SEARCH

SEARCH BY CITATION

Keywords:

  • Ageing;
  • body size;
  • C-reactive protein;
  • distribution;
  • epidemiology;
  • inflammation

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Address
  9. References
  10. Supporting Information

Background

C-reactive protein (CRP) is a well-documented predictor of cardiovascular diseases and mortality. We aimed to better understand the distribution and determinants of CRP in the population.

Materials and methods

Study participants were men and women aged 40–79 in the UK-based EPIC-Norfolk population-based cohort study. CRP was measured in 18 586 available serum samples (8334 men and 10 252 women) and remeasured in 6087 individuals on average 13 years later using a high-sensitivity assay.

Results

In cross-sectional analyses, the range of serum CRP was 0·1–188·3 mg/L and the median 1·6 mg/L. A third of the population had serum CRP levels above 3 mg/L. Serum CRP levels were comparable in men and women who were not taking postmenopausal hormone replacement therapy (HRT). Women who were taking HRT had double CRP levels compared with HRT nonusers. Smoking was also strongly related to CRP in men and women. Serum CRP was positively and independently associated with age, body mass index and waist circumference and inversely with height. A stronger association with serum CRP measured concurrently than on average 13 years later indicated a short-term rather than long-term association with smoking and HRT use. Social class and alcohol intake were not independently related to CRP, but there was a strong inverse association with educational status.

Conclusion

The distribution of serum CRP in the population is similar in men and women after taking into account smoking and HRT use. Anthropometric factors as well as educational status are strongly related to serum CRP.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Address
  9. References
  10. Supporting Information

C-reactive protein (CRP), synthesized by the liver, is an acute-phase reactant and a systemic marker of acute and chronic inflammation [1, 2]. CRP has been widely used in clinical practice as a nonspecific marker in diagnosis and monitoring of various conditions such as infectious diseases, autoimmune and rheumatologic disorders.

The use of CRP in clinical practice has expanded with the availability of high-sensitivity assays that can measure CRP levels as low as 0·1 mg/L. Elevated levels of high-sensitivity CRP are associated with higher risk of coronary heart disease [3], ischemic stroke [3] and type 2 diabetes [4]. Chronic inflammation of infectious or noninfectious origin has a role in the pathogenesis of cancers [5, 6], and elevated levels of CRP are associated with higher risk of cancer [7-9] and specifically lung [8-11] and colon [8, 12-14] cancers. Moreover, CRP is associated with survival in patients with many types of cancers [5, 15].

C-reactive protein has potential to be studied as a marker for prediction of risk or prognosis of various noncommunicable and communicable diseases. CRP may be involved in the pathogenesis of various conditions, as it has a role in innate and adaptive immunity and has both anti- and pro-inflammatory properties [2]. The level of CRP can rise rapidly (within 24–72 h) from < 1 to over 500 mg/L in response to inflammation, infection, tissue injury, etc. However, in large population studies, the within-person variability of CRP is similar to that of plasma cholesterol level and systolic blood pressure [16, 17]. Availability of standardized assays of this protein adds to the utility of this potential marker for risk prediction.

To be able to study the properties of this marker as a prognostic or risk factor for noncommunicable conditions and to assess the use of different cut-offs for different population strata for risk prediction purposes, we need to understand the population distribution and determinants of this marker. Although determinants of serum CRP have been reported previously, we aimed to explore and to quantify the effect of various factors on serum CRP in a large sample of the older adult population and especially explore potential differences in men and women.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Address
  9. References
  10. Supporting Information

Study participants were recruited as part of the European Prospective Investigation into Cancer in Norfolk (EPIC-Norfolk) population study. The detailed design of this ongoing cohort study has been previously described [18]. Briefly, 25 639 men and women, aged 40–79, resident in Norfolk, England, were recruited to the study using general practice age and sex registers which were approximately similar to the age–sex population structure in the United Kingdom. They attended the baseline health examination in 1993–1997, a second health examination on average 3·6 years and the third health examination at approximately 13 years after recruitment in the study. The EPIC-Norfolk study was approved by the Norfolk Local Research Ethics Committee, and all volunteers gave written informed consent.

Smoking status (current, former, never), alcohol consumption (units per week) and other lifestyle factors were assessed as part of a health and lifestyle questionnaire administered at baseline examination and at the third health examination. History of diseases and medication use were assessed by self-report on the health and lifestyle questionnaire plus by examining medications brought by the participants to the clinic visit. Social class was classified according to the Registrar General's occupation-based classification scheme into six categories [19].

