- Email: [email protected]
Five factor personality traits and inflammatory biomarkers in the English longitudinal study of aging
Personality and Individual Differences 111 (2017) 205–210
Contents lists available at ScienceDirect
Personality and Individual Differences journal homepage: www.elsevier.com/locate/paid
Five factor personality traits and inflammatory biomarkers in the English longitudinal study of aging Mark S. Allen a,⁎, Sylvain Laborde b,c a b c
School of Psychology, Faculty of Social Sciences, University of Wollongong, Northfields Avenue, Wollongong 2522, NSW, Australia EA 4260, University of Caen, France Department of Performance Psychology, German Sport University Cologne, Germany
a r t i c l e
i n f o
Article history: Received 13 November 2016 Received in revised form 7 February 2017 Accepted 10 February 2017 Available online xxxx Keywords: C–reactive protein Fibrinogen Gerontology Immune function Neuroticism White blood cell
a b s t r a c t Inflammatory processes are putative mechanisms underlying chronic disease. In this study we explore linear and non-linear associations between personality and change in inflammatory markers (C–reactive protein [CRP], fibrinogen, and white blood cell count [WBC]). In total, 5294 older adults (Mage = 64.51 ± 8.34 years) provided blood samples in 2008 with 3751 providing follow-up samples four year later. Midway between the two assessments, participants completed a measure of personality. After controlling for participant demographics (e.g., age, gender) and health-related lifestyle factors (e.g., exercise, cigarette smoking), we found that higher levels of agreeableness and lower levels of conscientiousness were associated with higher CRP levels, and (for conscientiousness) an increase in CRP over time. Age moderation effects indicated that agreeableness and conscientiousness were related to WBC among older participants in the sample (over ~ 70 years of age). These findings provide evidence that agreeableness and conscientiousness traits are important for inflammatory biomarkers in older adulthood. © 2017 Elsevier Ltd. All rights reserved.
1. Introduction Inflammation is one of the first responses of the immune system to harmful stimuli. Inflammatory abnormalities underlie biological processes that contribute to cardiovascular disease and other age-related diseases (Emerging Risk Factors Collaboration, 2010, 2012; Grivennikov, Greten, & Karin, 2010; Libby, 2002). Inflammatory markers, including C–reactive protein (CRP) and fibrinogen, reflect demographic, anthropometric, behavioural, and psychosocial factors (McDade, Hawkley, & Cacioppo, 2006) including trait personality (Luchetti, Barkley, Stephan, Terracciano, & Sutin, 2014). The five factor model (McCrae & John, 1992) considers that personality is best captured through the assessment of five broad trait dimensions: neuroticism, extraversion, openness, agreeableness and conscientiousness. Of these, there is emerging evidence that high conscientiousness might have a protective role in inflammatory processes independent of behavioural and anthropometric factors such as cigarette smoking and body mass (Luchetti et al., 2014). However, the potential involvement of other trait dimensions is less clear. The current study sought to explore associations between the big five trait dimensions and three markers of
⁎ Corresponding author at: School of Psychology, University of Wollongong, Northfields Avenue, NSW 2522, Australia. E-mail address: [email protected] (M.S. Allen).
http://dx.doi.org/10.1016/j.paid.2017.02.028 0191-8869/© 2017 Elsevier Ltd. All rights reserved.
chronic inflammation (CRP, fibrinogen, and white blood cell count [WBC]) in a large sample of older adults from England. Personality is thought to relate to inflammatory biomarkers through behavioural and psychophysiological mechanisms (Segerstrom, 2000). For example, personality is a strong predictor of stress appraisal and coping (Carver & Connor-Smith, 2010), and psychological stress can activate acute phase reactants associated with inflammation (Black, 2002, 2003). Inflammation (and activation of the immune system) can be detected by an increase in several biomarkers including WBC, fibrinogen, CRP, and interleukin-6 (IL-6). The most consistent finding across studies is that lower levels of conscientiousness relate to higher levels of IL-6 and CRP (Chapman et al., 2011; Mõttus, Luciano, Starr, Pollard, & Deary, 2013; Stephan, Sutin, Luchetti, & Terracciano, 2016; Sutin et al., 2010; Turiano, Mroczek, Moynihan, & Chapman, 2013), as well as increases in these markers over time (Chapman et al., 2011; Turiano et al., 2013), and this was supported in a recent meta-analysis of seven (CRP) and six (IL-6) independent samples (Luchetti et al., 2014). In addition to conscientiousness, the meta-analysis also found that lower levels of openness related to higher CRP levels, and that effects remained significant after controlling for some behavioural and anthropometric factors (cigarette smoking and body mass). There were no significant effects for neuroticism, extraversion or agreeableness (but notable heterogeneity across study samples), and sample age was not found to moderate associations between personality and CRP. The authors pointed towards the small number of samples limiting statistical
206
M.S. Allen, S. Laborde Personality and Individual Differences 111 (2017) 205–210
power to detect age moderation effects, and recommended researchers continue to search for important moderators that might explain heterogeneity across studies (Luchetti et al., 2014). Less research has explored fibrinogen and WBC as correlates of trait personality. In a sample of Israeli adults, after controlling for body fat and health behaviour, lower levels of openness and higher levels of neuroticism were found to be associated with higher fibrinogen levels, and (for neuroticism) an increase in fibrinogen over four years (Armon, Melamed, Shirom, Berliner, & Shapira, 2013). However, a further study of Scottish older adults, that included two measures of personality, found that agreeableness had a negative association with fibrinogen for one personality measure, but overall, fibrinogen (and change in fibrinogen over time) was unrelated to personality (Mõttus et al., 2013). Similar findings were observed in a study exploring two of the big five traits, in which neuroticism and extraversion were unrelated to fibrinogen after controlling for body mass and health behaviour (Millar et al., 2013). For WBC, a large sample study of Sardinian adults found that broad trait dimensions were not consistently associated with WBC and were unrelated to change in WBC over three years (Sutin et al., 2012). However, the study did find that higher levels of openness and lower levels of conscientiousness were related to a WBC range associated with increased risk of mortality (≥7.6 × 109 cells/L). Taken together, there is good (but not robust) evidence that conscientiousness is important for CRP and IL-6 (but associations for other dimensions remain unclear), mixed findings for an association between personality and fibrinogen, and a shortage of research into personality and WBC meaning generalizable conclusions about important associations cannot be made at this stage. One explanation for some of the inconsistent findings might be that personality becomes more or less important for inflammatory processes among particular subgroups. Indeed, psychological stress is often the mechanism considered most responsible for connecting personality to inflammation (Armon et al., 2013) and psychological stress becomes more important for inflammatory responses as people become older (Segerstrom & Miller, 2004). This would suggest that personality might have an important role in chronic inflammation but only among older adults. The current study is interested in the association between personality and inflammatory markers independent of health behaviour, and therefore we control for health-related factors such as exercise, cigarette smoking and body mass. We hypothesised that lower levels of conscientiousness would be associated with higher inflammation, and increases in inflammation over time, as measured via CRP, fibrinogen and WBC. We also explore age as a potential moderator of associations between personality and inflammatory markers, and hypothesised stronger associations between personality and inflammatory markers among older adults. 2. Method 2.1. Sample The English Longitudinal Study of Aging (ELSA) is a biannual survey that collects information on health, social, and economic circumstances of the English population aged 50 and older. Postcode sectors are selected from the national Postcode Address File, and addresses are selected randomly from each sector. At wave 1 there was a 77.5% response rate. At waves 3, 4, and 6 the sample was replenished with new participants to maintain the size and representativeness of the panel. The data is broadly considered to be representative of older adults in England. Detailed information on sampling and attrition across waves is available elsewhere (Steptoe, Breeze, Banks, & Nazroo, 2013). Ethical approval was granted by the National Research and Ethics Committee UK. We refer to data collected at wave 4 as Time 1 (2008) and data collected at wave 6 as Time 2 (2012). In total, 10,749 adults aged 50 and over were sampled at Time 1 with 6384 (59.4%) providing blood samples. Compared to those that provided blood samples, non-sampled participants were older, t(10,747) = 4.65, p b 0.001, d = 0.09, less
physically active, t(10,737) = 15.43, p b 0.001, d = 0.30, and had lower levels of extraversion, t(7884) = 3.42, p = 0.001, d = 0.08, and conscientiousness, t(7866) = 3.75, p b 0.001, d = 0.08. Personality data were available for 5294 of the participants that provided blood samples (82.9%). Compared to participants with personality data, those without personality data were older, t(6382) = 17.39, p b 0.001, d = 0.44, less physically active, t(6381) = 12.81, p b 0.001, d = 0.32, had a lower body mass, t(6382) = 5.84, p b 0.001, d = 0.15, higher blood CRP levels, t(6382) = 5.88, p b 0.001, d = 0.15, and a higher WBC count, t(6233) = 6.65, p b 0.001, d = 0.17. At Time 2, 10,372 participants were included, with 6039 (58.2%) providing blood samples. Compared to those that provided blood samples, non-sampled participants were older, t(10,370) = 4.90, p b 0.001, d = 0.10, had a higher body mass, t(7742) = 6.43, p b 0.001, d = 0.15, were less physically active, t(10,365) = 13.46, p b 0.001, d = 0.26, and had lower levels of extraversion, t(7662) = 4.36, p b 0.001, d = 0.10, and conscientiousness, t(7652) = 4.19, p b 0.001, d = 0.10. There were personality data available for 4911 participants that provided blood samples (81.3%). Compared to participants with personality data, those without personality data were younger, t(6037) = 21.27, p b 0.001, d = 0.55, less physically active, t(6036) = 4.94, p b 0.001, d = 0.13, and had a higher WBC count, t(6037) = 3.52, p b 0.001, d = 0.09. In total, there were 5294 participants available at Time 1 (2908 women, 2386 men; Mage = 64.51 ± 8.34 years), 4911 participants at Time 2 (2702 women, 2209 men; Mage = 67.90 ± 8.20 years), and 3751 participants sampled at both Time 1 and Time 2 (2065 women, 1686 men; Mage (Time 1) = 64.14 ± 7.94 years). 2.2. Measures 2.2.1. Personality Data on personality were collected midway between Time 1 and Time 2 (at wave 5) using a modified version of the Midlife Development Inventory (MDI; Lachman & Weaver, 1997). ELSA included an additional item (“thorough”) to the original 25 item measure to increase internal consistency of the conscientiousness subscale. The questionnaire uses 26 adjectives to assess five personality traits: neuroticism (e.g., “nervous”), extraversion (e.g., “outgoing”), openness (e.g., “creative”), agreeableness (e.g., “helpful”), and conscientiousness (e.g., “organised”). Participants indicate how well each adjective described them on a fourpoint scale: 1 (a lot), 2 (some), 3 (a little) or 4 (not at all). Scales were reversed so that higher scores represent higher levels of neuroticism (α = 0.68), extraversion (α = 0.76), openness (α = 0.79), agreeableness (α = 0.80), and conscientiousness (α = 0.68). The MDI has demonstrated evidence of criterion validity in adult samples (e.g., Joshanloo, 2017). 2.2.2. Clinical assessment A trained nurse collected anthropometric data (body mass and height) and blood samples. Body mass was measured using Seca 877 scales and height was measured using a portable Stadiometer. Participants did not wear shoes for anthropometric measures but wore light clothing. Body mass index (BMI) was calculated as mass (kg)/height squared (m2). Blood samples were taken from the left or right arm using a Vacutainer or Butterfly needle, and a maximum of two attempts were made to extract blood. Measurement of CRP was carried out using the N Latex CRP mono Immunoassay on the Behring Nephelometer II Analyzer. Fibrinogen analysis was carried out using the Organon Teknika MDA 180 analyser, using a modification of the Clauss thrombin clotting method. WBC was measured using the Sysmex XE2100 analyser that uses electrical impedance technology (Coulter principle) to count the white cells combined with hydrodynamic focusing and flow cytometry to differentiate the blood cells. All analyses were carried out according to Standard Operating Procedures by state registered medical laboratory scientific officers.
