In an effort to investigate the specific effects of inflammation on cognitive function, several studies have examined inflammatory markers and their association with cognition, cognitive decline, and onset of dementia (see summary of these studies in ). In a cross-sectional study, Wright and colleagues (2006)
examined serum levels of CRP, cytokines IL-1, IL-2, IL-6, TNF-α, and cytokine receptors (abbreviated ‘R’) IL-2R, TNFR-1, TNFR-2 in relation to Mini-Mental State Examination (MMSE) performance in a subset of participants enrolled in the Northern Manhattan Study. Findings indicated that IL-6 was negatively associated with MMSE score even after adjusting for sociodemographic and vascular risk factors, suggesting that high levels of inflammation may have direct negative effects on cognition. As the researchers pointed out, this study was limited by its cross-sectional nature and use of MMSE as the sole outcome measure of cognitive function (due to its lack of sensitivity to detect mild changes in cognition, especially for highly educated participants) (Wright et al., 2006
Summary of Studies Examining Inflammatory Biomarkers in Relation to Cognitive Function
A longitudinal examination of inflammatory markers and global cognitive decline was conducted by Yaffe and colleagues (2003)
in a large sample of older adults (42% African American) from the prospective Health, Aging, and Body Study. Using the Modified Mini-Mental State Examination (3MS) as a measure of cognitive function, these researchers evaluated serum levels of IL-6, CRP, and TNF-α at baseline in relation to baseline cognition and risk of cognitive decline over two years. Findings indicated that participants in the highest tertile of IL-6 or CRP serum concentrations performed significantly worse at baseline and follow-up 3MS, with a 24% increase in risk of cognitive decline over the two-year period in comparison to those participants in the lowest tertile (Yaffe et al., 2003
). TNF-α did not emerge as a significant marker of cognitive function or decline in this study. Additionally, there was no interaction between race and inflammation on cognition.
While the previously mentioned studies suggest that inflammation is associated with global cognitive function, they were unable to offer insight as to which domains of cognition may be most affected by inflammation. Baune and colleagues (2008)
examined circulating cytokines (serum IL-1β, IL-4R, IL-6, IL-8, IL-10, IL-12, and TNF-α) and specific neuropsychological domains of cognitive functioning – short-term memory, attention, cognitive, and motor speed – in a cross-sectional sample of 369 community-dwelling older adults enrolled in the Memory and Morbidity in Augsburg Elderly (MEMO) Study (Augsberg, Germany). These researchers found that higher levels of IL-8 were associated with poorer memory and motor function and with slower cognitive/perceptual speed performance. No significant associations were found for other cytokines or for other cognitive domains.
Ravaglia and colleagues (2003)
examined predictors of performance on the Clock Drawing Test, a measure of executive function and visuospatial ability, in a cohort from the large-scale Conselice Study comprised of community-dwelling Italian participants aged 65 and older. In a cross-sectional analysis, these researchers found a significant inverse association between performance on the Clock Drawing Test and serum CRP levels despite no evidence of overt inflammatory disease or acute infection in participants, which appears consistent with CRP’s role as a marker of underlying inflammation.
While results from the two studies previously discussed (Baune et al., 2008
; Ravaglia et al., 2003
) supported associations between proinflammatory biomarkers and different domains of cognitive function, they were limited by use of cross-sectional designs. Teunissen and colleagues (2003)
utilized longitudinal data from the Maastricht Aging Study in the Netherlands to investigate the relationship between serum inflammatory protein levels and change in cognitive performance. Haptoglobin and CRP (protein markers of inflammation) levels at baseline were examined in relation to individual cognitive performance over six years. The sample included cognitively healthy persons aged 30 – 82 years. The sample was evaluated as a whole, as well as stratified to include a subset of participants aged 50 and older. Measures of cognitive function included the Word Learning Task (a measure of word list learning and recall), Letter-Digit Coding Test (a measure of processing speed), and the Stroop Test (an attentional measure of perceptual interference and response inhibition where low scores are indicative of better performance). Results yielded a significant negative correlation between serum haptoglobin levels at baseline and performance on the Stroop Test and the Delayed Recall portion of the Word Learning Test at baseline and throughout the six-year follow-up period. Additionally, among the subset of participants aged 50 and older, CRP concentration was negatively correlated with performance on Delayed Recall and Word List Total Learning from the Auditory Verbal Learning Test at baseline and through six-year follow-up (Teunissen et al., 2003
). While participants differed significantly in cognitive function at baseline, there were no significant differences between persons in the course of cognitive function during follow-up, possibly due to the use of a relatively young sample that would not be expected to show much change in cognitive function; therefore, interactions between time and serum inflammation markers were not examined.