At the clinic visit, trained nurses took anthropometric measurements on individuals in light clothing without shoes. Height was measured to the nearest millimetre using a free-standing stadiometer, and weight was measured to the nearest 100gr using a digital scale. Waist and hip circumferences were measured to the nearest millimetre using a D-loop tape measure. Body mass index (BMI) was calculated as weight in kilograms divided by the square of height in metres. Waist-to-hip ratio (WHR) was obtained by dividing waist circumference by hip circumference both in centimetres [18].

Serum high-sensitivity CRP was measured at baseline and at the third health examination. At baseline, nonfasting blood samples were obtained and stored overnight in a dark box in a refrigerator at 4–7 °C and then centrifuged at 2100 g for 15 min at 4 °C. Aliquots of serum used for the CRP analyses were stored in −80 °C freezers. In 2008, frozen samples from the baseline study were retrieved, thawed and assayed for serum high-sensitivity CRP using the Olympus AU640 chemistry analyzer (Olympus Diagnostics, Watford, UK). Serum CRP at the third health examination was measured in fresh blood samples using the Siemens Dimension clinical chemistry analyzer (Newark, DE, USA). Serum concentrations of total cholesterol, HDL cholesterol and triglycerides were measured in fresh serum samples at baseline and third health examination with the RA1000 Tecnicon analyser (Bayer Diagnostics, Basingstoke, UK), and LDL cholesterol concentrations were calculated with the Friedewald formula.

Statistical analysis

Serum CRP levels had a skewed distribution and were log-transformed to obtain a normal distribution. Linear regression analyses were performed to assess the independent association between predictors and loge CRP. For categorical variables, P-values were derived using F-test statistics. We checked for cases with probable large influence on the model by examining leverage versus squared residuals plots, and adjusted-variable plots. Multicollinearity was ruled out by calculating tolerance and variance inflation factor. Linear relationships between independent variables and outcome were explored using scatterplot matrices, plot of standardized residuals against each predictor variable and lowess smoothed regression lines.

In cross-sectional analysis of baseline values, waist circumference and height were categorized into sex-specific quintiles. General linear modelling was used to calculate mean loge CRP values across categories of predictors adjusted at the population average age, BMI, waist circumference, height, smoking (nonsmoking), alcohol intake, baseline prevalent disease and medication use in men, and additionally adjusted for postmenopausal hormone replacement therapy (HRT) (nonuse) and menopausal status in women. Geometric mean (95% confidence interval) serum CRP values (mg/L) were calculated by exponentiating crude- and adjusted- mean (95% CI) loge CRP values. Analyses were repeated to compare the magnitude of the association between baseline and third health examination values of anthropometric and lifestyle factors in prediction of serum CRP measured at the third health examination.

Pairwise Pearson's correlation coefficients were calculated for CRP, systolic and diastolic blood pressure, triglycerides, LDL and HDL cholesterol measurements at the first and third health examination. All statistical analyses were performed using Stata 10.0 (StataCorp. 2007. Stata Statistical Software: Release 10. College Station, TX, USA: StataCorp LP). Reporting of the study conforms to STROBE statement [20].

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Address
  9. References
  10. Supporting Information

Serum high-sensitivity CRP was measured in 18 586 available serum samples (8334 men and 10 252 women) from baseline. The baseline characteristics of the study population in whom CRP was measured are summarized in Table 1. There was no significant difference in baseline characteristics between the subgroups in whom CRP was and was not measured (data not shown). Serum CRP ranged from 0·1 to 188·3 mg/L. The median CRP was 1·6 mg/L in men and women. The 5th and 95th percentile of the CRP distribution was 0·3 and 9·8 mg/L, respectively. The percentage of the population with CRP levels ≤ 1, 1–3 and > 3 mg/L were 36%, 37% and 27%, respectively. Only 4·7% had CRP levels > 10 mg/L. The distributions of loge CRP were similar in men and women. Men and women with lower levels of CRP tended to be younger, leaner and taller than those with higher serum CRP levels were less likely to smoke, to take medications or postmenopausal HRT and had a lower prevalence of cancer, diabetes, myocardial infarction (MI), stroke or arthritis (Table 2).

Table 1. Baseline characteristics in men and women with available baseline C-reactive protein measurement in the European Prospective Investigation into Cancer in Norfolk cohort studya
 Men (N = 8334)Women (N = 10 252)
  1. a

    Values are mean (standard deviation) or number (percentage) unless otherwise stated.