M.S. Allen, S. Laborde Personality and Individual Differences 111 (2017) 205–210
207
Table 1 Bivariate correlations between biomarkers and other study variables. Time 1
Age Gender Ethnicity Past smoker Current smoker Exercise BMI Neuroticism Extraversion Openness Agreeableness Conscientiousness
Time 2
CRP
Fibrinogen
WBC
CRP
Fibrinogen
WBC
0.08⁎⁎ 0.04⁎ 0.01 0.08⁎⁎ 0.11⁎⁎
0.14⁎⁎ 0.08⁎⁎ 0.01 0.08⁎⁎ 0.16⁎⁎
0.01 −0.05⁎⁎ 0.01 0.15⁎⁎ 0.30⁎⁎ −0.11⁎⁎ 0.13⁎⁎
0.09⁎⁎ 0.06⁎⁎ −0.01 0.07⁎⁎ 0.09⁎⁎ −0.21⁎⁎ 0.31⁎⁎
0.12⁎⁎ 0.07⁎⁎ −0.01 0.06⁎⁎ 0.12⁎⁎ −0.15⁎⁎ 0.18⁎⁎
0.06⁎⁎ −0.06⁎⁎ −0.01 0.12⁎⁎ 0.26⁎⁎ −0.13⁎⁎ 0.12⁎⁎
0.02 −0.06⁎⁎ −0.01 0.00 −0.06⁎⁎
−0.00 −0.04⁎ −0.04⁎ 0.05⁎ −0.09⁎⁎
−0.02 −0.04⁎ −0.04⁎ 0.03 −0.07⁎⁎
0.01 −0.06⁎⁎ −0.03 −0.01 −0.07⁎⁎
−0.20⁎⁎ 0.37⁎⁎ 0.02 −0.05⁎⁎ −0.01 0.06⁎⁎ −0.08⁎⁎
−0.16⁎⁎ 0.20⁎⁎ 0.01 −0.07⁎⁎ −0.05⁎ 0.02 −0.08⁎⁎
Note: Gender was coded as 1 (men) or 2 (women), ethnicity was coded as 1 (White) or 2 (non-White), past and current smoker were coded as 1 (no) or 2 (yes). ⁎ p b 0.01, ⁎⁎ p b 0.001.
To explore associations between personality and inflammatory markers we ran nine regression models. CRP, fibrinogen and WBC were set as dependent variables and we explored these markers at Time 1, Time 2 and the change in these markers over time. In all models we controlled for age, gender, ethnicity, physical activity, past smoking, current smoking, and BMI. For change scores, we entered the Time 1 score as an additional covariate with the Time 2 score set as the dependent variable. Because CRP showed extreme positive skew (see Supplementary File), raw data were transformed into ordered categories according to relative risk (Ridker & Cook, 2004). We further tested whether associations between personality and inflammatory markers varied as a function of age and gender. Interaction terms were computed from standardised data, and significant effects were followed up using simple slope analyses (Hayes, 2013). Consistent with previous research (Sutin et al., 2012) we ran additional logistic regression models for WBC at Time 1 and Time 2 using a threshold of ≥ 7.6 × 109 cells/L that has been found to be associated with an increased risk of cardiovascular mortality (Brown, Giles, & Croft, 2001). Prior to analyses, missing data were imputed for 809 empty cells at Time 1 (1.3% of the data set) and 659 empty cells at Time 2 (1.1% of the data set) using multiple imputation (Rubin, 1987). For the linear regression models, there was no evidence of multicollinearity (VIF values b 3.00) and no multivariate outliers (Cook's distance values b 0.05). For the logistic regression, the Hosmer–Lemeshow test was non-significant at all steps of the model (p N 0.001). For all statistical tests, significance was set at p b 0.01.
Time 1, Time 2 and for CRP change. There were no moderation effects by gender, but some associations were moderated by age. At Time 1, there was a significant age moderation of agreeableness on WBC (βinteraction = 0.05, p = 0.004). Simple slope analyses demonstrated that agreeableness was unrelated to WBC among younger participants (tmean age 56.2 = − 0.47, p = 0.635) and the sample average (tmean age 64.5 = 2.15, p = 0.032), but had a positive association among older participants (tmean age 72.8 = 3.53, p b 0.001). This age moderation effect also emerged at Time 2 (βinteraction = 0.05, p = 0.005). Computation of Johnson-Neyman significance regions showed that agreeableness became a significant predictor of WBC for participants aged over 66.7 years at Time 1, and over 74.9 years at Time 2. At Time 2, there was also a significant age moderation of conscientiousness on WBC (βinteraction = −0.05, p = 0.006), with conscientiousness unrelated to WBC among younger participants (tmean age 59.7 = 1.30, p = 0.191) and the sample average (tmean age 67.9 = − 0.70, p = 0.483), but negatively related among older participants (tmean age 76.1 = − 2.33, p = 0.019). The effect for conscientiousness on WBC was significant for participants aged over 75.0 years. This age moderation effect also emerged for conscientiousness on WBC change (βinteraction = −0.04, p = 0.004), with conscientiousness becoming a significant predictor of change in WBC for participants aged over 72.2 years (at Time 1). No age or gender moderation effects emerged for CRP or fibrinogen. Findings from the logistic regression models are reported in the Supplementary File. No significant main effects emerged and there were no gender moderations. There was a significant age moderation of agreeableness on WBC at Time 1 (b = 0.14, s.e. = 0.04, p = 0.001) and Time 2 (b = 0.15, s.e. = 0.04, p b 0.001). Similar to the linear regression model, agreeableness was unrelated to WBC among younger participants (zmean age 56.2 = −1.67, p = 0.096; zmean age 59.7 = −1.87, p = 0.060) and the sample average (zmean age 64.5 = 0.95, p = 0.343; zmean age 67.9 = 0.78, p = 0.437), but had a positive association among older participants (zmean age 72.8 = 2.96, p = 0.003; zmean age 76.1 = 3.02, p = 0.002). Computation of significance regions showed that agreeableness became a significant predictor of WBC categories for participants aged over 70.1 years at Time 1, and over 73.4 years at Time 2.1
3. Results
4. Discussion
Table 1 provides correlations between inflammatory markers and other measured variables, and findings from the linear regression models are reported in Table 2. The data show that extraversion had negative associations with inflammatory markers at Time 1, but these associations did not re-emerge at Time 2. A positive association emerged between agreeableness and CRP at Time 1 and Time 2, and a negative association emerged between conscientiousness and CRP at
This study explored associations between personality and inflammatory markers in a large sample of English adults aged 50 and over. After controlling for behavioural and anthropometric factors (e.g., cigarette
2.2.3. Control variables Participants reported their date of birth, gender, ethnicity, whether they currently smoked cigarettes, and whether they had ever smoked cigarettes. Participants also reported whether they “take part in sports or activities that are moderately energetic” with response categories of 1 (hardly ever, or never), 2 (one to three times a month), 3 (once a week) or 4 (more than once a week). 2.3. Data analysis
1 Similar (but non-significant) effects emerged for conscientiousness at Time 1 (b = −0.10, s.e. = 0.04, p = 0.014) and Time 2 (b = −0.10, s.e. = 0.04, p = 0.013) – indicative of stronger associations among older participants.
