Is Memory Decline Associated With Inflammatory Response?

Abstract

Objective: To examine whether changes in memory over a 10-year period could predict a change in C-reactive protein (CRP) levels. Method: A mixed model analysis was first conducted to obtain the estimates for change in memory over the 10-year period using data from the Health and Retirement Study. Then a multivariate regression to determine whether a change in episodic memory could predict subsequent CRP levels was conducted. Furthermore, a general linear model was conducted to determine differences in CRP levels among different rates of change in episodic memory. Results: Greater declines in episodic memory were associated with higher levels of subsequent CRP (Estimate = −0.32, SE = 0.12, β = −.03, p = .008). The general linear model revealed that those with greater memory declines were more likely to have higher levels of CRP, F = 26.50, p < .001. Discussion: These results highlight the notion that memory decline and inflammation may be intertwined, and we discuss various avenues that warrant further investigation.

Keywords

cognitive function cognitive status biomarkers

C-reactive protein (CRP) is an acute-phase protein that is released early in the body’s inflammatory response (Du Clos & Mold, 2004; Pepys & Hirschfield, 2003). Research has identified that CRP is predictive of a variety of chronic and acute health conditions, such as diabetes (Dehghan et al., 2007; Pradhan, Manson, Rifai, Buring, & Ridker, 2001), hypertension (Dehghan et al., 2007), heart failure (Athilingam et al., 2013), chronic obstructive pulmonary disease (Broekhuizen, Wouters, Creutzberg, & Schols, 2006), upper respiratory tract infection (Melbye, Hvidsten, Holm, Nordbø, & Brox, 2004), and depression (Danner, Kasl, Abramson, & Vaccarino, 2003; Valkanova, Ebmeier, & Allan, 2013). In general, higher CRP levels are suggested to signal higher inflammatory response.

When activated, CRP induces the synthesis of additional proinflammatory markers, yet some studies have shown that CRP is a more sensitive measure for identifying active inflammation (Frost, Roach, Kushner, & Schreiber, 2005; Sheldon, Riches, Gooding, Soni, & Hobbs, 1993). The assessment of CRP is relatively easy and noninvasive, thus making it a relatively useful indicator of overall physical health (Kao, Shiesh, & Wu, 2006).

In addition to the relationships among CRP and health conditions, researchers have also identified a link between CRP and cognition, with higher levels of CRP being associated with greater decline in various measures of cognitive functioning (Dimopoulos et al., 2006; McGeer & McGeer, 2001; Yasojima, Schwab, McGeer, & McGeer, 2000). However, to the best of our knowledge, no study to date has assessed whether memory decline may reflect subsequent higher levels of CRP, whereby indicating more global, physiological change rather than cognitive decline alone. The notion that memory decline may represent an important indicator of subsequent global health changes has been given more attention in recent research. For example, Terhorst, Holm, Toto, and Rogers (2017) reported that global cognition and executive function predicted mobility difficulties 6 months later. Yaffe et al. (2016) recently reported that steeper trajectories of change in the same domains over about two decades were predictive of a subsequent greater risk of mortality. Finally, Batty, Deary, and Zaninotto (2016) reported that poor performance in executive function, memory, and speed was a significant indicator of early death due to a wide variety of chronic conditions 9 years later.

Building on previous research, the purpose was to use data from the nationally representative Health and Retirement Study (HRS) to conduct an exploratory analysis of the potential relationship between changes in episodic memory, a known early marker of cognitive impairment (Bäckman, Jones, Berger, Laukka, & Small, 2005), over a 10-year period and subsequent CRP levels. It was hypothesized that steeper declines in episodic memory would predict higher levels of CRP. In addition, the interactions among episodic memory and CRP with depressive symptoms (Danner et al., 2003; Valkanova et al., 2013), a greater waistline circumference (Behan & Mbizo, 2007), and the number of health conditions were examined. It was hypothesized that higher depressive symptoms, greater waistline circumference, and increased amount of health conditions would strengthen the association between greater decline in episodic memory and higher CRP levels.

Method

Participants

Data were obtained from 1996 to 2006 biannual waves of the HRS. The HRS is a nationally representative longitudinal panel study conducted at the University of Michigan. The current study included 6,213 participants aged 30 years and older with an average age of 68 years who were predominantly Caucasian and female with at least a high school education. Regarding the number of health conditions, the participants reported, approximately only 14% reported no health conditions, 25% reported one health condition, 26% reported two, and 38% reported three or more. See Table 1 for additional baseline characteristics.