  2. b

    Geometric mean CRP (95% CI).

Age59·1 (9·1)58·4 (9·2)
C-reactive protein (mg/L)b1·56 (1·53-1·59)1·64 (1·60-1·67)
Body mass index (kg/m2)26·5 (3·2)26·1 (4·2)
Waist circumference (cm)95·5 (9·5)81·8 (10·7)
Waist–hip ratio0·9 (0·1)0·8 (0·1)
Height (cm)174·0 (6·6)161·0 (6·2)
Smoking
Current984 (11·9%)1105 (10·9%)
Former4505 (54·4%)3235 (31·9%)
Never2788 (33·7%)5817 (57·3%)
Alcohol intake (units/week)10·2 (11·6)4·4 (5·5)
Social class
[I] Professionals630 (7·7%)627 (6·3%)
[II] Manager3140 (38·2%)3543 (35·4%)
[III M] Skilled nonmanual worker1019 (12·4%)2009 (20·0%)
[III NM] Skilled manual worker2086 (25·4%)2104 (20·1%)
[IV] Semi-skilled worker1084 (13·2%)1322 (13·2%)
[V] Nonskilled worker231 (2·8%)384 (3·8%)
Education level
Degree1282 (15·4%)1106 (10·8%)
A-level3794 (45·6%)3652 (35·6%)
O-level735 (8·8%)1170 (11·4%)
No qualification2517 (30·2%)4319 (42·2%)
Prevalent disease
Cancer303 (3·6%)656 (6·4%)
Diabetes258 (3·1%)147 (1·4%)
Myocardial infarction476 (5·7%)126 (1·2%)
Stroke154 (1·9%)106 (1·0%)
Arthritis1529 (18·4%)2746 (26·8%)
Medication use
Corticosteroids246 (3·0%)335 (3·3%)
Lipid-lowering drugs128 (1·5%)140 (1·4%)
NSAID1359 (16·3%)1376 (13·4%)
Aspirin1021 (12·3%)721 (7%)
Hormone replacement therapy
Current 2138 (20·9%)
Former 1164 (11·4%)
Never 6942 (67·8%)
Menopausal status
Premenopause 1684 (16·4%)
Perimenopause 564 (5·5%)
Postmenopause (2–5 years) 1913 (18·7%)
Postmenopause (> 5 years) 6083 (59·4%)
Table 2. Baseline characteristics in categories of C-reactive proteina
 CRP categories
< 1 mg/L1·1–3 mg/L3·1–10 mg/L> 10 mg/L
  1. a

    Values are mean (standard deviation) or number (percentage).

Men
N 302231691759384
Age56·3 (8·7)59·8 (9·0)61·6 (8·9)63·0 (8·4)
Body mass index25·4 (2·8)26·9 (3·0)27·5 (3·5)26·9 (3·7)
Waist circumference92·0 (8·7)96·7 (8·8)99·0 (9·9)97·9 (10·7)
Height175·0 (6·5)173·8 (6·7)173·1 (6·6)172·5 (6·7)
Current smokers (%)230 (7·7%)341 (10·8%)337 (19·3%)76 (19·8%)
Alcohol intake (units/week)10·1 (10·6)10·3 (12·1)10·1 (12·5)9·6 (11·3)
Prevalent disease
Cancer76 (2·5%)132 (4·2%)72 (4·1%)23 (6·0%)
Diabetes55 (1·8%)99 (3·1%)83 (4·7%)21 (5·5%)
Myocardial infarction95 (3·1%)186 (5·9%)153 (8·7%)42 (10·9%)
Stroke33 (1·1%)57 (1·8%)50 (2·8%)14 (3·7%)
Arthritis466 (15·4%)550 (17·4%)402 (22·9%)111 (28·9%)
Medication use
Corticosteroids54 (1·8%)93 (2·9%)75 (4·3%)24 (6·3%)
NSAID389 (12·9%)508 (16·0%)357 (20·3%)105 (27·3%)
Aspirin275 (9·1%)400 (12·6%)277 (15·8%)69 (18·0%)
Lipid-lowering drugs31 (1·0%)54 (1·7%)36 (2·1%)7 (1·8%)
Women
N 365036912420491
Age55·9 (9·1)59·3 (9·0)60·4 (8·8)60·2 (8·9)
Body mass index24·2 (3·1)26·2 (3·8)28·3 (4·7)28·4 (5·7)
Waist circumference76·9 (8·0)82·2 (9·5)87·6 (11·7)87·4 (12·9)
Height161·7 (6·1)160·8 (6·1)160·3 (6·1)160·3 (6·4)
Current smokers (%)359 (9·9%)388 (10·6%)293 (12·2%)65 (13·4%)
Alcohol intake (units/week)4·7 (5·5)4·3 (5·5)4·0 (5·4)4·1 (6·0)
Prevalent disease
Cancer217 (6·0%)238 (6·5%)172 (7·1%)29 (5·9%)
Diabetes26 (0·7%)55 (1·5%)54 (2·2%)12 (2·4%)
Myocardial infarction21 (0·6%)56 (1·5%)36 (1·5%)13 (2·7%)
Stroke27 (0·7%)37 (1·0%)31 (1·3%)11 (2·2%)
Arthritis784 (21·5%)1005 (27·2%)787 (32·5%)170 (34·6%)
Medication use
Corticosteroids74 (2·0%)115 (3·1%)115 (4·8%)31 (6·3%)
NSAID388 (10·6%)499 (13·5%)368 (15·2%)121 (24·6%)
Aspirin193 (5·3%)263 (7·1%)211 (8·7%)54 (11·0%)
Lipid-lowering drugs40 (1·1%)47 (1·3%)45 (1·9%)8 (1·6%)
Hormone replacement therapy515 (14·1%)791 (21·4%)669 (27·6%)163 (33·2%)
Postmenopause (vs. premenopause)2492 (68·3%)3044 (82·5%)2125 (87·8%)422 (86·0%)

