Association between C reactive proteinlevel and depressive symptoms in anelderly Korean population: KoreanSocial Life, Health and Aging Project

Bo Mi Song,1,2 Ju-Mi Lee,2,3 Wungrak Choi,4 Yoosik Youm,5 Sang Hui Chu,6

Yeong-Ran Park,7 Hyeon Chang Kim2,3

To cite: Song BM, Lee J-M,Choi W, et al. Associationbetween C reactive proteinlevel and depressivesymptoms in an elderlyKorean population: KoreanSocial Life, Health and AgingProject. BMJ Open 2015;5:e006429. doi:10.1136/bmjopen-2014-006429

Prepublication history andadditional material isavailable. To view please visitthe journal (http://dx.doi.org/10.1136/bmjopen-2014-006429).

Received 20 August 2014Revised 23 January 2015Accepted 3 February 2015

For numbered affiliations seeend of article.

Correspondence toDr Hyeon Chang Kim;hckim@yuhs.ac

ABSTRACTObjective: The relationship between inflammatorymarkers and depressive symptoms has been reportedinconsistently. Moreover, there were only a few studiesconducted in an Asian population. The purpose of thisstudy was to examine the association between Creactive protein (CRP) and depressive symptoms in anelderly Korean population.Design, setting and participants: This study useddata from the Korean Social Life, Health and AgingProject Health Examination Cohort, which started in2011. Among participants aged 60 or over recruitedfrom a rural community, 569 (224 men and 345women) without a history of stroke, angina pectoris,myocardial infarction or CRP20 mg/L were employedfor cross-sectional analyses. As a marker of systemicinflammation, CRP was measured. Depressivesymptoms were measured with the Center forEpidemiologic Studies Depression (CES-D) scale.Multiple linear regression analysis was used toinvestigate the relationship between CRP anddepressive symptoms.Results: In men, CRP levels had significantassociations with depressive symptoms before(=0.420, p=0.010) and after (=0.336, p=0.025)adjusting for age, body mass index, systolic bloodpressure, number of comorbidities, smoking status,alcohol intake, marital status, education and sleepduration. However, in women, the association betweenCRP and depressive symptoms was not significantbefore (p=0.250) and after (p=0.256) adjustment.Conclusions: Our findings suggest that elevated CRPlevels are independently associated with the presenceof depressive symptoms in elderly Korean men.

INTRODUCTIONThe elderly population is growing worldwideand has been accompanied by a concurrentincrease in physical or psychological disabil-ities.1 Depression is among the highly preva-lent disorders in elderly people, and leads toincreased risk of mortality.2 The underlying

mechanisms for depressive symptoms in oldage remain unclear, but the inammatory hostresponse is repeatedly inferred in the patho-genesis of neuropsychiatric conditions.3 4 Inseveral population studies including olderparticipants, the relationship between inam-matory markers and depressive symptomshas been described.2 513 However, those studieshave generated conicting results. Somestudies observed positive associations betweeninammatory markers and depressive symp-toms,2 59 11 but others did not.12 13 Thisinconsistency might be due to differences instudy population, measurements of depres-sion and control for important confounders.Moreover, most of the reported data arebased on studies conducted in western coun-tries, with only a few exceptions targeting theAsian population.12 14

Thus, we investigated the independent asso-ciation between serum C reactive protein(CRP), a marker of systemic inammation,and depressive symptoms in an elderly Koreanpopulation.

METHODSStudy populationThe present study used data from the KoreanSocial Life, Health and Aging Project

Strengths and limitations of this study

We controlled sociodemographic, lifestyle, phys-ical and psychological factors to evaluate theindependent association between serum CRPlevels and depressive symptoms.

This study is the first to demonstrate that serumCRP levels are associated with depressive symp-toms in elderly Koreans.

Limitations include a single measurement ofserum CRP and use of self-report questionnairesfor depressive symptoms.