Table 1.

Descriptive Statistics.

Variable	M (SD) or %	Range
CRP (mg/L)	4.69 (9.51)	0.03-280
Age (years)	67.54 (10.63)	30-98
Sex (female)	57.97
Race (Caucasian)	85.98
Education (years)	12.59 (3.13)	0-17
Depressive symptoms	1.46 (1.97)	0-8
Health conditions	2.06 (1.44)	0-8
Waistline circumference (inches)	39.71 (5.98)	24.50-66

Note. N = 6,212. CRP = C-reactive protein.

Measures

CRP

CRP was analyzed by a dried blood spot test to measure the levels of inflammation (Crimmins et al., 2008). The dried blood spot samples were mailed to Biosafe labs in Chicago and the University of Vermont where the assays were completed. Only the 2006 wave included measures of CRP, and these data were included in the analyses.

Episodic memory

Episodic memory was measured by the total amount of words recalled in both immediate and delayed tasks (Ofstedal, Fisher, & Herzog, 2005). Biennial waves from 1996 to 2006 of episodic memory were included in the analysis. To measure a change in cognition, the slope estimates that were saved were subsequently used in multivariate regression models to investigate whether steeper changes in cognition predict levels of CRP.

Covariates

Covariates included age (continuous), gender (male/female), ethnicity/race (Caucasian, African American, or other), education (number of school years), the number of depressive symptoms, waistline circumference (continuous), and the number of health conditions (ordinal). Depressive symptoms were assessed by the Center for Epidemiologic Studies Depression Scale (CES-D), an eight-item scale where higher scores indicate more depressive symptoms (Radloff, 1977; Steffick, 2000).

Statistical Analysis

SAS 9.4 software was used to conduct all analyses. The analytical procedure included two steps to examine whether changes in episodic memory predict levels of CRP. First, mixed effects models were used to obtain intercept and slope estimates for episodic memory performance measured biennially between 1996 and 2006. More specifically, a linear model of change was estimated, with the intercept representing levels of episodic memory being set in 2006 and the slope representing changes over the 10-year period. Participants with missing data for episodic memory were excluded from further analysis. A total of 216 participants had missing data for waistline circumference and the sample mean was imputed. Z scores were used to standardize the rate of change in episodic memory and for cognition in 1996. Second, multivariate regression analyses were conducted to determine whether changes in cognition between 1996 and 2006 would predict CRP levels in 2006. In addition, the covariates were examined for any interactions with a change in episodic memory and CRP.

To further examine the rates of declines, a general linear model was conducted to determine group differences in CRP based on the different trajectories of change in episodic memory. Change in episodic memory was categorized based on quartiles values. The first quartile was −0.50 or lower, second quartile ranged from −0.50 to 0.07, third quartile ranged from 0.07 to 0.60, and the fourth quartile was 0.60 and greater. Ad hoc analyses were conducted with Tukey’s HSD (honest significant difference) to determine significant differences among the groups.

Results

Of the 6,213 participants, one was removed due to a coding error in the database for CES-D; thus, resulting in a total of 6,212 participants included for analyses.

Results from the multivariate regression model adjusted for cognition in 2006, age, sex, education, race, depression, health conditions, and waistline circumference are presented in Table 2. These results indicated that a change in cognition over a 10-year period did predict subsequent CRP levels. The rate of change, sex, education, depression, health conditions, and waistline circumference covariates were significant to the model, whereas age, race, and cognition in 2006 were not (Table 2). When examining interactions, there were no significant interactions found among any of the covariates (ps > .05).

Table 2.

Parameter Estimates for Change in Cognition to Predict CRP Levels in a Multivariate Regression.

	Estimate	SE	β	p value
Intercept	−4.13	1.44	.00	.004
Rate of change from 1996 until 2006	−0.32	0.12	−.03	.008
Episodic memory in 2006	−0.01	0.15	−.001	.968
Age	−0.02	0.01	−.01	.180
Sex	1.51	0.26	.08	<.001
Race	0.22	0.12	.02	.068
Education	−0.14	0.04	−.04	<.001
Depression	0.23	0.07	.05	<.001
Health conditions	0.41	0.09	.05	<.001
Waistline circumference	0.24	0.02	.15	<.001

Note. CRP = C-reactive protein; β = standardized beta coefficient.