208
M.S. Allen, S. Laborde Personality and Individual Differences 111 (2017) 205–210
Table 2 Regression models exploring linear associations between personality and inflammatory biomarkers. C–reactive protein
Fibrinogen
White blood cell count
Time 1
Time 2
Change
Time 1
Time 2
Change
Time 1
Time 2
Change
0.03 0.09⁎⁎ 0.02 0.02 0.13⁎⁎
0.04⁎ 0.09⁎⁎ −0.01 0.03 0.10⁎⁎
0.02 0.04⁎ −0.01 0.02 0.04 −0.04⁎
0.07⁎⁎ 0.15⁎⁎ 0.01 0.02 0.18⁎⁎
0.03 0.05⁎⁎ −0.02 −0.00 0.05⁎⁎
−0.07⁎⁎ 0.08⁎⁎ 0.00 0.04 0.26⁎⁎ −0.07⁎⁎ 0.13⁎⁎
−0.04⁎ 0.07⁎⁎ −0.01 0.00 0.07⁎⁎
−0.03 0.07⁎⁎
−0.06⁎⁎ 0.04⁎ 0.01 0.05⁎ 0.30⁎⁎ −0.05⁎ 0.15⁎⁎
−0.03 0.04⁎
Step 1 Gender Age Ethnicity Past smoker Current smoker Exercise BMI
−0.10⁎⁎ 0.35⁎⁎
−0.11⁎⁎ 0.30⁎⁎
0.03
−0.07⁎⁎ 0.20⁎⁎
0.06⁎⁎ 0.12⁎⁎ −0.01 0.02 0.14⁎⁎ −0.07⁎⁎ 0.18⁎⁎
Step 2 Neuroticism Extraversion Openness Agreeableness Conscientiousness
0.01 −0.06⁎⁎ 0.04 0.07⁎⁎ −0.04⁎
−0.02 −0.01 −0.01 0.05⁎ −0.04⁎
−0.01 −0.01 0.00 0.01 −0.05⁎
−0.00 −0.06⁎⁎ 0.02 0.04 −0.02
−0.04 −0.03 0.00 0.04 −0.04
−0.01 −0.02 0.02 0.01 −0.04
−0.00 −0.06⁎⁎ 0.02 0.04 −0.01
0.00 −0.02 −0.01 0.02 −0.01
0.01 −0.01 −0.01 −0.00 −0.00
Step 3 Age × neuroticism Age × extraversion Age × openness Age × agreeableness Age × conscientiousness R2
0.01 0.01 0.02 0.01 0.00 0.18
0.01 0.03 −0.03 0.01 −0.01 0.14
0.00 0.02 −0.03 0.01 −0.01 0.35
0.00 0.01 0.02 −0.01 0.01 0.11
0.01 0.01 −0.02 0.01 −0.00 0.08
0.00 0.03 −0.04 0.00 0.00 0.32
−0.01 −0.01 0.01 0.05⁎ −0.03 0.13
−0.00 0.02 −0.02 0.05⁎ −0.04⁎ 0.11
0.00 0.01 −0.02 0.02 −0.04⁎ 0.40
Note: Standardised regression coefficients reported. ⁎ p b 0.01, ⁎⁎ p b 0.001.
smoking, body mass), we found that higher levels of agreeableness and lower levels of conscientiousness were associated with higher CRP levels, and (for conscientiousness) an increase in CRP levels over four years. Extraversion had inconsistent associations with inflammatory markers (negative associations at Time 1 but not Time 2), and neuroticism and openness were unrelated to inflammatory markers. Age moderation effects showed that agreeableness and conscientiousness were important for WBC among older participants in the sample. These findings add to the growing body of literature that is beginning to uncover important connections between personality and markers of chronic inflammation. The finding that lower levels of conscientiousness were related to higher levels of CRP and increases in CRP over time is consistent with previous research (Luchetti et al., 2014; Mõttus et al., 2013; Stephan et al., 2016; Sutin et al., 2010). Conscientious individuals are characterised by organisation, discipline and self-control (McCrae & John, 1992), and tend to respond to daily life stress with more adaptive coping responses (Carver & Connor-Smith, 2010). We can speculate that conscientiousness might offer protection against elevated CRP levels (indicative of chronic inflammation) because conscientious individuals have more resources available to cope with exposure to stress. The size of the effect was small (see Gignac & Szodorai, 2016) and this is consistent with findings in other large sample studies (Luchetti et al., 2014). Small effects are common in research on personality and health outcomes (Bogg & Roberts, 2004) but can, nonetheless, have substantial practical consequences at the population level. We also found that higher levels of agreeableness related to higher CRP levels. This association was not hypothesised and is inconsistent with findings from previous studies that show agreeableness is generally unrelated to CRP and other inflammatory biomarkers (Luchetti et al., 2014; Mõttus et al., 2013; Stephan et al., 2016; Sutin et al., 2010). Of the seven samples included in the meta-analysis by Luchetti et al. (2014), only one (the Midlife in the United States [MIDUS] sample) showed a positive association between agreeableness and CRP. The MIDUS uses the same personality measure as that used in the current study – the MIDI (Lachman & Weaver, 1997) – and the extreme heterogeneity across study samples might be a function of differences between personality measures (see Luchetti et al., 2014; Mõttus et al., 2013). To speculate on the positive association found here, individuals scoring high on
agreeableness tend to be more cooperative, considerate, kind and sympathetic (McCrae & John, 1992), but are eager to avoid unsettling others and, despite being more support giving, are less support seeking (often using disengagement coping strategies such as withdrawal; Carver & ConnorSmith, 2010). This might, under certain circumstances, place agreeable individuals at higher risk of prolonged stress, contributing to elevated CRP levels. Nevertheless, given the mixed findings observed for CRP and agreeableness in previous research, the positive association found here should be interpreted with caution. Findings showed that agreeableness and conscientiousness were unrelated to WBC and change in WBC over time, and this is consistent with findings from a large sample of Sardinian adults (Sutin et al., 2012). We also found that WBC was unrelated to personality dimensions in our non-linear analyses of WBC and this is inconsistent with findings from Sutin et al. who found that openness and conscientiousness differed between high-risk and low-risk categories of WBC. Our findings did, nonetheless, provide evidence for an association between WBC and the traits of agreeableness and conscientiousness among the oldest participants in the sample. This result was hypothesised based on findings that age increases vulnerability to stress-related decreases in functional immune measures (Segerstrom & Miller, 2004), and suggests that high conscientiousness and low agreeableness might offer a protective role against elevated WBC among the elderly. It is unclear whether previous research tested for age moderation effects, and further research is needed to verify this finding given the problems that come with testing interaction effects in multiple regression field tests (see McClelland & Judd, 1993). The differences in findings (for the non-linear analyses) might reflect differences in lifestyle factors between the population sampled by Sutin et al. (adults from the Mediterranean island of Sardinia) and those sampled in the current study. That neuroticism, extraversion and openness were unimportant for fibrinogen and CRP is supportive of most previous research (Luchetti et al., 2014; Millar et al., 2013; Mõttus et al., 2013). However, findings of this study did show that higher extraversion related to higher inflammation (as measured by all three markers) at Time 1, but the association did not re-emerge at Time 2. Similar to previous studies (e.g., Sutin et al., 2012), we consider inconsistent findings as not supportive of an association. However, it is worth noting that extraversion was found to have a
M.S. Allen, S. Laborde Personality and Individual Differences 111 (2017) 205–210
negative association with CRP in a demographically similar sample of older adults from the United States (Stephan et al., 2016), and therefore a potential connection between extraversion and inflammatory biomarkers might warrant further investigation. Some research has also found that neuroticism relates to fibrinogen (Armon et al., 2013) and CRP (Sutin et al., 2010) and, given the importance of neuroticism in stress appraisal and coping (Carver & Connor-Smith, 2010), this association might also warrant further scrutiny. Strengths of this study include the large sample, measurement of inflammatory biomarkers over two time points, and control of anthropometric and health-related factors. However, there are some important limitations that require consideration when interpreting study findings. The most obvious limitation is that the measurement of personality was not concurrent with sampling of blood. We have interpreted the measure of personality (assessed midway between blood sample measures) as providing an approximation of older adults' personality at both time points – two years earlier and two years later – but it is possible (and likely) that personality changed, at least to some degree, over the four years. Personality is more stable in later adulthood than early adulthood, but traits (and agreeableness in particular) have been found to change meaningfully in later life (see Roberts, Walton, & Viechtbauer, 2006). This means that our assessment of personality might not entirely reflect personality at Time 1 or Time 2, and this might have contributed additional measurement error attenuating some associations. A second limitation is that the final sample was not entirely representative of the English older adult population. Attrition analyses showed that excluded participants (with missing blood sample or personality data) were older, had higher CRP and WBC levels, and lower levels of extraversion and conscientiousness. This might also have attenuated some associations, and findings should be considered a reflection of sampled participants that might not necessarily be representative of all older adults living in England. A third limitation is that the study does not address the processes connecting personality to inflammatory biomarkers. Future research might look to explore stress and other lifestyle factors as potential mediators of the associations between personality and chronic inflammation. To conclude, a number of important findings have emerged from this study. First, the study reinforces findings from previous research demonstrating a negative association between conscientiousness and change in CRP. The study also provides evidence that higher levels of agreeableness relate to higher CRP levels, and that higher levels of agreeableness and lower levels of conscientiousness relate to a higher WBC among elderly individuals. These findings all controlled for demographic, anthropometric and health-related factors. Effect sizes for significant associations were small, and below the RMPE (recommended minimum effect size representing a ‘practically’ significant effect) for social science research (Ferguson, 2009), but small effects can have meaningful practical consequences at the population level. More research is required before practical applications of this research can be made in confidence, but should findings be replicated in subsequent independent research, they might have implications in terms of early identification of individuals at risk of chronic inflammation. We recommend prospective studies explore personality and inflammation in alternative cultures, and aim to identify potential factors that might explain some of the inconsistent findings observed across study samples. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.paid.2017.02.028.