In similar research, Dik and colleagues (2005)
assessed serum inflammation markers associated with risk of cognitive decline over a three-year time period in a large sample (N
= 1,284) of older adults enrolled in the Longitudinal Aging Study Amsterdam. Inflammatory proteins included α1
-antichymotrypsin (ACT), CRP, IL-6, and albumin. Cognition was assessed on the following categories: general cognition (MMSE); memory (Auditory Verbal Learning Task, AVLT); fluid intelligence (Raven’s Colored Progressive Matrices); and information-processing speed (Coding Task). After adjusting for age, sex, and education, ACT was associated with baseline memory performance on the AVLT and 60% increased risk of clinically relevant decline on the MMSE. ACT was not predictive of decline on other cognitive measures. Neither CRP nor IL-6 was associated with cognitive performance or decline, which is in contrast to prior findings. The researchers suggested that low assay sensitivities for IL-6 (which was dichotomized around a selected detection limit), selective nonresponse of participants, and loss of follow-up of frail elders may have contributed to these discrepancies.
Using data from the MacArthur Study of Successful Aging, a longitudinal cohort study of 851 high-functioning older adults aged 70–79 years at baseline (1988) with follow-up assessments in 1991 and 1995, Alley, Crimmins, Karlamangla, Hu, and Seeman (2008)
examined individual growth curves from baseline and follow-up performances on measures of abstraction (Wechsler Adult Intelligence Scale-Revised [WAIS-R] subtests), language (abbreviated Boston Naming Test), spatial ability (copying geometric figures), verbal recall (recall of naming items from the Boston Naming Test), spatial recognition (WAIS-R Spatial Span), and global cognitive function (summary measure derived from all subtest scores) as well as performance on the nine-item Short Portable Mental Status Questionnaire based on age, serum levels of IL-6 and CRP, and sociodemographic and health characteristics. These researchers found evidence for a cross-sectional linear relationship between inflammation and cognition, with higher levels of inflammation related to lower levels of baseline cognitive function (specifically, IL-6 with language and the global summary score, and both IL-6 and CRP with abstraction). After controlling for potential confounding variables, growth curve analysis yielded no effect of inflammation on baseline cognitive function or the rate of longitudinal cognitive change. However, when these researchers stratified inflammatory biomarkers into high-risk categories, participants in the top IL-6 tertile had 52% increased odds of decline in abstraction, 62% increased odds of decline in global cognitive function, and 88% increased odds of cognitive impairment as indicated by a decline of more than two points on the Short Portable Mental Status Questionnaire relative to those in the lowest IL-6 tertile.
In an effort to investigate predictors of maintenance of cognitive function (as opposed to decline), Yaffe, Fiocco, and colleagues (2009)
studied data for 2,509 well-functioning African American and white older adults enrolled in the large-scale longitudinal Health, Aging and Body Composition (Health ABC) Study. Cognitive function was measured using the 3MS at baseline and years three, five, and eight, and cytokines IL-6, CRP, and TNF-α were used as serum markers of inflammation. These researchers found that older adults with lower levels of IL-6 and CRP were more likely to maintain their baseline levels of cognitive function over the eight-year follow-up period. Additionally, behavioral factors played an important role in maintenance of cognitive function moderate exercisers, non-smokers, infrequent drinkers of alcohol, and those who worked or volunteered regularly were more likely to maintain baseline levels of cognitive function. Notably, these contributing behavioral factors can be classified as potentially modifiable (Yaffe, Fiocco, et al., 2009
In the above-mentioned studies, cognitive outcome measures and mean age of study groups were heterogeneous and protocols varied in terms of study design – including candidate serum proteins and stratification of assays (e.g., dichotomization versus tertiles) – and length of follow-up, making individual comparisons difficult. Furthermore, while studies may include measures of cognition in relation to inflammation, there is a paucity of research examining the interrelationships between psychological and health factors, inflammation, and cognitive function, especially with respect to changes over time. Overall, findings from existing studies generally do support an association between increased risk of cognitive decline and high inflammatory concentrations, but there is a need for further analysis of potential interrelationships between inflammation, cognitive function, and psychological and health factors.