To compare the group differences in CRP based on the different trajectories of change in cognition, a general linear model was conducted that adjusted for age, sex, education, race, depression, health conditions, and waistline circumference. There was a main effect for group, F = 26.50, p < .001. Tukey’s HSD post hoc analysis indicated that individuals who had the greatest amount of cognitive decline had significantly higher levels of CRP compared with the two groups that had the lowest amounts of change in cognition (see Figure 1).

Figure 1.

Tukey’s HSD post hoc analysis of predicted CRP levels based on the different trajectories of change in episodic memory.

Discussion

The current study examined whether changes in episodic memory over a 10-year period would be associated with a subsequent inflammatory response as assessed by levels of CRP in a nationally representative sample. The results indicated that a steeper decline in episodic memory over a decade prior was associated with greater CRP levels. These findings provide unique evidence suggesting that changes in episodic memory may mark not only oncoming cognitive problems but may also indicate broader physiological changes. The findings support recent research suggesting that cognitive performance and decline reflect the risk of mobility problems (Terhorst et al., 2017) and mortality from a wide variety of causes (Batty et al., 2016; Yaffe et al., 2016).

Previous research has focused on identifying the relationship between CRP levels and subsequent cognitive decline (Bettcher et al., 2012; Dlugaj et al., 2012; Gunstad et al., 2006; Komulainen et al., 2007; Mancinella et al., 2009; Noble et al., 2010; Schuitemaker et al., 2009; Xu, Zhou, Zhu, Fan, & Liu, 2009), finding that a relatively consistent significant association between higher CRP and cognitive decline may exist. Together with the findings reported here, one can assume that the link between cognition and CRP may have important implications for assessing and treating individuals experiencing cognitive decline. Specifically, for individuals coming to memory clinics with complaints about cognitive decline, a comprehensive assessment that goes beyond cognitive assessment may be useful to help identify underlying causes and/or consequences of cognitive decline among older adults.

The strengths of the study included, first, the use of a nationally representative sample to examine the prediction of CRP which allows for greater generalizability, and second, the use of biennial total recall scores over a 10-year period, thus allowing for an examination of episodic memory over time.

The study is not without limitations. First, CRP was the only measure of inflammation, and it is known for CRP to fluctuate for a number of reasons (Du Clos & Mold, 2004; Macy, Hayes, & Tracy, 1997). Given this, it may be that the association between cognition and inflammation would be stronger with a more sensitive and stable measurement of inflammation. However, CRP has been shown to reflect active inflammation (Frost et al., 2005; Sheldon et al., 1993). Further research conducted by Ballou and Lozanski (1992) further found that when CRP levels were equivalent to moderate or severe inflammation, there was a rapid induction of cytokines. However, when there were low levels of CRP, there was a noticeable decrease in the amount of cytokines produced; therefore, supporting CRP as a useful indicator of systemic inflammation.

Second, episodic memory was the only measure of cognition accessed for cognitive decline. At the same time, episodic memory is a well-established early indicator of cognitive impairment, hence making it particularly relevant for this study. As evidenced by a meta-analysis conducted by Bäckman et al. (2005), episodic memory is a cognitive domain that is affected in the early stages of cognitive impairment that precedes a diagnosis of Alzheimer’s disease.

Third, the measure of health conditions was only for 2006. This measure does not allow for the potential changes in the number of health conditions during the duration of the current study. Future studies need to examine the potential changes in the number of health conditions in coordination with changes in CRP.

In conclusion, when examined over a 10-year period, changes in episodic memory were associated with particularly high levels of CRP. It may be that age-related memory decline coincides with a global physiological decline expressed by a heightened inflammatory response, creating a new avenue for a potentially more effective intervention that targets overall health rather than cognition alone. These notions may deserve the attention of future studies.

Footnotes

Declaration of Conflicting Interests

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

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

References

Athilingam

Moynihan

Chen

D’Aoust

Groer

Kip

(2013). Elevated levels of interleukin 6 and C-reactive protein associated with cognitive impairment in heart failure. Congestive Heart Failure, 19, 92-98.