References Armon, G., Melamed, S., Shirom, A., Berliner, S., & Shapira, I. (2013). The associations of the Five Factor Model of personality with inflammatory biomarkers: A four-year prospective study. Personality and Individual Differences, 54, 750–755http://dx.doi.org/10.1016/ j.paid.2012.11.035
209
Black, P. H. (2002). Stress and the inflammatory response: A review of neurogenic inflammation. Brain, Behavior, and Immunity, 16, 622–653http://dx.doi.org/10.1016/S08891591(02)00021-1 Black, P. H. (2003). The inflammatory response is an integral part of the stress response: Implications for atherosclerosis, insulin resistance, type II diabetes and metabolic syndrome X. Brain, Behavior, and Immunity, 17, 350–364http://dx.doi.org/10.1016/ S0889-1591(03)00048-5 Bogg, T., & Roberts, B. W. (2004). Conscientiousness and health-related behaviors: A meta-analysis of the leading behavioral contributors to mortality. Psychological Bulletin, 130, 887–919http://psycnet.apa.org/doi/10.1037/0033-2909.130.6.887 Brown, D. W., Giles, W. H., & Croft, J. B. (2001). White blood cell count: An independent predictor of coronary heart disease mortality among a national cohort. Journal of Clinical Epidemiology, 54, 316–322http://dx.doi.org/10.1016/S0895-4356(00)00296-1 Carver, C. S., & Connor-Smith, J. (2010). Personality and coping. Annual Review of Psychology, 61, 679–704. http://dx.doi.org/10.1146/annurev.psych.093008. 100352. Chapman, B. P., van Wijngaarden, E., Seplaki, C. L., Talbot, N., Duberstein, P., & Moynihan, J. (2011). Openness and conscientiousness predict 34-week patterns of interleukin-6 in older persons. Brain, Behavior, and Immunity, 25, 667–673http://dx.doi.org/10.1016/j. bbi.2011.01.003 Emerging Risk Factors Collaboration. (2010). C-reactive protein concentration and risk of coronary heart disease, stroke, and mortality: An individual participant meta-analysis. The Lancet, 375, 132–140http://dx.doi.org/10.1016/S01406736(09)61717-7 Emerging Risk Factors Collaboration. (2012). C-reactive protein, fibrinogen, and cardiovascular disease prediction. New England Journal of Medicine, 367, 1310–1320. http://dx.doi.org/10.1056/NEJMoa1107477. Ferguson, C. J. (2009). An effect size primer: A guide for clinicians and researchers. Professional Psychology, 40, 532–538http://psycnet.apa.org/doi/10.1037/ a0015808 Gignac, G. E., & Szodorai, E. T. (2016). Effect size guidelines for individual differences researchers. Personality and Individual Differences, 102, 74–78http://dx.doi.org/10. 1016/j.paid.2016.06.069 Grivennikov, S. I., Greten, F. R., & Karin, M. (2010). Immunity, inflammation, and cancer. Cell, 140, 883–899http://dx.doi.org/10.1016/j.cell.2010.01.025 Hayes, A. F. (2013). Introduction to mediation, moderation, and conditional process analysis. New York, NY: Guilford. Joshanloo, M. (2017). Factor structure and criterion validity of original and short versions of the Negative and Positive Affect Scale (NAPAS). Personality and Individual Differences, 105, 233–237http://dx.doi.org/10.1016/j.paid.2016.09.060 Lachman, M. E., & Weaver, S. L. (1997). The Midlife Development Inventory (MIDI) personality scales. Waltham, MA: Brandeis University. Libby, P. (2002). Inflammation in atherosclerosis. Nature, 420, 868–874. http://dx.doi.org/ 10.1038/nature01323. Luchetti, M., Barkley, J. M., Stephan, Y., Terracciano, A., & Sutin, A. R. (2014). Five-factor model personality traits and inflammatory markers: New data and a meta-analysis. Psychoneuroendocrinology, 50, 181–193http://dx.doi.org/10.1016/j.psyneuen.2014. 08.014 McClelland, G. H., & Judd, C. M. (1993). Statistical difficulties of detecting interactions and moderator effects. Psychological Bulletin, 114, 376–390http://psycnet.apa.org/doi/10. 1037/0033-2909.114.2.376 McCrae, R. R., & John, O. P. (1992). An introduction to the five-factor model and its applications. Journal of Personality, 60, 175–215. http://dx.doi.org/10.1111/j.1467-6494. 1992.tb00970.x. McDade, T. W., Hawkley, L. C., & Cacioppo, J. T. (2006). Psychosocial and behavioral predictors of inflammation in middle-aged and older adults: The Chicago health, aging, and social relations study. Psychosomatic Medicine, 68, 376–381. http://dx.doi.org/ 10.1097/01.psy.0000221371.43607.64. Millar, K., Lloyd, S. M., McLean, J. S., Batty, G. D., Burns, H., Cavanagh, J., ... Mõttus, R. (2013). Personality, socio-economic status and inflammation: Cross-sectional, population-based study. PloS One, 8, e58256http://dx.doi.org/10.1371/journal.pone. 0058256 Mõttus, R., Luciano, M., Starr, J. M., Pollard, M. C., & Deary, I. J. (2013). Personality traits and inflammation in men and women in their early 70s: The Lothian Birth Cohort 1936 study of healthy aging. Psychosomatic Medicine, 75, 11–19. http://dx.doi.org/ 10.1097/PSY.0b013e31827576cc. Ridker, P. M., & Cook, N. (2004). Clinical usefulness of very high and very low levels of Creactive protein across the full range of Framingham risk scores. Circulation, 109, 1955–1959. http://dx.doi.org/10.1161/01.CIR.0000125690.80303.A8. Roberts, B. W., Walton, K. E., & Viechtbauer, W. (2006). Patterns of mean-level change in personality traits across the life course: A meta-analysis of longitudinal studies. Psychological Bulletin, 132, 1–25http://psycnet.apa.org/doi/10.1037/ 0033-2909.132.1.1 Rubin, D. B. (1987). Multiple imputation for nonresponse in surveys. New York, NY: Wiley. Segerstrom, S. C. (2000). Personality and the immune system: Models, methods, and mechanisms. Annals of Behavioral Medicine, 22, 180–190. http://dx.doi.org/10.1007/ BF02895112. Segerstrom, S. C., & Miller, G. E. (2004). Psychological stress and the human immune system: A meta-analytic study of 30 years of inquiry. Psychological Bulletin, 130, 601–630http://psycnet.apa.org/doi/10.1037/0033-2909.130.4.601 Stephan, Y., Sutin, A. R., Luchetti, M., & Terracciano, A. (2016). Allostatic load and personality: A 4-year longitudinal study. Psychosomatic Medicine, 78, 302–310. http://dx.doi. org/10.1097/PSY.0000000000000281. Steptoe, A., Breeze, E., Banks, J., & Nazroo, J. (2013). Cohort profile: The English longitudinal study of ageing. International Journal of Epidemiology, 42, 1640–1648. http://dx. doi.org/10.1093/ije/dys168.
210
M.S. Allen, S. Laborde Personality and Individual Differences 111 (2017) 205–210
Sutin, A. R., Terracciano, A., Deiana, B., Naitza, S., Ferrucci, L., Uda, M., ... Costa, P. T. (2010). High neuroticism and low conscientiousness are associated with interleukin-6. Psychological Medicine, 40, 1485–1493. http://dx.doi.org/10. 1017/S0033291709992029. Sutin, A. R., Milaneschi, Y., Cannas, A., Ferrucci, L., Uda, M., Schlessinger, D., ... Terracciano, A. (2012). Impulsivity-related traits are associated with higher white blood cell
counts. Journal of Behavioral Medicine, 35, 616–623. http://dx.doi.org/10.1007/ s10865-011-9390-0. Turiano, N. A., Mroczek, D. K., Moynihan, J., & Chapman, B. P. (2013). Big 5 personality traits and interleukin-6: Evidence for “healthy neuroticism” in a US population sample. Brain, Behavior, and Immunity, 28, 83–89.
