Characteristics of Cognitively Normal Mexican-Americans with Cognitive Complaints

INTRODUCTION

Individuals with subjective cognitive complaints (SCC) (also referred to as subjective cognitive decline) perceive themselves as having experienced some decline in cognitive functioning in the absence of objective findings on neuropsychological testing. Subjective cognitive complaints have been viewed as a risk factor [1, 2] for, and a very early symptom of, later cognitive decline and dementia [3]. SCC has been associated with an increased risk for incident cognitive loss over time though there remains insufficient evidence to suggest SCC should be a required component for defining prodromal Alzheimer’s disease [4]. Kaup et al. [5] investigated the long-term risk for cognitive impairment for woman who had SCC and found SCC strongly related to cognitive impairment almost two decades later. Although by definition cognitively healthy individuals with SCC score in the normal range for age and education on neuropsychological testing, differences have been found between healthy older adults with and without SCC. Studies on non-Hispanics have found that individuals with SCC have poorer temporal order memory [6] and poorer immediate and delayed verbal recall [3]. Subtle changes in the ability to carry out everyday activities have also been related to the presence of SCC [6, 7].

A number of affective symptoms, health risks, and biomarkers have been related to SCC. A strong link has been found between affective states, especially depression and SCC [8–10]. Health risks including cardiovascular risks [12, 13] and diabetes [14] have been related to SCC. Specific imaging and cerebrospinal fluid biomarkers related to Alzheimer’s disease have been found in individuals with SCC [15] and blood-based biomarkers [16] have been shown to predict conversion to dementia.

The research that has been done on SCC has been conducted on a variety of clinical and community based populations. To date, none of this work has focused on the rapidly expanding population of Mexican-American elders. Previously our group has identified the distinct characteristics of Mexican-Americans with mild cognitive impairment (MCI) [17] and Alzheimer’s disease [18]. These studies found that the risk factors for MCI are distinct for Mexican-Americans and include more symptoms of depression and increased likelihood of being diagnosed with diabetes. Age was the only variable consistently linked to cognitive impairment (MCI and Alzheimer’s disease) among Mexican-Americans. A study of Mexican-Americans with amnestic MCI found a biomarker profile that was primarily inflammatory with TNFα, IL 10, and TARC being the top three biomarkers [19]. Multiple metabolic and vascular conditions have been linked to cognitive decline among Mexican-Americans [20]. None of the studies have focused on Mexican-Americans with subjective cognitive complaints. The present study sought to identify the characteristics of community-dwelling older Mexican-Americans who voice concerns about changes in cognition in the absence of objective evidence of cognitive impairment.

METHODS

Data from 319 participants (254 females, 65 males) from the Health and Aging Brain among Latino Elders (HABLE) study who had been diagnosed as cognitively normal were analyzed. HABLE is an ongoing epidemiological study of cognitive aging among community-dwelling Mexican Americans that has been described in detail elsewhere [21]. The current sample is predominately female. The HABLE protocol was designed to collect a wide range of data including clinical characteristics, anthropomorphic measures, affective status, neuropsychological functioning, and proteomic markers. In brief, the HABLE participants undergo a medical examination, clinical interview, neuropsychological assessment, and blood draw for clinical and biomarker analysis. An informant interview is conducted to assess functional level. Data on alcohol consumption [22] and sleep [23] are collected during the clinical interview. Participants complete the Geriatric Depression Scale (GDS) [24], the Beck Anxiety Inventory (BAI) [25], and the Penn State Worry Questionnaire (PSWQ) [26] for measures of affective distress. In addition, anthropomorphic measures such as height and weight are collected to calculated participant’s body mass index.