Bäckman

Jones

Berger

A. K.

Laukka

E. J.

Small

B. J.

(2005). Cognitive impairment in preclinical Alzheimer’s disease: A meta-analysis. Neuropsychology, 19, 520-531. doi:10.1037/0894-4105.19.4.520

Ballou

S. P.

Lozanski

(1992). Induction of inflammatory cytokine release from cultured human monocytes by C-reactive protein. Cytokine, 4, 361-368.

Batty

G. D.

Deary

I. J.

Zaninotto

(2016). Association of cognitive function with cause-specific mortality in middle and older age: Follow-up of participants in the English longitudinal study of ageing. American Journal of Epidemiology, 183, 183-190.

Behan

K. J.

Mbizo

(2007). The relationship between waist circumference and biomarkers for diabetes and CVD in healthy non-obese women. The Pensacola study. Laboratory Medicine, 38, 422-427. doi:10.1309/VV1N2FGHNT84JKJF

Bettcher

B. M.

Wilheim

Rigby

Green

Miller

J. W.

Racine

C. A.

. . . Kramer

J. H.

(2012). C-reactive protein is related to memory and medial temporal brain volume in older adults. Brain, Behavior, and Immunity, 26, 103-108. doi:10.1016/j.bbi.2011.07.240

Broekhuizen

Wouters

E. F. M.

Creutzberg

E. C.

Schols

A. M. W. J

. (2006). Raised CRP levels mark metabolic and functional impairment in advanced COPD. Thorax, 61, 17-22. doi:10.1136/thx.2005.041996

Crimmins

Guyer

Langa

Ofstedal

M. B.

Wallace

Weir

(2008). Documentation of biomarkers in the Health and Retirement Study. Ann Arbor: University of Michigan.

Danner

Kasl

S. V.

Abramson

J. L.

Vaccarino

(2003). Association between depression and elevated C-reactive protein. Psychosomatic Medicine, 65, 347-356. doi:10.1097/01.PSY.0000041542.29808.01

10.

Dehghan

Kardys

de Maat

M. P. M.

Uitterlinden

A. G.

Sijbrands

E. J. G.

Bootsma

A. H.

. . . Witteman

J. C. M.

(2007). Genetic variation, C-reactive protein levels, and incidence of diabetes. Diabetes, 56, 872-878. doi:10.2337/db06-0922

11.

Dimopoulos

Piperi

Salonicioti

Mitropoulos

Kallai

Liappas

. . . Kalofoutis

(2006). Indices of low-grade chronic inflammation correlate with early cognitive deterioration in an elderly Greek population. Neuroscience Letters, 398, 118-123. doi:10.1016/j.neulet.2005.12.064

12.

Dlugaj

Gerwig

Wege

Siegrist

Mann

Bröcker-Preuß

. . . Bokhof

(2012). Elevated levels of high-sensitivity C-reactive protein are associated with mild cognitive impairment and its subtypes: Results of a population-based case-control study. Journal of Alzheimer’s Disease, 28, 503-514. doi:10.3233/JAD-2011-111352

13.

Du Clos

T. W.

Mold

. (2004). C-reactive protein. Immunologic Research, 30, 261-277. doi:10.1385/IR:30:3:261

14.

Frost

Roach

M. J.

Kushner

Schreiber

(2005). Inflammatory C-reactive protein and cytokine levels in asymptomatic people with chronic spinal cord injury. Archives of Physical Medicine and Rehabilitation, 86, 312-317.

15.

Gunstad

Bausserman

Paul

R. H.

Tate

D. F.

Hoth

Poppas

. . . Cohen

R. A.

(2006). C-reactive protein, but not homocysteine, is related to cognitive dysfunction in older adults with cardiovascular disease. Journal of Clinical Neuroscience, 13, 540-546. doi:10.1016/j.jocn.2005.08.010

16.

Kao

P. C.

Shiesh

S. C.

T. J.

(2006). Serum C-reactive protein as a marker for wellness assessment. Annals of Clinical & Laboratory Science, 36, 163-169.

17.

Komulainen

Lakka

T. A.

Kivipelto

Hassinen

Penttilä

I. M.

Helkala

. . . Rauramaa

(2007). Serum high sensitivity C-reactive protein and cognitive function in elderly women. Age and Ageing, 36, 443-448. doi:10.1093/ageing/afm051

18.