Contents lists available at ScienceDirect
Personality and Individual Differences journal homepage: www.elsevier.com/locate/paid
Five factor personality traits and inflammatory biomarkers in the English longitudinal study of aging Mark S. Allen a,⁎, Sylvain Laborde b,c a b c
School of Psychology, Faculty of Social Sciences, University of Wollongong, Northfields Avenue, Wollongong 2522, NSW, Australia EA 4260, University of Caen, France Department of Performance Psychology, German Sport University Cologne, Germany
a r t i c l e
i n f o
Article history: Received 13 November 2016 Received in revised form 7 February 2017 Accepted 10 February 2017 Available online xxxx Keywords: C–reactive protein Fibrinogen Gerontology Immune function Neuroticism White blood cell
a b s t r a c t Inflammatory processes are putative mechanisms underlying chronic disease. In this study we explore linear and non-linear associations between personality and change in inflammatory markers (C–reactive protein [CRP], fibrinogen, and white blood cell count [WBC]). In total, 5294 older adults (Mage = 64.51 ± 8.34 years) provided blood samples in 2008 with 3751 providing follow-up samples four year later. Midway between the two assessments, participants completed a measure of personality. After controlling for participant demographics (e.g., age, gender) and health-related lifestyle factors (e.g., exercise, cigarette smoking), we found that higher levels of agreeableness and lower levels of conscientiousness were associated with higher CRP levels, and (for conscientiousness) an increase in CRP over time. Age moderation effects indicated that agreeableness and conscientiousness were related to WBC among older participants in the sample (over ~ 70 years of age). These findings provide evidence that agreeableness and conscientiousness traits are important for inflammatory biomarkers in older adulthood. © 2017 Elsevier Ltd. All rights reserved.
1. Introduction Inflammation is one of the first responses of the immune system to harmful stimuli. Inflammatory abnormalities underlie biological processes that contribute to cardiovascular disease and other age-related diseases (Emerging Risk Factors Collaboration, 2010, 2012; Grivennikov, Greten, & Karin, 2010; Libby, 2002). Inflammatory markers, including C–reactive protein (CRP) and fibrinogen, reflect demographic, anthropometric, behavioural, and psychosocial factors (McDade, Hawkley, & Cacioppo, 2006) including trait personality (Luchetti, Barkley, Stephan, Terracciano, & Sutin, 2014). The five factor model (McCrae & John, 1992) considers that personality is best captured through the assessment of five broad trait dimensions: neuroticism, extraversion, openness, agreeableness and conscientiousness. Of these, there is emerging evidence that high conscientiousness might have a protective role in inflammatory processes independent of behavioural and anthropometric factors such as cigarette smoking and body mass (Luchetti et al., 2014). However, the potential involvement of other trait dimensions is less clear. The current study sought to explore associations between the big five trait dimensions and three markers of
⁎ Corresponding author at: School of Psychology, University of Wollongong, Northfields Avenue, NSW 2522, Australia. E-mail address: [email protected] (M.S. Allen).
http://dx.doi.org/10.1016/j.paid.2017.02.028 0191-8869/© 2017 Elsevier Ltd. All rights reserved.
chronic inflammation (CRP, fibrinogen, and white blood cell count [WBC]) in a large sample of older adults from England. Personality is thought to relate to inflammatory biomarkers through behavioural and psychophysiological mechanisms (Segerstrom, 2000). For example, personality is a strong predictor of stress appraisal and coping (Carver & Connor-Smith, 2010), and psychological stress can activate acute phase reactants associated with inflammation (Black, 2002, 2003). Inflammation (and activation of the immune system) can be detected by an increase in several biomarkers including WBC, fibrinogen, CRP, and interleukin-6 (IL-6). The most consistent finding across studies is that lower levels of conscientiousness relate to higher levels of IL-6 and CRP (Chapman et al., 2011; Mõttus, Luciano, Starr, Pollard, & Deary, 2013; Stephan, Sutin, Luchetti, & Terracciano, 2016; Sutin et al., 2010; Turiano, Mroczek, Moynihan, & Chapman, 2013), as well as increases in these markers over time (Chapman et al., 2011; Turiano et al., 2013), and this was supported in a recent meta-analysis of seven (CRP) and six (IL-6) independent samples (Luchetti et al., 2014). In addition to conscientiousness, the meta-analysis also found that lower levels of openness related to higher CRP levels, and that effects remained significant after controlling for some behavioural and anthropometric factors (cigarette smoking and body mass). There were no significant effects for neuroticism, extraversion or agreeableness (but notable heterogeneity across study samples), and sample age was not found to moderate associations between personality and CRP. The authors pointed towards the small number of samples limiting statistical
206
M.S. Allen, S. Laborde Personality and Individual Differences 111 (2017) 205–210
power to detect age moderation effects, and recommended researchers continue to search for important moderators that might explain heterogeneity across studies (Luchetti et al., 2014). Less research has explored fibrinogen and WBC as correlates of trait personality. In a sample of Israeli adults, after controlling for body fat and health behaviour, lower levels of openness and higher levels of neuroticism were found to be associated with higher fibrinogen levels, and (for neuroticism) an increase in fibrinogen over four years (Armon, Melamed, Shirom, Berliner, & Shapira, 2013). However, a further study of Scottish older adults, that included two measures of personality, found that agreeableness had a negative association with fibrinogen for one personality measure, but overall, fibrinogen (and change in fibrinogen over time) was unrelated to personality (Mõttus et al., 2013). Similar findings were observed in a study exploring two of the big five traits, in which neuroticism and extraversion were unrelated to fibrinogen after controlling for body mass and health behaviour (Millar et al., 2013). For WBC, a large sample study of Sardinian adults found that broad trait dimensions were not consistently associated with WBC and were unrelated to change in WBC over three years (Sutin et al., 2012). However, the study did find that higher levels of openness and lower levels of conscientiousness were related to a WBC range associated with increased risk of mortality (≥7.6 × 109 cells/L). Taken together, there is good (but not robust) evidence that conscientiousness is important for CRP and IL-6 (but associations for other dimensions remain unclear), mixed findings for an association between personality and fibrinogen, and a shortage of research into personality and WBC meaning generalizable conclusions about important associations cannot be made at this stage. One explanation for some of the inconsistent findings might be that personality becomes more or less important for inflammatory processes among particular subgroups. Indeed, psychological stress is often the mechanism considered most responsible for connecting personality to inflammation (Armon et al., 2013) and psychological stress becomes more important for inflammatory responses as people become older (Segerstrom & Miller, 2004). This would suggest that personality might have an important role in chronic inflammation but only among older adults. The current study is interested in the association between personality and inflammatory markers independent of health behaviour, and therefore we control for health-related factors such as exercise, cigarette smoking and body mass. We hypothesised that lower levels of conscientiousness would be associated with higher inflammation, and increases in inflammation over time, as measured via CRP, fibrinogen and WBC. We also explore age as a potential moderator of associations between personality and inflammatory markers, and hypothesised stronger associations between personality and inflammatory markers among older adults. 2. Method 2.1. Sample The English Longitudinal Study of Aging (ELSA) is a biannual survey that collects information on health, social, and economic circumstances of the English population aged 50 and older. Postcode sectors are selected from the national Postcode Address File, and addresses are selected randomly from each sector. At wave 1 there was a 77.5% response rate. At waves 3, 4, and 6 the sample was replenished with new participants to maintain the size and representativeness of the panel. The data is broadly considered to be representative of older adults in England. Detailed information on sampling and attrition across waves is available elsewhere (Steptoe, Breeze, Banks, & Nazroo, 2013). Ethical approval was granted by the National Research and Ethics Committee UK. We refer to data collected at wave 4 as Time 1 (2008) and data collected at wave 6 as Time 2 (2012). In total, 10,749 adults aged 50 and over were sampled at Time 1 with 6384 (59.4%) providing blood samples. Compared to those that provided blood samples, non-sampled participants were older, t(10,747) = 4.65, p b 0.001, d = 0.09, less
physically active, t(10,737) = 15.43, p b 0.001, d = 0.30, and had lower levels of extraversion, t(7884) = 3.42, p = 0.001, d = 0.08, and conscientiousness, t(7866) = 3.75, p b 0.001, d = 0.08. Personality data were available for 5294 of the participants that provided blood samples (82.9%). Compared to participants with personality data, those without personality data were older, t(6382) = 17.39, p b 0.001, d = 0.44, less physically active, t(6381) = 12.81, p b 0.001, d = 0.32, had a lower body mass, t(6382) = 5.84, p b 0.001, d = 0.15, higher blood CRP levels, t(6382) = 5.88, p b 0.001, d = 0.15, and a higher WBC count, t(6233) = 6.65, p b 0.001, d = 0.17. At Time 2, 10,372 participants were included, with 6039 (58.2%) providing blood samples. Compared to those that provided blood samples, non-sampled participants were older, t(10,370) = 4.90, p b 0.001, d = 0.10, had a higher body mass, t(7742) = 6.43, p b 0.001, d = 0.15, were less physically active, t(10,365) = 13.46, p b 0.001, d = 0.26, and had lower levels of extraversion, t(7662) = 4.36, p b 0.001, d = 0.10, and conscientiousness, t(7652) = 4.19, p b 0.001, d = 0.10. There were personality data available for 4911 participants that provided blood samples (81.3%). Compared to participants with personality data, those without personality data were younger, t(6037) = 21.27, p b 0.001, d = 0.55, less physically active, t(6036) = 4.94, p b 0.001, d = 0.13, and had a higher WBC count, t(6037) = 3.52, p b 0.001, d = 0.09. In total, there were 5294 participants available at Time 1 (2908 women, 2386 men; Mage = 64.51 ± 8.34 years), 4911 participants at Time 2 (2702 women, 2209 men; Mage = 67.90 ± 8.20 years), and 3751 participants sampled at both Time 1 and Time 2 (2065 women, 1686 men; Mage (Time 1) = 64.14 ± 7.94 years). 2.2. Measures 2.2.1. Personality Data on personality were collected midway between Time 1 and Time 2 (at wave 5) using a modified version of the Midlife Development Inventory (MDI; Lachman & Weaver, 1997). ELSA included an additional item (“thorough”) to the original 25 item measure to increase internal consistency of the conscientiousness subscale. The questionnaire uses 26 adjectives to assess five personality traits: neuroticism (e.g., “nervous”), extraversion (e.g., “outgoing”), openness (e.g., “creative”), agreeableness (e.g., “helpful”), and conscientiousness (e.g., “organised”). Participants indicate how well each adjective described them on a fourpoint scale: 1 (a lot), 2 (some), 3 (a little) or 4 (not at all). Scales were reversed so that higher scores represent higher levels of neuroticism (α = 0.68), extraversion (α = 0.76), openness (α = 0.79), agreeableness (α = 0.80), and conscientiousness (α = 0.68). The MDI has demonstrated evidence of criterion validity in adult samples (e.g., Joshanloo, 2017). 2.2.2. Clinical assessment A trained nurse collected anthropometric data (body mass and height) and blood samples. Body mass was measured using Seca 877 scales and height was measured using a portable Stadiometer. Participants did not wear shoes for anthropometric measures but wore light clothing. Body mass index (BMI) was calculated as mass (kg)/height squared (m2). Blood samples were taken from the left or right arm using a Vacutainer or Butterfly needle, and a maximum of two attempts were made to extract blood. Measurement of CRP was carried out using the N Latex CRP mono Immunoassay on the Behring Nephelometer II Analyzer. Fibrinogen analysis was carried out using the Organon Teknika MDA 180 analyser, using a modification of the Clauss thrombin clotting method. WBC was measured using the Sysmex XE2100 analyser that uses electrical impedance technology (Coulter principle) to count the white cells combined with hydrodynamic focusing and flow cytometry to differentiate the blood cells. All analyses were carried out according to Standard Operating Procedures by state registered medical laboratory scientific officers.
