bivariate analysis showed strong associations with effect size (η 2 > 13) between the poor general cognition with higher age and lower educational level. The higher economic difficulties, poorer selfperceived physical health, the presence of diseases, ADL deficits, and depressive symptoms were also associated with poor cognition, with a moderate effect size. The other variables examined, ie, gender, marital status, and diseases, showed a lower association with a weak effect size.
The correlation between general cognition and the Motivation factor was higher than that between general cognition and the Suffering factor, although both were significant. Full results are shown in Table S1 (Supporting Information).

| Cognition and variables: Multiple linear regression analysis
The variables that presented a higher association with higher general    Table 3.

| Multivariate regression analysis: EURO-D depression scale, factors, and cognition
Three linear regression analyses were adjusted for the EURO-D depression scale and for each factor, ie, Suffering and Motivation. All independent variables were introduced in a single step, and those that showed a high degree of collinearity and/or a low contribution coefficient were eliminated, thereby resulting in a final model that included cognition, loneliness, physical health, gender, diseases, ADL deficits, and age.
In the EURO-D and in both factors, a higher depression level was associated with higher loneliness and poorer self-perceived physical health. In the Suffering factor, diseases (β = .127, P < .001) and younger age (β = −.105, P < .001) also had betas higher than 0.10, associated with higher depression, while in the Motivation factor, ADL deficits (β = .126, P < .001) were more relevant.   Table 4.
Data regarding age are concordant with those of the bivariate analysis (Table S2)

| Relationship between cognition and clinical and sociodemographic variables
The first aim of the study was to explore the relationship between clinical and sociodemographic variables and the general cognition of the participants. The analysis showed that the variables most associated with poor cognition were older age, lower educational level, higher economic difficulties, and higher depression level. The EURO-D factors Suffering and Motivation 34 were also associated with poorer cognition, to a greater extent in the Motivation factor than in the Suffering factor.
These results corroborate previous studies that indicate that older age was associated with poorer cognitive performance, 1,23 fact that some authors explain by the lower physical activity, exercise, and cognitive training that is common in older adults. 48,49 In addition, lower educational levels and the consequent poorer cognitive reserve 19 would be associated with poorer cognition 31 and faster cognitive decline. 48 Similarly, economic difficulties would be related to a higher frequency of depressive symptoms and poorer cognitive performance because of a cumulative effect, whereby the deficiencies over the years due to economic difficulties cause poorer access to health services. 1,50,51 The stressors inherent to economic concerns would also affect the appearance of depressive symptomatology and affect the neural circuits. 52 It seems, therefore, that education and economic aspects act as mediators in the influence of older age on  cognition, given that the educational level is closely related to occupational complexity and the economic status in adulthood. 53

| Relationship between cognition and EURO-D factors
The second aim of the study was to analyse the relationship between cognition and the EURO-D and its factors, ie, Suffering and Motivation. The analysis showed that the correlations between the cognitive tests and the EURO-D factors were always negative and higher for the Motivation factor, with significant differences between factors in each cognitive test.
The main differential aspects of general cognition between the Suffering and Motivation factors were general cognition, verbal fluency, and immediate recall. These subscales, which measure mental flexibility and the capacity to consolidate and recover memories, were the most strongly influenced by depressive symptoms, both in the overall EURO-D score and in the Suffering and Motivation factors. 14,34,54 Other authors who have used the same EURO-D depression scale have found verbal fluency performance to be more strongly associated with motivational symptoms of depression than with affective suffering symptoms. [36][37][38]55 In our study, verbal fluency also presented the highest correlation with the Motivation factor, although the correlations with this factor were higher in all cognitive subscales.

| Variables associated with the EURO-D and suffering and motivation factors
The third objective of this study was to determine the relevance of cognition with respect to the other clinical and sociodemographic variables in the EURO-D and the Suffering and Motivation factor scores.
Loneliness and health problems were common variables in the EURO-D and Suffering and Motivation factors. The association between higher loneliness and higher depressive scores in both factors might be because of the lack of social interaction 28 and social support 29 that people with depression generally experience. With regard to health, in both self-perceived physical condition and ADL deficits and the presence of diseases, our results are consistent with those usually reported in the literature, [56][57][58] thus confirming that the poorest perception of health and/or diseases or ADL deficits were associated with a higher presence of depressive symptoms.
The main differences revealed in the analysis were that general cognition was nonsignificant in the Suffering factor, but significant in the Motivation factor, and gender was more relevant in the Suffering factor.
Our results corroborate the findings of previous studies, ie, that poorer cognition was associated with higher depression levels. [13][14][15][16][17] The most relevant of our findings was that the association between a poor cognition and a higher depression only occurred in the Motivation factor. One possible explanation is that the Motivation factor includes some specific cognitive items such as disinterest, lack of concentration, or lack of enjoyment, while the items of the Suffering factor are not specifically cognitive. Some authors have hypothesized that vascular depression could explain this stronger association between the motivational factor and poor cognition, 59 while others have suggested that high scores on motivational symptoms would be more present in subjects with mild cognitive impairment 60 or even be a prodrome of dementia. 61 A second difference was that women were more relevant in the Suffering factor, according to several studies, 34,[36][37][38] and usually present higher percentages of widowhood, loneliness, and social isolation. 35,36,58,62 A key topic addressed by several authors is the influence of age and cognition on Motivation factor scores. 36,38 They consider that Motivation factor scores could be related to the depression-executive dysfunction hypothesis, with the presence of motivation-related symptoms caused by the deterioration of executive function in ageing, 39 because older age may lead to a reduction in cognitive function, as well as less activity and motivation. 37 In our study, eliminating cognition from the regression analysis, in the Motivation factor, was associated with an increase in ADL deficits and higher age, and therefore, the hypothesis that executive deficits are more frequent in age-related depression could be related to higher scores in the Motivation factor.
However, while the older age was related to a higher Motivation factor, 38 the younger age was related to a higher Suffering factor.
It is possible that the differences between the Suffering and

| Clinical implications
Our results corroborate the association between depression and cognitive decline. [13][14][15][16][17] The presence of depression favours lower cognitive performance, in the same way that cognitive performance influences and conditions the possible presence of depressive symptomatology. Nevertheless, different sociodemographic variables would mediate and modify the type and direction of these associations. It is therefore necessary to assess the presence of both depression and cognitive impairment at the time of diagnosis to establish appropriate therapeutic criteria.
In light of the results of our study and previous research, it would be convenient to specifically assess the Motivation factor items because of their remarkable association with lower cognition and possible cognitive deterioration.

| Limitations and future research
An important limitation of our study has been the absence of a clinical examination of the participants with regard to depression, as well as not being able to use other psychometric instruments on depression to compare the results.
Given the importance of education in cognition, it would be advisable to carry out further research on the education system in each of the countries of origin of the SHARE participants with a view to assessing this specific aspect. Moreover, given the impact that health has on depression, it would be useful to identify the diseases and ADL deficits that condition self-perception of health status and the associated depression.
Although this is a comprehensive and cross-cutting analysis, a longitudinal study that reflects the differences between countries and/or European regions would be useful.

| CONCLUSIONS
This analysis showed an association mainly between lower cognitive status and higher age and lower educational levels. Greater economic difficulties and the presence of depressive symptoms were also related to lower cognitive performance.
The scores of the depression scale (EURO-D) were associated with greater loneliness, poorer self-perceived health, female gender, and lower cognition.

CONFLICT OF INTEREST
None declared.