Subjective Cognitive Complaints in Cognitively Healthy Older Adults and Their Relationship to Cognitive Performance and Depressive Symptoms

INTRODUCTION

Subjective cognitive complaints (SCCs) are common in older adults. In population-based studies, including both cognitively normal subjects and those with probable cognitive impairment, the prevalence of SCCs ranges from 10% to 81% [1–3]. A higher rate of SCCs is usually associated with higher age [4, 5], poorer cognitive performance [6], and a higher level of depressive symptomatology [7]. Generally, SCCs have been associated with normal aging, cognitive decline beyond normal aging, and other non-cognitive conditions such as depression and anxiety [7–9]. However, little is known about how specific SCCs may relate to each of these conditions, and this might account for some authors remaining skeptical about the importance of SCCs in preclinical stages of Alzheimer’s disease (AD) [10].

The most common cause of cognitive decline beyond normal aging is AD. The pathological changes associated with AD start decades before clinical diagnosis of cognitive impairment and biomarkers can be detected by positron emission tomography imaging of amyloid-β (Aβ), or magnetic resonance imaging of the hippocampus or cerebrospinal fluid Aβ₄₂, together with subtle cognitive decline [11]. In early stages, highly educated subjects can compensate for the subtle cognitive decline, resulting in a normal performance on standard cognitive tests [12]. Still, using compensatory cognitive mechanisms does not prevent such individuals from self-experience of cognitive decline [13].

On the other hand, a number of cross-sectional studies found SCCs in subjects without dementia to be associated with depressive symptomatology rather than with objective cognitive status [7–9], and therefore the relevance of SCCs as a marker of cognitive decline due to neurodegeneration has been questioned by some authors [10]. However, it seems that the relationships between SCCs, objective cognition, and depressive symptomatology are much more complex. This complexity and absence of long-term longitudinal studies using biomarkers could be one of the reasons why the causal relationship between these factors remains poorly understood. It has been proposed that depressive symptoms can be an early prodrome of dementia, that depression may lead to hippocampal dysfunction similarly as AD does [10, 15], or that depressive symptoms may rather be a consequence of awareness of SCCs [16]. Nevertheless, most recent studies investigating AD biomarkers and neuroimaging markers either exclude subjects with depression, which precludes the study of the potential causal relationship, or statistically account for depressive symptoms, which may bias thefindings [17].

Other possible sources of incongruent findings among existing SCC studies are different approaches how to assess SCCs and cognition. Some studies have applied screening tools to evaluate global cognitive status [6, 18]. Only a few studies have administered complex cognitive batteries comprising traditional neuropsychological tests at a population level [2, 5]. The use of screening tools to evaluate global cognitive function could have led to an underpowering of the relationship between SCCs and cognition.

In previous studies, SCCs assessment ranged from a single question or a small number of items with varied content and unknown measurement properties to more comprehensive and validated questionnaires [19]. The differences in approaches might explain major discrepancies among studies estimating the prevalence and predictive value of SCCs in non-demented older adults. To date, there is no consensus or guidelines for instrument selection to assess SCCs. The QPC (Le Questionnaire de Plainte Cognitive; in English, Questionnaire of Cognitive Complaints) has been used to evaluate SCCs. The QPC is a relatively brief and easy to administer tool for assessing SCCs. The QPC was originally developed to help physicians in primary care identify subjects with cognitive impairment or depression and has been applied to populations of patients with mild cognitive impairment (MCI), mild to moderate dementia due to AD, and patients with depressive syndrome [20, 21]. The QPC is comprised of three types of questions that have recently been proposed as useful for differentiating between normal aging and prodromal AD [13]: 1) cognitive decline in comparison to the previous level of cognitive functioning; 2) cognitive functioning in comparison to individuals of the same age group; and 3) other specific complaints potentially specific for pathological aging or depression.

Using the Czech version of the French QPC, we aimed to evaluate the prevalence of SCCs among cognitively healthy Czech volunteers. We also assessed the distribution of SCCs according to demographic variables, such as age, education, and gender. Finally, we aimed to study the relation of the number of SCCs reported in QPC and specific SCCs in relation to depressive symptomology and cognitiveperformance.

Based on previous findings, we hypothesized that the number of complaints would increase with age and that higher number of complaints would be associated with more depressive symptoms and lower cognitive performance. We also expected to find specific complaints linked to cognitive functioning (especially memory) and depressivesymptomatology.

MATERIALS AND METHODS

MEASURES

Measurement of SCCs: QPC

All participants were administered the Czech version of the QPC, which was based on an original French 10-item yes/no questionnaire assessing the presence of cognitive difficulties in the last six months [20, 21]. The Czech version was made based on a translation from French to Czech by an experienced translator, followed by a back-translation from Czech to French by a translator blinded to the original French version. The discrepancies were consulted and based on the consensus between the two translations the final version was developed to preserve the original meaning as accurately as possible. For the purpose of the publication, the QPC was translated into English in cooperation with a native English speaker and the final English version was approved by the author of the questionnaire.