M.S. Allen, S. Laborde Personality and Individual Differences 111 (2017) 205–210
207
Table 1 Bivariate correlations between biomarkers and other study variables. Time 1
Age Gender Ethnicity Past smoker Current smoker Exercise BMI Neuroticism Extraversion Openness Agreeableness Conscientiousness
Time 2
CRP
Fibrinogen
WBC
CRP
Fibrinogen
WBC
0.08⁎⁎ 0.04⁎ 0.01 0.08⁎⁎ 0.11⁎⁎
0.14⁎⁎ 0.08⁎⁎ 0.01 0.08⁎⁎ 0.16⁎⁎
0.01 −0.05⁎⁎ 0.01 0.15⁎⁎ 0.30⁎⁎ −0.11⁎⁎ 0.13⁎⁎
0.09⁎⁎ 0.06⁎⁎ −0.01 0.07⁎⁎ 0.09⁎⁎ −0.21⁎⁎ 0.31⁎⁎
0.12⁎⁎ 0.07⁎⁎ −0.01 0.06⁎⁎ 0.12⁎⁎ −0.15⁎⁎ 0.18⁎⁎
0.06⁎⁎ −0.06⁎⁎ −0.01 0.12⁎⁎ 0.26⁎⁎ −0.13⁎⁎ 0.12⁎⁎
0.02 −0.06⁎⁎ −0.01 0.00 −0.06⁎⁎
−0.00 −0.04⁎ −0.04⁎ 0.05⁎ −0.09⁎⁎
−0.02 −0.04⁎ −0.04⁎ 0.03 −0.07⁎⁎
0.01 −0.06⁎⁎ −0.03 −0.01 −0.07⁎⁎
−0.20⁎⁎ 0.37⁎⁎ 0.02 −0.05⁎⁎ −0.01 0.06⁎⁎ −0.08⁎⁎
−0.16⁎⁎ 0.20⁎⁎ 0.01 −0.07⁎⁎ −0.05⁎ 0.02 −0.08⁎⁎
Note: Gender was coded as 1 (men) or 2 (women), ethnicity was coded as 1 (White) or 2 (non-White), past and current smoker were coded as 1 (no) or 2 (yes). ⁎ p b 0.01, ⁎⁎ p b 0.001.
To explore associations between personality and inflammatory markers we ran nine regression models. CRP, fibrinogen and WBC were set as dependent variables and we explored these markers at Time 1, Time 2 and the change in these markers over time. In all models we controlled for age, gender, ethnicity, physical activity, past smoking, current smoking, and BMI. For change scores, we entered the Time 1 score as an additional covariate with the Time 2 score set as the dependent variable. Because CRP showed extreme positive skew (see Supplementary File), raw data were transformed into ordered categories according to relative risk (Ridker & Cook, 2004). We further tested whether associations between personality and inflammatory markers varied as a function of age and gender. Interaction terms were computed from standardised data, and significant effects were followed up using simple slope analyses (Hayes, 2013). Consistent with previous research (Sutin et al., 2012) we ran additional logistic regression models for WBC at Time 1 and Time 2 using a threshold of ≥ 7.6 × 109 cells/L that has been found to be associated with an increased risk of cardiovascular mortality (Brown, Giles, & Croft, 2001). Prior to analyses, missing data were imputed for 809 empty cells at Time 1 (1.3% of the data set) and 659 empty cells at Time 2 (1.1% of the data set) using multiple imputation (Rubin, 1987). For the linear regression models, there was no evidence of multicollinearity (VIF values b 3.00) and no multivariate outliers (Cook's distance values b 0.05). For the logistic regression, the Hosmer–Lemeshow test was non-significant at all steps of the model (p N 0.001). For all statistical tests, significance was set at p b 0.01.
Time 1, Time 2 and for CRP change. There were no moderation effects by gender, but some associations were moderated by age. At Time 1, there was a significant age moderation of agreeableness on WBC (βinteraction = 0.05, p = 0.004). Simple slope analyses demonstrated that agreeableness was unrelated to WBC among younger participants (tmean age 56.2 = − 0.47, p = 0.635) and the sample average (tmean age 64.5 = 2.15, p = 0.032), but had a positive association among older participants (tmean age 72.8 = 3.53, p b 0.001). This age moderation effect also emerged at Time 2 (βinteraction = 0.05, p = 0.005). Computation of Johnson-Neyman significance regions showed that agreeableness became a significant predictor of WBC for participants aged over 66.7 years at Time 1, and over 74.9 years at Time 2. At Time 2, there was also a significant age moderation of conscientiousness on WBC (βinteraction = −0.05, p = 0.006), with conscientiousness unrelated to WBC among younger participants (tmean age 59.7 = 1.30, p = 0.191) and the sample average (tmean age 67.9 = − 0.70, p = 0.483), but negatively related among older participants (tmean age 76.1 = − 2.33, p = 0.019). The effect for conscientiousness on WBC was significant for participants aged over 75.0 years. This age moderation effect also emerged for conscientiousness on WBC change (βinteraction = −0.04, p = 0.004), with conscientiousness becoming a significant predictor of change in WBC for participants aged over 72.2 years (at Time 1). No age or gender moderation effects emerged for CRP or fibrinogen. Findings from the logistic regression models are reported in the Supplementary File. No significant main effects emerged and there were no gender moderations. There was a significant age moderation of agreeableness on WBC at Time 1 (b = 0.14, s.e. = 0.04, p = 0.001) and Time 2 (b = 0.15, s.e. = 0.04, p b 0.001). Similar to the linear regression model, agreeableness was unrelated to WBC among younger participants (zmean age 56.2 = −1.67, p = 0.096; zmean age 59.7 = −1.87, p = 0.060) and the sample average (zmean age 64.5 = 0.95, p = 0.343; zmean age 67.9 = 0.78, p = 0.437), but had a positive association among older participants (zmean age 72.8 = 2.96, p = 0.003; zmean age 76.1 = 3.02, p = 0.002). Computation of significance regions showed that agreeableness became a significant predictor of WBC categories for participants aged over 70.1 years at Time 1, and over 73.4 years at Time 2.1
3. Results
4. Discussion
Table 1 provides correlations between inflammatory markers and other measured variables, and findings from the linear regression models are reported in Table 2. The data show that extraversion had negative associations with inflammatory markers at Time 1, but these associations did not re-emerge at Time 2. A positive association emerged between agreeableness and CRP at Time 1 and Time 2, and a negative association emerged between conscientiousness and CRP at
This study explored associations between personality and inflammatory markers in a large sample of English adults aged 50 and over. After controlling for behavioural and anthropometric factors (e.g., cigarette
2.2.3. Control variables Participants reported their date of birth, gender, ethnicity, whether they currently smoked cigarettes, and whether they had ever smoked cigarettes. Participants also reported whether they “take part in sports or activities that are moderately energetic” with response categories of 1 (hardly ever, or never), 2 (one to three times a month), 3 (once a week) or 4 (more than once a week). 2.3. Data analysis
1 Similar (but non-significant) effects emerged for conscientiousness at Time 1 (b = −0.10, s.e. = 0.04, p = 0.014) and Time 2 (b = −0.10, s.e. = 0.04, p = 0.013) – indicative of stronger associations among older participants.