Neuropsychological testing is completed in English or Spanish depending on the participant’s preference. In the current study, 72 participants completed the protocol in English and 247 in Spanish. The neuropsychological battery consisted of tests of global cognition (Mini-Mental State Examination (MMSE) [27]; executive functioning (Trail-Making Test Part B [28], Executive Interview (EXIT25) [29], clock drawing (CLOX1) [32]); language (FAS and Animal Naming) [31]; visuospatial skills (CLOX2) [32]; memory (Wechsler Memory Scale, Third Edition (WMS-3) Logical Memory 1 & 2 [33], Consortium for the Establishment of Registry for Alzheimer’s Disease (CERAD) List Learning [34], AVLT [35]); and attention (WMS-III Digit Span, Trail-Making Test Part A). A consensus diagnosis was assigned based on NINCDS-ADRDA [36] criteria for Alzheimer’s disease and on the Petersen [37] criteria for MCI. Those participants who perform within normal parameters on psychometric testing and have a Clinical Dementia Rating [38] (CDR) = 0 were classified as cognitively normal and the presence of subjective cognitive complaints was determined from the clinical interview. Each participant was asked if they were concerned about changes in memory and thinking. Those responding in the positive were classified as having subjective cognitive complaints. Institutional Review Board approval was attained for HABLE and written informed consent was obtained from all participants included in this study.

RESULTS

Table 1 presents the descriptive statistics for the sample. The two groups did not differ on age (F = 2.05, p = 0.153) or education (F [1, 317] = 0.452, p = 0.502). Both groups were composed primarily of females although the percent of females was significantly higher in the SCC group (χ² = 3.85, p = 0.04). The two groups did not differ on body mass index (F [1, 317] = 2.32, p = 0.130). The SCC group scored significantly lower on the MMSE (F [1, 317] = 5.63, p = 0.018) although both groups scored within a range that would be defined as normal. The SCC endorsed significantly more depressive symptoms on the GDS than did the No SCC group (F [1, 317] = 28.94, p = 0.000). A significantly higher percent of SCC scored in the depressed range (≥10) on the GDS (χ² = 8.88, p = 0.002; OR = 2.014, 95% CI = 1.27 – 3.30). The SCC endorsed significantly more symptoms of anxiety (F [1, 317] = 37.71, p = 0.000) and a significantly higher percent of SCC scored in the anxious range (≥10) on the BAI (χ² = 29.85, p < 0.0001; OR = 5.03, 95% CI = 2.72–9.28). The SCC scored significantly higher on the PSWQ (F [1, 317] = 33.06, p = 0.000). The SCC reported significantly more symptoms of sleep difficulties (F [1, 317] = 9.47, p = 0.002).

Tables 2–4 present the results of neuropsychological testing. The memory measures from the battery are presented in Table 2. As shown there were no significant differences between the SCC group and the No SCC group on any of the tests of memory. Table 3 presents the language measures. The SCC group produced significantly fewer total words on FAS (F [1, 306] = 7.41, p = 0.007) and had a significantly lower scale score (F [1, 306] = 11.44, p = 0.001) than the No SCC group. No differences were found on any of the other language measures. Table 4 presents the results for measures of attention and executive functioning. The SCC group took significantly longer (F [1, 270] = 5.62, p = 0.018) to complete Trails A and had a significantly lower scale score (F [1, 270] = 5.34, p = 0.021) than the No SCC group. No differences were found on any of the other measures of attention and executive functioning.

In order to evaluate the potential impact of health, the results of an extensive clinical blood work panel were analyzed comparing the SCC group with the No SCC group. Age was co-varied in all analyses. The two groups did not differ on any of the measures of liver function (albumin, bilirubin, albumin/globulin ratio, alkaline phosphatase, AST, ALT). No significant differences were found on measures of kidney function (GFR, creatinine and BUN). No differences were found on measures of thyroid functioning (TSH, T4). The two groups did differ significantly on blood glucose (SCC M = 129.96 SD = 61.21; No SCC M = 118.15 SD = 49.18, F = 3.89, p = 0.049) and HβA1 C (SCC M = 6.93 SD = 1.84; No SCC M = 6.56 SD = 1.64, F [1, 316] = 4.17, p = 0.042). Analysis of the cholesterol panel (total cholesterol, LDL, HDL, triglycerides) revealed no significant differences between the groups on any of the measures although a significantly higher percent of SCC individuals reported having high cholesterol (χ² = 5.21, p = 0.02; OR = 1.677, 95% CI = 1.074 – 2.617). The groups did not differ on self-reported heart disease or hypertension. A significantly higher percent of SCC individuals reported having been told by a physician that they had diabetes (χ² = 6.74, p = 0.009; OR = 1.845, 95% CI = 1.159–2.935).