Song BM, et al. BMJ Open 2015;5:e006429. doi:10.1136/bmjopen-2014-006429 1

Open Access Research

group.bmj.com on June 28, 2015 - Published by http://bmjopen.bmj.com/Downloaded from

(KSHAP), which is a population-based longitudinal studyof health determinants among elderly Koreans. Thetarget populations of the KSHAP are people aged60 years or older and their spouses living within onetownship (myeon) of Ganghwa-gun, Incheon, Korea.This township is a typical rural Korean village wherefarming is the main industry. A complete enumerationsurvey was conducted in the rst wave of the KSHAP on94.7% (n=814) of the target population (n=860) betweenDecember 2011 and July 2012. The KSHAP-HealthExamination (KSHAP-HE) cohort consists of 698 peoplewho completed additional health examinations at apublic health centre or at their home. The sampling anddata collection procedures have been described in detailelsewhere.15 In this study, 129 participants were excludedfor missing key variables (n=85); history of stroke, anginapectoris or myocardial infarction (n=32); or CRP levelsabove 20 mg/L (n=12). Ultimately, a total of 569 partici-pants (224 men and 345 women) were included for cross-sectional analyses. All participants provided writteninformed consent forms.

Questionnaire dataAll participants were interviewed by trained personnelusing standardised questionnaires according to the pre-developed protocol. In addition, the designated eld dir-ector double-checked whether answers were missing orinappropriate and continuously monitored the entireinterview process. Sociodemographic variables includedage, education (uneducated, elementary school, middleschool or high school+) and marital status (with orwithout spouse). Health behaviours included smokingstatus, alcohol intake and sleep duration. Smoking statuswas classied into two groups: current smokers orcurrent non-smokers (past smokers or never smokers).Alcohol intake was categorised as follows: regularalcohol drinkers or others (participants who drinkunder once a week or never alcohol drinkers). Sleepduration was collected in hours per day (h/day) as theaverage during the past year. The presence of comorbid-ities were assessed by self-report and included diseases asfollows: hypertension, diabetes mellitus, metabolic syn-drome, dyslipidaemia, osteoporosis, cancer, stroke, myo-cardial infarction, angina pectoris, arthritis, pulmonarytuberculosis, asthma, cataract, glaucoma, hepatitis B,depression, urinary incontinence and benign prostatichyperplasia. Depressive symptoms were measured usingthe Korean version of the Center for EpidemiologicStudies Depression (CES-D) scale, a 20-item, self-reportscale designed to measure depressive symptoms experi-enced during the previous week.16 The scale, rangingfrom 0 to 60, has been shown to be a valid and reliableinstrument in older populations.17 The underlyingstructure of the CES-D consists of four dimensions ofdepression: seven items measure depressed affect, fouritems measure positive affect, two items measure inter-personal difculties and seven items measure somaticsymptoms.

Physical examinationAll participants wore lightweight hospital gowns for con-venient and reliable examinations. Standing height wasmeasured to the nearest 0.1 cm using a stadiometer andbody weight was measured to the nearest 0.1 kg on adigital scale. Body mass index (BMI) was calculated asbody weight in kilograms divided by standing height inmetres squared (kg/m2). Participants were seated for atleast 5 min before blood pressure measurement, andtwo measurements were taken within at least 5 min usingan automatic sphygmomanometer (Dinamap 1846 SX/P;GE Healthcare, Waukesha, Wisconsin, USA). If the twomeasurements differed by 10 mmHg for systolic bloodpressure (SBP) or diastolic blood pressure (DBP), a thirdmeasurement was taken after 5 min, and the last two mea-surements were averaged for analysis.

Laboratory testBlood samples were collected from the antecubital veinof participants after at least 8 h of fasting. Collectedblood samples were analysed at a central research labora-tory for serum CRP, total cholesterol, high-density lipo-protein cholesterol, triglycerides, glucose and so on. CRPlevels, a marker of acute systemic inammation, weredetermined in accordance with turbidimetric immuno-assay with an ADVIA1800 Auto Analyzer (Siemensmedical Sol., USA). The interassay coefcient of variationfor CRP was from 0.38% to 1.53% and results wereexpressed as mg/L.