208
M.S. Allen, S. Laborde Personality and Individual Differences 111 (2017) 205–210
Table 2 Regression models exploring linear associations between personality and inflammatory biomarkers. C–reactive protein
Fibrinogen
White blood cell count
Time 1
Time 2
Change
Time 1
Time 2
Change
Time 1
Time 2
Change
0.03 0.09⁎⁎ 0.02 0.02 0.13⁎⁎
0.04⁎ 0.09⁎⁎ −0.01 0.03 0.10⁎⁎
0.02 0.04⁎ −0.01 0.02 0.04 −0.04⁎
0.07⁎⁎ 0.15⁎⁎ 0.01 0.02 0.18⁎⁎
0.03 0.05⁎⁎ −0.02 −0.00 0.05⁎⁎
−0.07⁎⁎ 0.08⁎⁎ 0.00 0.04 0.26⁎⁎ −0.07⁎⁎ 0.13⁎⁎
−0.04⁎ 0.07⁎⁎ −0.01 0.00 0.07⁎⁎
−0.03 0.07⁎⁎
−0.06⁎⁎ 0.04⁎ 0.01 0.05⁎ 0.30⁎⁎ −0.05⁎ 0.15⁎⁎
−0.03 0.04⁎
Step 1 Gender Age Ethnicity Past smoker Current smoker Exercise BMI
−0.10⁎⁎ 0.35⁎⁎
−0.11⁎⁎ 0.30⁎⁎
0.03
−0.07⁎⁎ 0.20⁎⁎
0.06⁎⁎ 0.12⁎⁎ −0.01 0.02 0.14⁎⁎ −0.07⁎⁎ 0.18⁎⁎
Step 2 Neuroticism Extraversion Openness Agreeableness Conscientiousness
0.01 −0.06⁎⁎ 0.04 0.07⁎⁎ −0.04⁎
−0.02 −0.01 −0.01 0.05⁎ −0.04⁎
−0.01 −0.01 0.00 0.01 −0.05⁎
−0.00 −0.06⁎⁎ 0.02 0.04 −0.02
−0.04 −0.03 0.00 0.04 −0.04
−0.01 −0.02 0.02 0.01 −0.04
−0.00 −0.06⁎⁎ 0.02 0.04 −0.01
0.00 −0.02 −0.01 0.02 −0.01
0.01 −0.01 −0.01 −0.00 −0.00
Step 3 Age × neuroticism Age × extraversion Age × openness Age × agreeableness Age × conscientiousness R2
0.01 0.01 0.02 0.01 0.00 0.18
0.01 0.03 −0.03 0.01 −0.01 0.14
0.00 0.02 −0.03 0.01 −0.01 0.35
0.00 0.01 0.02 −0.01 0.01 0.11
0.01 0.01 −0.02 0.01 −0.00 0.08
0.00 0.03 −0.04 0.00 0.00 0.32
−0.01 −0.01 0.01 0.05⁎ −0.03 0.13
−0.00 0.02 −0.02 0.05⁎ −0.04⁎ 0.11
0.00 0.01 −0.02 0.02 −0.04⁎ 0.40
Note: Standardised regression coefficients reported. ⁎ p b 0.01, ⁎⁎ p b 0.001.
smoking, body mass), we found that higher levels of agreeableness and lower levels of conscientiousness were associated with higher CRP levels, and (for conscientiousness) an increase in CRP levels over four years. Extraversion had inconsistent associations with inflammatory markers (negative associations at Time 1 but not Time 2), and neuroticism and openness were unrelated to inflammatory markers. Age moderation effects showed that agreeableness and conscientiousness were important for WBC among older participants in the sample. These findings add to the growing body of literature that is beginning to uncover important connections between personality and markers of chronic inflammation. The finding that lower levels of conscientiousness were related to higher levels of CRP and increases in CRP over time is consistent with previous research (Luchetti et al., 2014; Mõttus et al., 2013; Stephan et al., 2016; Sutin et al., 2010). Conscientious individuals are characterised by organisation, discipline and self-control (McCrae & John, 1992), and tend to respond to daily life stress with more adaptive coping responses (Carver & Connor-Smith, 2010). We can speculate that conscientiousness might offer protection against elevated CRP levels (indicative of chronic inflammation) because conscientious individuals have more resources available to cope with exposure to stress. The size of the effect was small (see Gignac & Szodorai, 2016) and this is consistent with findings in other large sample studies (Luchetti et al., 2014). Small effects are common in research on personality and health outcomes (Bogg & Roberts, 2004) but can, nonetheless, have substantial practical consequences at the population level. We also found that higher levels of agreeableness related to higher CRP levels. This association was not hypothesised and is inconsistent with findings from previous studies that show agreeableness is generally unrelated to CRP and other inflammatory biomarkers (Luchetti et al., 2014; Mõttus et al., 2013; Stephan et al., 2016; Sutin et al., 2010). Of the seven samples included in the meta-analysis by Luchetti et al. (2014), only one (the Midlife in the United States [MIDUS] sample) showed a positive association between agreeableness and CRP. The MIDUS uses the same personality measure as that used in the current study – the MIDI (Lachman & Weaver, 1997) – and the extreme heterogeneity across study samples might be a function of differences between personality measures (see Luchetti et al., 2014; Mõttus et al., 2013). To speculate on the positive association found here, individuals scoring high on
agreeableness tend to be more cooperative, considerate, kind and sympathetic (McCrae & John, 1992), but are eager to avoid unsettling others and, despite being more support giving, are less support seeking (often using disengagement coping strategies such as withdrawal; Carver & ConnorSmith, 2010). This might, under certain circumstances, place agreeable individuals at higher risk of prolonged stress, contributing to elevated CRP levels. Nevertheless, given the mixed findings observed for CRP and agreeableness in previous research, the positive association found here should be interpreted with caution. Findings showed that agreeableness and conscientiousness were unrelated to WBC and change in WBC over time, and this is consistent with findings from a large sample of Sardinian adults (Sutin et al., 2012). We also found that WBC was unrelated to personality dimensions in our non-linear analyses of WBC and this is inconsistent with findings from Sutin et al. who found that openness and conscientiousness differed between high-risk and low-risk categories of WBC. Our findings did, nonetheless, provide evidence for an association between WBC and the traits of agreeableness and conscientiousness among the oldest participants in the sample. This result was hypothesised based on findings that age increases vulnerability to stress-related decreases in functional immune measures (Segerstrom & Miller, 2004), and suggests that high conscientiousness and low agreeableness might offer a protective role against elevated WBC among the elderly. It is unclear whether previous research tested for age moderation effects, and further research is needed to verify this finding given the problems that come with testing interaction effects in multiple regression field tests (see McClelland & Judd, 1993). The differences in findings (for the non-linear analyses) might reflect differences in lifestyle factors between the population sampled by Sutin et al. (adults from the Mediterranean island of Sardinia) and those sampled in the current study. That neuroticism, extraversion and openness were unimportant for fibrinogen and CRP is supportive of most previous research (Luchetti et al., 2014; Millar et al., 2013; Mõttus et al., 2013). However, findings of this study did show that higher extraversion related to higher inflammation (as measured by all three markers) at Time 1, but the association did not re-emerge at Time 2. Similar to previous studies (e.g., Sutin et al., 2012), we consider inconsistent findings as not supportive of an association. However, it is worth noting that extraversion was found to have a
M.S. Allen, S. Laborde Personality and Individual Differences 111 (2017) 205–210
negative association with CRP in a demographically similar sample of older adults from the United States (Stephan et al., 2016), and therefore a potential connection between extraversion and inflammatory biomarkers might warrant further investigation. Some research has also found that neuroticism relates to fibrinogen (Armon et al., 2013) and CRP (Sutin et al., 2010) and, given the importance of neuroticism in stress appraisal and coping (Carver & Connor-Smith, 2010), this association might also warrant further scrutiny. Strengths of this study include the large sample, measurement of inflammatory biomarkers over two time points, and control of anthropometric and health-related factors. However, there are some important limitations that require consideration when interpreting study findings. The most obvious limitation is that the measurement of personality was not concurrent with sampling of blood. We have interpreted the measure of personality (assessed midway between blood sample measures) as providing an approximation of older adults' personality at both time points – two years earlier and two years later – but it is possible (and likely) that personality changed, at least to some degree, over the four years. Personality is more stable in later adulthood than early adulthood, but traits (and agreeableness in particular) have been found to change meaningfully in later life (see Roberts, Walton, & Viechtbauer, 2006). This means that our assessment of personality might not entirely reflect personality at Time 1 or Time 2, and this might have contributed additional measurement error attenuating some associations. A second limitation is that the final sample was not entirely representative of the English older adult population. Attrition analyses showed that excluded participants (with missing blood sample or personality data) were older, had higher CRP and WBC levels, and lower levels of extraversion and conscientiousness. This might also have attenuated some associations, and findings should be considered a reflection of sampled participants that might not necessarily be representative of all older adults living in England. A third limitation is that the study does not address the processes connecting personality to inflammatory biomarkers. Future research might look to explore stress and other lifestyle factors as potential mediators of the associations between personality and chronic inflammation. To conclude, a number of important findings have emerged from this study. First, the study reinforces findings from previous research demonstrating a negative association between conscientiousness and change in CRP. The study also provides evidence that higher levels of agreeableness relate to higher CRP levels, and that higher levels of agreeableness and lower levels of conscientiousness relate to a higher WBC among elderly individuals. These findings all controlled for demographic, anthropometric and health-related factors. Effect sizes for significant associations were small, and below the RMPE (recommended minimum effect size representing a ‘practically’ significant effect) for social science research (Ferguson, 2009), but small effects can have meaningful practical consequences at the population level. More research is required before practical applications of this research can be made in confidence, but should findings be replicated in subsequent independent research, they might have implications in terms of early identification of individuals at risk of chronic inflammation. We recommend prospective studies explore personality and inflammation in alternative cultures, and aim to identify potential factors that might explain some of the inconsistent findings observed across study samples. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.paid.2017.02.028.