The first two items inquire about general memory abilities, while the remaining eight items inquire about more particular cognitive complaints with a focus on memory. The items also cover difficulties with spatial orientation, language, instrumental activities, or personality change. The exact wording of all items is presented in Fig. 1. For each item, we indicate a key name for ease in reporting findings.

OPEN IN VIEWER

Fig.1

QPC, Questionnaire of cognitive complaints. Exact phrasing and key names. Questions are related to last six months.

RESULTS

DISCUSSION

Using the QPC, we found a high prevalence of SCCs in a volunteer sample of Czech cognitively healthy older adults. SCC levels gradually increased with higher level of depressive symptomatology. There was an association between higher number of complaints and lower memory and global cognitive function when not controlling for depressive symptomatology. Two SCCs were significantly related to lower memory performance (Spatial orientation difficulties and Impression of worse memory in comparison to peers). Another two SCCs were related to higher level of depressive symptomatology (Limitation in daily activities and Personality change). Interestingly, none of the three most common SCCs (Word finding difficulties, Difficulties with recalling past events, and Impression of memory change) were related to lower memory performance or higher level of depressive symptomatology.

The reported prevalence of SCCs in older adults ranges from 10% to 81% [1–3]. This broad range underscores the difficulty of reliably estimating the prevalence of SCCs. In our study, prevalence of at least one SCC was 71% and prevalence of individual specific complaints ranged from 4% to 40%.

The majority of existing studies report a higher rate of SCCs with advancing age [30, 31], which is congruent with results of our correlation analysis. We have not found any association between years of education and the number of SCCs, which agrees with the previous study using the QPC [21]. Several community-based studies have reported a negative association between level of education and severity of SCCs in older adults, indicating an increasing the number of SCCs with lower formal education [5, 32]. In clinically-based studies the presence of subjective memory complaints is usually associated with a higher level of education [33]. We assume that the specificity of QPC complaints together with low variability in years of education of participants could explain why no education effects were demonstrated. Because of the absence of a significant education effect, education was not subsequently included in main analyses exploring the relationship between SCCs, depressive symptomatology, and cognition.

Previous findings about the association between gender and SCCs are incongruent. Some community-based studies reported a higher rate of SCCs to be associated with female gender [32, 34]. Other studies including the previous one that used the QPC did not find any effect of gender on the number of SCCs [21, 35], which is congruent with our findings. The previous studies included different SCC items. It is possible that some SCCs are more likely to be endorsed by women which could be one explanation for the inconsistency in findings.

The level of depressive symptomatology (measured by GDS-15) was the best indicator of SCCs, as reported on the QPC. This is consistent with findings of several previous studies [2, 18]. Although the overall percentage of elderly people with depressive disorder might seem low (between 1% and 5% of individuals aged 65 or older [36, 37]), the number of elderly individuals exhibiting minor individual depressive symptoms not meeting the criteria for major depressive disorder is much higher, ranging from 15% to 25% [38]. Depressed older adults often display cognitive complaints and somatic symptoms rather than affective symptoms [39]. Some of these cognitive complaints are even included in the GDS-15: “Do you feel you have more problems with memory than most people?” (compared with Impression of worse memory in comparison to peers), and “Have you dropped many of your activities and interests?” (compared with Limitation in daily activities). However, to our knowledge, no study has investigated individual items of the GDS-15 or any other depression scale in relation to individual SCC items. We identified two specific complaints related to higher GDS-15 score (Limitation in daily activities and Personality change). The latter was additionally related to lower executive performance, which often coexists with late-life depression (LLD), as shown previously [40]. Interestingly, in the previous study using the QPC [21], the two most common SCCs in participants with clinically manifested depression syndrome were different (Difficulties with recalling past events and Word finding difficulties). However, the sample with depression in that study differs from the study sample used here. Participants included in the current study were free of a clinical diagnosis of depression and reported only minor individual depressive symptoms in the GDS-15.

The relationship between SCCs and objective cognitive performance has been a matter of controversy. Several previous cross-sectional studies showed an association between SCCs and objective cognitive performance [6, 41], while others did not [3, 7]. The two studies without significant findings used only screening tools to measure cognitive performance, which could fail to reveal subtle cognitive changes and result in low variability of performance in the study sample, and thus the potential association between SCCs and cognition may be hidden by a stronger relation between SCCs and depression. In addition, in a community-based study [3], participants were classified as complainers or non-complainers based on a single question about memory. This simple classification could mislead the results. In our study applying a comprehensive neuropsychological battery and a SCC questionnaire, we have found a statistically significant association between higher number of SCCs and lower MMSE and memory performance. Interestingly, this relation became nonsignificant when controlling for depressive symptomatology.