Statistical analysisWe evaluated differences in clinical and demographiccharacteristics between men and women. CRP level,CES-D scale and several confounding factors were signi-cantly different between men and women. Therefore,we performed all analyses separately by sex. Continuousvariables were described as mean and SD (for normallydistributed variables) or as median and IQR (for skewedvariables), and tested by independent t test and Wilcoxonrank sum test, respectively. Categorical variables weredescribed as numbers with percentage and tested by a 2

test. General characteristics and concentrations ofselected biomarkers were also analysed according tothree categories of depression score. One-way analysis ofvariance was used for continuous variables and 2 test wasused for categorical variables. For this analysis, the trendtest was used. For the continuous variables, p values werecalculated by using a contrast to test for a linear trend.For the categorical variables, the Cochran-Armitage testand the Mantel-Haenszel test were used. The relation-ships between depression score and other variables wereevaluated using Pearsons correlation coefcients.Multiple linear regression analyses were used to assess theindependent association between serum CRP and depres-sion score in the unadjusted model and two adjustedmodels: (1) adjusting for age, BMI, SBP, number ofcomorbidities, smoking status and alcohol intake; (2)additional adjustment for marital status, sleep duration

2 Song BM, et al. BMJ Open 2015;5:e006429. doi:10.1136/bmjopen-2014-006429

Open Access

group.bmj.com on June 28, 2015 - Published by http://bmjopen.bmj.com/Downloaded from

and education. All statistical analyses were performedusing SAS V.9.2 (SAS Institute, Cary, North Carolina,USA), and statistical signicance was dened as two-sidedp value less than 0.05.

RESULTSClinical and demographic characteristics of the studyparticipants are presented in table 1. This study con-sisted of 224 men with a mean age of 72.8 years and 345women with a mean age of 71.6 years. The median CRPlevel was signicantly higher in men than in women (1.0vs 0.8 mg/L, p

Table 2 Characteristics of study participants by CES-D tertile

Variables Men (n=224) Women (n=345)03 (n=65) 410 (n=85) 11+ (n=74) p Trend 05 (n=123) 612 (n=107) 13+ (n=115) p Trend

Age, year 69.75.8 73.06.2 75.27.1

Health, Aging and Body Composition Study, high levelsof inammatory markers including CRP, interleukin 6(IL-6) and tumour necrosis factor (TNF-) were asso-ciated with depressive symptoms which are measuredwith the CES-D scale.5 Moreover, in older people withmetabolic syndrome, CRP levels were independentlyassociated with depressive symptoms after adjustingfor sex, age, education, smoking status, cognitive func-tion, metabolic syndrome, hypertension, diabetes and

medication.11 Lastly, a systemic review and meta-analysisof longitudinal studies which are conducted across allage groups found that raised inammatory markersincluding CRP and IL-6 have a small but signicant asso-ciation with the subsequent development of depressivesymptoms after adjustment for age and a wide range offactors associated with risk for depression.9 Contrary tothese studies, however, several studies did not nd inde-pendent associations between CRP levels and depressivesymptoms.10 19 20 In two elderly population studies, ele-vated CRP levels were associated with higher CES-Dscore, but not after adjustment for confoundingfactors.10 19 In the prospective Sydney Memory andAging Study, CRP levels were not associated with depres-sive symptoms, whereas IL-8 was associated with depres-sive symptoms at baseline and at 2-year follow-up.20

The relationship between inammation and depressionis potentially bidirectional, and the underlying mechan-isms are still poorly understood. First, inammation maylead to depression. Increased inammatory markers con-tribute to decreases in serotonin, an important neurotrans-mitter in the pathogenesis of depression.2123 Second,depression may lead to inammation. Psychological stressactivates the sympathetic nervous system and releases stresshormones. These hormones in turn initiate acute-phaseresponses triggering inammation.2426