Macy

E. M.

Hayes

T. E.

Tracy

R. P.

(1997). Variability in the measurement of C-reactive protein in healthy subjects: Implications for reference intervals and epidemiological applications. Clinical Chemistry, 43, 52-58.

19.

Mancinella

Carpinteri

Bellomo

Fossati

Gianturco

. . . Marigliano

(2009). Is there a relationship between high C-reactive protein (CRP) levels and dementia? Archives of Gerontology and Geriatrics, 49, 185-194. doi:10.1016/j.archger.2009.09.028

20.

McGeer

P. L.

McGeer

E. G.

(2001). Inflammation, autotoxicity and Alzheimer disease. Neurobiology of Aging, 22, 799-809. doi:10.1016/S0197-4580(01)00289-5

21.

Melbye

Hvidsten

Holm

Nordbø

S. A.

Brox

(2004). The course of C-reactive protein response in untreated upper respiratory tract infection. The British Journal of General Practice, 54, 653-658.

22.

Noble

J. M.

Manly

J. J.

Schupf

Tang

M. X.

Mayeux

Luchsinger

J. A.

(2010). Association of C-reactive protein with cognitive impairment. Archives of Neurology, 67, 87-92. doi:10.1001/archneurol.2009.308

23.

Ofstedal

M. B.

Fisher

G. G.

Herzog

A. R.

(2005). Documentation of cognitive functioning measures in the Health and Retirement Study. Ann Arbor: University of Michigan.

24.

Pepys

M. B.

Hirschfield

G. M.

(2003). C-reactive protein: A critical update. Journal of Clinical Investigation, 111, 1805-1812. doi:10.1172/JCI200318921

25.

Pradhan

A. D.

Manson

J. E.

Rifai

Buring

J. E.

Ridker

P. M.

(2001). C-reactive protein, interleukin 6, and risk of developing type 2 diabetes mellitus. Journal of the American Medical Association, 286, 327-334. doi:10.1001/jama.286.3.327

26.

Radloff

L. S.

(1977). The CES-D Scale: A self-report depression scale for research in the general population. Applied Psychological Measurement, 1, 385-401. doi:10.1177/014662167700100306

27.

Schuitemaker

Dik

M. G.

Veerhuis

Scheltens

Schoonenboom

N. S. M.

Hack

C. E.

. . . Jonker

(2009). Inflammatory markers in AD and MCI patients with different biomarker profiles. Neurobiology of Aging, 30, 1885-1889. doi:10.1016/j.neurobiolaging.2008.01.014

28.

Sheldon

Riches

Gooding

Soni

Hobbs

J. R.

(1993). C-reactive protein and its cytokine mediators in intensive-care patients. Clinical Chemistry, 39, 147-150.

29.

Steffick

D. E.

(2000). Documentation of affective functioning measures in the Health and Retirement Study. Ann Arbor: University of Michigan.

30.

Terhorst

Holm

M. B.

Toto

P. E.

Rogers

J. C.

(2017). Performance-based impairment measures as predictors of early-stage activity limitations in community-dwelling older adults. Journal of Aging and Health, 29, 880-892.

31.

Valkanova

Ebmeier

K. P.

Allan

C. L.

(2013). CRP, IL-6 and depression: A systematic review and meta-analysis of longitudinal studies. Journal of Affective Disorders, 150, 736-744. doi:10.1016/j.jad.2013.06.004

32.

Zhou

Zhu

Fan

Liu

(2009). Plasma C-reactive protein is related to cognitive deterioration and dementia in patients with mild cognitive impairment. Journal of the Neurological Sciences, 284, 77-80. doi:10.1016/j.jns.2009.04.018

33.

Yaffe

Peltz

C. B.

Ewing

S. K.

McCulloch

C. E.

Cummings

S. R.

Cauley

J. A.

. . . Ensrud

K. E.

(2016). Long-term cognitive trajectories and mortality in older women. Journals of Gerontology, Series A: Biomedical Sciences and Medical Sciences, 71, 1074-1080.

34.

Yasojima

Schwab

McGeer

E. G.

McGeer

P. L.

(2000). Human neurons generate C-reactive protein and amyloid P: Upregulation in Alzheimer’s disease. Brain Research, 887, 80-89. doi:10.1016/S0006-8993(00)02970-X