The interpretation of these results should take into account the presumed complex relationship between depression, SCCs, and cognition. We propose three possibilities: 1) the depressive symptomatology can lead to lower cognitive performance, which may trigger the SCCs [40]; 2) the minor individual depressive symptoms in cognitively healthy older adults aretriggered by their awareness of cognitive capacity decline not captured by the cross-sectional cognitive testing [16]; or 3) both the depressive symptomatology and the subtle cognitive impairment can be a direct concomitant consequence of early neuropathological changes in AD [42, 43]. Therefore, we cannot interpret the decrease in relation between the SCCs and cognition, after adjustment for depression, as depression causing the SCCs.

Another source of the weak association of SCCs and cognition could lie in the discrepancy between the subjective and objective type of data, as proposed by a recent review [13]. SCCs, in general, refer to longitudinal change, whereas cross-sectional neuropsychological testing describes the cognitive capacity at a single time point, which may not be sensitive enough to capture a possible intraindividual cognitive change. In such cases, the subjective report does not correlate well with objective performance.

Specific complaints should be taken into account when trying to identify associations between SCCs and cognition. In our sample, poor memory was associated with two specific SCCs: Impression of worse memory in comparison to peers and Spatial orientation difficulties. This is congruent with former findings [6, 44].

Impression of worse memory in comparison peers has been considered a feature that increases the likelihood of preclinical AD and should raise concerns [13, 20]. In a previous study using the QPC, this complaint was commonly reported by patients with AD dementia (41.8%), but not by cognitively normal subjects. Spatial orientation difficulties was additionally associated with the lower MMSE score and was the rarest complaint in our sample. This is in agreement with the findings of previous studies. Amariglio et al. [6] also reported Spatial orientation difficulties as the rarest complaint, but associated with impaired memory and language. In a previous study using the QPC [21], no cognitively normal individual reported this difficulty, however, it was present in subjects with AD dementia. The early presence of spatial orientation impairment has already been observed [45] but probably is not usually detected when using standard neuropsychological methods. According to some recent findings, spatial navigation may be an independent cognitive domain which was not assessed in our study [46]. Based on our results, subjectively reported difficulties with spatial orientation in otherwise cognitively normal elderly should be heeded.

A previous population study which also included subjects with cognitive impairment [6] has shown a positive association between two memory complaints (difficulties in remembering recent events, experience of change in memory) and poor memory performance. Interestingly, in our sample of cognitively normal older adults, these two common memory complaints did not relate to memory. It has been postulated that during the course of AD, objective intraindividual change in memory precedes objective memory impairment [13]. From this point of view, subjective feeling of memory change may demonstrate subtle memory decline not captured by standard memory tests in a cross-sectional study. In our sample, Impression of memory change was reported by almost one-third of cognitively healthy older adults. With respect to its high prevalence, this complaint does not seem to be particularly specific for pathological aging. Still, further research in a clinical setting is needed.

The three remaining complaints, Losing things, Forgetting about appointments, and Forgetting about past experiences, were rarely reported and were not related to cognition. We assume that these complaints do not appear until advanced stages of cognitive deficit, which may explain our findings. Also, in a previous study using the QPC [21], these difficulties were reported by an insignificant number of controls in contrast to subjects with AD dementia.

Our current study is one of a few European projects with an ethnically homogeneous study sample. To overcome the limitations of previous studies, we applied a validated questionnaire for assessing SCCs and administered a complex cognitive battery that covered all cognitive domains for evaluation of cognitive performance. There are still several limitations that should be considered in the interpretation of the current results.

When recruiting cognitively healthy older participants for the study of healthy aging, intact activities of daily living and the absence of significant cognitive concerns were the basic inclusion criteria. These are also two of the core criteria of the revised Petersen’s criteria for MCI [47]. However, it is likely that when using a different approach to define cognitive impairment, as was recently proposed in the literature (e.g., [48]), participants with early cognitive deficit might have been included. Although we demonstrated associations between specific complaints and lower objective cognitive performance in this sample, these findings are not easily transferred to the clinical setting. As such, our findings should be confirmed in a future longitudinal study. Further research of specific complaints in patients with MCI and those meeting criteria for subjective cognitive decline [13] is needed.

In conclusion, cognitive complaints in cognitively healthy older adults are common. A higher number of SCCs in these well cognitively functioning individuals appears to be most closely related to depressive symptomatology, and this may hide a more subtle association between the number of SCCs and cognition. However, the causal relationship of SCCs and depressive symptoms needs further clarification, so the possible pathological significance of the SCCs in individuals with depressive symptoms should not be diminished.

Our findings support the notion that some complaints are part of normal aging while others could be specific for non-normative cognitive decline and should raise concerns. Therefore, when looking for early signs of prodromal AD, inquiring about specific cognitive complaints seems to be more relevant. Typical complaints may differ during the course of the disease. Therefore, the specific complaints rarely reported in our cognitively normal sample may be common in subjects with developed objective cognitive impairment. Using the QPC may help identify subjects that will need further neuropsychiatric examination.