In our study, CRP levels were signicantly associatedwith depressive symptoms in men but not in women. Thisnding is consistent with a few previous reports.2729 In anobese clinical population, depressive symptoms were sig-nicantly associated with CRP in men only, even afteradjusting for age, obesity, metabolic variables and medica-tions known to affect inammation. This nding suggeststhat there are biological differences between men andwomen that may modify the relationship between CRPlevels and depression.28 In the third National Health andNutrition Examination Survey (NHANES III), depressionand elevated CRP levels were more common in womenthan in men, but the association between CRP levels anddepression was signicant only in men.29 They hypothe-sised that CRP levels might vary by the hormonal environ-ment. In the Health 2000 Survey, depressive symptomswere independently associated with systemic inammationprocesses in men, but not in women.27 They suggest that

Table 3 Pearsons correlation coefficients between depression score and other variables

VariablesMen (n=224) Women (n=345)Correlation coefficient p Value Correlation coefficient p Value

Age 0.333

depression may have a different psychological and physio-logical meaning and background for men and women.Our study participants were the elderly population aged60 years or older. Therefore, our ndings that there was asignicant association between CRP levels and depressivesymptoms in men only are probably because the mainfactors inuencing on depression differ rather than hor-monal environments or biological features differ betweenmen and women.Our study has a few distinctions compared with previ-

ous studies. The study participants were recruited from acommunity of single ethnic background. Furthermore,we carefully controlled various covariates known to berelated to inammation and depression in the analyses.These enabled us to evaluate the independent associ-ation between the inammatory marker and depressivesymptoms. To the best of our knowledge, this study isthe rst to demonstrate that serum CRP levels are asso-ciated with depressive symptoms in elderly Koreans. Wealso observed the gender difference in the associationbetween CRP levels and depressive symptoms amonggeneral elderly Koreans.The study has some limitations. First, since this is a cross-

sectional study in which all information was collected atthe same point in time, no denitive conclusion about thedirection of the associations between CRP levels anddepressive symptoms could be reached. Second, we reliedon a single blood sample per individual. However, circulat-ing CRP levels obtained from one time point have beenshown to represent systemic low-grade inammation.30

Third, inammatory markers other than CRP were notmeasured in this study. Further studies including otherinammatory markers needed to conrm our results.Fourth, we investigated depressive symptoms using theCES-D scale, which has been widely used to assess depres-sion and its risk factors in population-based studies.17 16

However, the validity of CES-D is inferior to the psychiatricdiagnostic interview, the gold standard.31 Fifth, we did notuse a validated measure of comorbidity such as theCharlson Comorbidity index because comorbidity compo-nents of the Charlson Comorbidity index were notmatched with the comorbidities asked for in this study.However, many epidemiological studies have used anumber of selected chronic disorders to measure generalhealth conditions of elderly people.1 20 32 Sixth, we couldnot control the effects of medications such as hormone-replacement therapy, anti-inammatory medications andantidepressants. Alternatively, we additionally analysedafter excluding 11 people with a history of depression.A similar result was observed (online supplementarytable S1). Lastly, the study participants were recruited froma single rural community and the sample size was relativelysmall. Consequently, our ndings may not be generalisableto the other older population.In conclusion, our ndings suggest that high serum

CRP levels are independently associated with depressivesymptoms in community-dwelling elderly Korean men,even after adjusting for several physical and psychological

Table4

Associationbe

twee

nde

pression

scorean

dCRP

Variables

Men

(n=2

24)

Women

(n=3

45)

Unad

justed

Model1*

Model2

Unad

justed

Model1*

Model2

pValue

pValue

pValue

pValue

pvalue

pValue

CRP

0.42

00.01

00.36

50.01

70.33

60.02

50.25

90.25

00.25

10.25

40.25

30.25

6Age

0.27

7