References Armon, G., Melamed, S., Shirom, A., Berliner, S., & Shapira, I. (2013). The associations of the Five Factor Model of personality with inflammatory biomarkers: A four-year prospective study. Personality and Individual Differences, 54, 750–755http://dx.doi.org/10.1016/ j.paid.2012.11.035
209
Black, P. H. (2002). Stress and the inflammatory response: A review of neurogenic inflammation. Brain, Behavior, and Immunity, 16, 622–653http://dx.doi.org/10.1016/S08891591(02)00021-1 Black, P. H. (2003). The inflammatory response is an integral part of the stress response: Implications for atherosclerosis, insulin resistance, type II diabetes and metabolic syndrome X. Brain, Behavior, and Immunity, 17, 350–364http://dx.doi.org/10.1016/ S0889-1591(03)00048-5 Bogg, T., & Roberts, B. W. (2004). Conscientiousness and health-related behaviors: A meta-analysis of the leading behavioral contributors to mortality. Psychological Bulletin, 130, 887–919http://psycnet.apa.org/doi/10.1037/0033-2909.130.6.887 Brown, D. W., Giles, W. H., & Croft, J. B. (2001). White blood cell count: An independent predictor of coronary heart disease mortality among a national cohort. Journal of Clinical Epidemiology, 54, 316–322http://dx.doi.org/10.1016/S0895-4356(00)00296-1 Carver, C. S., & Connor-Smith, J. (2010). Personality and coping. Annual Review of Psychology, 61, 679–704. http://dx.doi.org/10.1146/annurev.psych.093008. 100352. Chapman, B. P., van Wijngaarden, E., Seplaki, C. L., Talbot, N., Duberstein, P., & Moynihan, J. (2011). Openness and conscientiousness predict 34-week patterns of interleukin-6 in older persons. Brain, Behavior, and Immunity, 25, 667–673http://dx.doi.org/10.1016/j. bbi.2011.01.003 Emerging Risk Factors Collaboration. (2010). C-reactive protein concentration and risk of coronary heart disease, stroke, and mortality: An individual participant meta-analysis. The Lancet, 375, 132–140http://dx.doi.org/10.1016/S01406736(09)61717-7 Emerging Risk Factors Collaboration. (2012). C-reactive protein, fibrinogen, and cardiovascular disease prediction. New England Journal of Medicine, 367, 1310–1320. http://dx.doi.org/10.1056/NEJMoa1107477. Ferguson, C. J. (2009). An effect size primer: A guide for clinicians and researchers. Professional Psychology, 40, 532–538http://psycnet.apa.org/doi/10.1037/ a0015808 Gignac, G. E., & Szodorai, E. T. (2016). Effect size guidelines for individual differences researchers. Personality and Individual Differences, 102, 74–78http://dx.doi.org/10. 1016/j.paid.2016.06.069 Grivennikov, S. I., Greten, F. R., & Karin, M. (2010). Immunity, inflammation, and cancer. Cell, 140, 883–899http://dx.doi.org/10.1016/j.cell.2010.01.025 Hayes, A. F. (2013). Introduction to mediation, moderation, and conditional process analysis. New York, NY: Guilford. Joshanloo, M. (2017). Factor structure and criterion validity of original and short versions of the Negative and Positive Affect Scale (NAPAS). Personality and Individual Differences, 105, 233–237http://dx.doi.org/10.1016/j.paid.2016.09.060 Lachman, M. E., & Weaver, S. L. (1997). The Midlife Development Inventory (MIDI) personality scales. Waltham, MA: Brandeis University. Libby, P. (2002). Inflammation in atherosclerosis. Nature, 420, 868–874. http://dx.doi.org/ 10.1038/nature01323. Luchetti, M., Barkley, J. M., Stephan, Y., Terracciano, A., & Sutin, A. R. (2014). Five-factor model personality traits and inflammatory markers: New data and a meta-analysis. Psychoneuroendocrinology, 50, 181–193http://dx.doi.org/10.1016/j.psyneuen.2014. 08.014 McClelland, G. H., & Judd, C. M. (1993). Statistical difficulties of detecting interactions and moderator effects. Psychological Bulletin, 114, 376–390http://psycnet.apa.org/doi/10. 1037/0033-2909.114.2.376 McCrae, R. R., & John, O. P. (1992). An introduction to the five-factor model and its applications. Journal of Personality, 60, 175–215. http://dx.doi.org/10.1111/j.1467-6494. 1992.tb00970.x. McDade, T. W., Hawkley, L. C., & Cacioppo, J. T. (2006). Psychosocial and behavioral predictors of inflammation in middle-aged and older adults: The Chicago health, aging, and social relations study. Psychosomatic Medicine, 68, 376–381. http://dx.doi.org/ 10.1097/01.psy.0000221371.43607.64. Millar, K., Lloyd, S. M., McLean, J. S., Batty, G. D., Burns, H., Cavanagh, J., ... Mõttus, R. (2013). Personality, socio-economic status and inflammation: Cross-sectional, population-based study. PloS One, 8, e58256http://dx.doi.org/10.1371/journal.pone. 0058256 Mõttus, R., Luciano, M., Starr, J. M., Pollard, M. C., & Deary, I. J. (2013). Personality traits and inflammation in men and women in their early 70s: The Lothian Birth Cohort 1936 study of healthy aging. Psychosomatic Medicine, 75, 11–19. http://dx.doi.org/ 10.1097/PSY.0b013e31827576cc. Ridker, P. M., & Cook, N. (2004). Clinical usefulness of very high and very low levels of Creactive protein across the full range of Framingham risk scores. Circulation, 109, 1955–1959. http://dx.doi.org/10.1161/01.CIR.0000125690.80303.A8. Roberts, B. W., Walton, K. E., & Viechtbauer, W. (2006). Patterns of mean-level change in personality traits across the life course: A meta-analysis of longitudinal studies. Psychological Bulletin, 132, 1–25http://psycnet.apa.org/doi/10.1037/ 0033-2909.132.1.1 Rubin, D. B. (1987). Multiple imputation for nonresponse in surveys. New York, NY: Wiley. Segerstrom, S. C. (2000). Personality and the immune system: Models, methods, and mechanisms. Annals of Behavioral Medicine, 22, 180–190. http://dx.doi.org/10.1007/ BF02895112. Segerstrom, S. C., & Miller, G. E. (2004). Psychological stress and the human immune system: A meta-analytic study of 30 years of inquiry. Psychological Bulletin, 130, 601–630http://psycnet.apa.org/doi/10.1037/0033-2909.130.4.601 Stephan, Y., Sutin, A. R., Luchetti, M., & Terracciano, A. (2016). Allostatic load and personality: A 4-year longitudinal study. Psychosomatic Medicine, 78, 302–310. http://dx.doi. org/10.1097/PSY.0000000000000281. Steptoe, A., Breeze, E., Banks, J., & Nazroo, J. (2013). Cohort profile: The English longitudinal study of ageing. International Journal of Epidemiology, 42, 1640–1648. http://dx. doi.org/10.1093/ije/dys168.
210
M.S. Allen, S. Laborde Personality and Individual Differences 111 (2017) 205–210
Sutin, A. R., Terracciano, A., Deiana, B., Naitza, S., Ferrucci, L., Uda, M., ... Costa, P. T. (2010). High neuroticism and low conscientiousness are associated with interleukin-6. Psychological Medicine, 40, 1485–1493. http://dx.doi.org/10. 1017/S0033291709992029. Sutin, A. R., Milaneschi, Y., Cannas, A., Ferrucci, L., Uda, M., Schlessinger, D., ... Terracciano, A. (2012). Impulsivity-related traits are associated with higher white blood cell
counts. Journal of Behavioral Medicine, 35, 616–623. http://dx.doi.org/10.1007/ s10865-011-9390-0. Turiano, N. A., Mroczek, D. K., Moynihan, J., & Chapman, B. P. (2013). Big 5 personality traits and interleukin-6: Evidence for “healthy neuroticism” in a US population sample. Brain, Behavior, and Immunity, 28, 83–89.