The results for the blood-based serum biomarkers assayed are presented in Table 5. Among the 21 biomarkers assayed significant differences were found for VCAM a marker of microvascular pathology and three inflammatory biomarkers – Serum Amyloid A (SAA), Interluekin 6 (IL-6), and Tumor Necrosis Factor alpha (TNFα). The SCC group had significantly higher levels of SAA (F [1, 299] = 4.347, p = 0.038), VCAM (F [1, 299] = 4.770, p = 0.041) and IL-6 (F [1, 299] = 4.980, p = 0.017). Unexpectedly, the SCC group had a significantly lower level of TNFα (F [1, 299] = 10.471, p = 0.001.

DISCUSSION

Prior research has shown a consistent relationship between SCC and affective symptoms [11] and quality of life [42]. Subjective concern with cognitive decline has been found to correlate with depression rather than with objective cognitive decline in both clinic and community-based samples [11, 47]. In our community-based sample of older Mexican-Americans, those with SCC experienced significantly higher levels of depression, anxiety and worry than those without concerns. They were two times more likely to score in the clinically depressed range and five times more likely to score in the clinically anxious range. SCC individuals with worry have been shown to be at increased risk for conversion to dementia [1, 48]. In our sample, it is likely that the depression and anxiety as well as the higher level of worry are related to fears of loss of function and decline in cognition. The greater number of sleep related difficulties may reflect the impact of the affective symptoms on sleep [43]. In earlier work, our group has shown a relationship between comorbid depression and diabetes as a risk for cognitive impairment in Mexican-Americans [21]. In the current research, we found an increased likelihood for those with normal cognition and SCC to be diagnosed with diabetes as well as having significantly higher levels of clinical markers of elevated blood sugar. This suggests that even before objective cognitive decline has occurred individuals with depression and diabetes perceive change in their cognitive ability.

Neuropsychological testing revealed few differences between the groups. We found no differences on memory measures although the difference found on MMSE appears attributable to performance on memory items. Differences between the two groups were found on two timed measures—FAS and Trails A. These tests that have been shown to assess executive functions as well as speed [44]. Heringa et al. [45] found that low grade inflammation and endothelial dysfunction were contributing factors in reduced information processing speed and executive functioning in older individuals. In our study, we found two inflammatory markers, SAA and IL-6, along with a marker of endothelial dysfunction (VCAM) were higher in individuals with SCC. The lower level of TNF-α is not consistent with the other findings but it is possible that some dysregulation that affects neuronal processes [46] may be accounting for the difference. Over time, it is possible that compensatory mechanisms in response to systemic inflammation may lead to an elevation in TNF-α.

The results of the current research have begun to identify the affective symptoms, health risks, and biomarkers that differentiate older Mexican-Americans who voice concerns about changes in cognition in the absence of objective evidence of cognitive decline from those who do not. A number of limitations impact the study. First, our sample size was relatively small. The sample was a community sample and the findings may well differ from those found in clinical cohorts [47]. Second, and most important for the application of the findings, the research was cross-sectional and therefore the impact of the findings on the development of cognitive impairment could not be assessed. We are currently following this cohort longitudinally, which will allow us to evaluate the association between SCC, diabetes, depression, anxiety, inflammatory processes, and cognitive decline in Mexican-Americans. Despite these limitations, it is possible that the features that differentiate SCC may be early indicators of later cognitive decline and should be evaluated as potential areas of intervention.