Unadjusted median CRP (mg/L) levels increased with age in both men and women (Fig. 1). Women taking HRT had higher CRP levels than men and HRT nonusers across all age groups regardless of menopausal status. After adjusting for age, BMI, sex-specific waist quintiles, smoking, prevalent disease and medication use in multiple regression analysis, there was no significant difference in serum CRP levels between men and women who were not taking HRT (Table 3). There was no significant difference in serum CRP between pre- and postmenopausal women either.

Table 3. Geometric mean C-reactive protein (mg/L) in men and women by menopausal status and hormone replacement therapy (HRT) with progressive adjustment for covariates
Level of adjustmentMen (N = 8334)Women
HRT nonusersHRT users
Premenopause (N = 1929)Postmenopause (N = 6177)Premenopause (N = 319)Postmenopause (N = 1819)
Age-adjusted (age = 60)

1·62

(1·58, 1·66)

1·39

(1·31, 1·47)

1·56

(1·52, 1·60)

2·70

(2·39, 3·05)

2·56

(2·44, 2·69)

+ medication use and prevalent disease

1·61

(1·57, 1·64)

1·38

(1·31, 1·46)

1·57

(1·52, 1·60)

2·65

(2·35, 2·99)

2·53

(2·42, 2·66)

+ Smoking (nonsmoking)

1·42

(1·38, 1·47)

1·27

(1·20, 1·34)

1·44

(1·39, 1·48)

2·41

(2·14, 2·72)

2·29

(2·18, 2·41)

+ Body mass index (BMI = 25 kg/m2)

1·28

(1·24, 1·31)

1·24

(1·18, 1·31)

1·27

(1·24, 1·31)

2·32

(2·07, 2·60)

2·19

(2·08, 2·29)

+Waist (sex-specific quintiles)

1·33

(1·30, 1·37)

1·26

(1·20, 1·33)

1·32

(1·28, 1·36)

2·38

(2·13, 2·67)

2·25

(2·15, 2·37)

or + Waist (cm)

1·20

(1·16, 1·24)

1·41

(1·32, 1·48)

1·47

(1·42, 1·52)

2·66

(2·36, 2·98)

2·52

(2·39, 2·66)

image

Figure 1. Crude median serum C-reactive protein (mg/L) in men and women by menopausal status and postmenopausal hormone replacement therapy across age groups.

Download figure to PowerPoint

In sex-specific multiple regression analyses with loge CRP as outcome, the single most important determinant of serum CRP was smoking in men and HRT use in women (Table 4). Each of these factors was associated with approximately double CRP levels compared with nonsmokers and HRT nonusers, respectively. The association between smoking and serum CRP was less prominent in women. CRP had a positive log-linear association with age, BMI and waist circumference, had an inverse association with height and was not strongly associated with alcohol intake, or social class in men and women (Table 4, Fig. 2).

Table 4. Linear regression coefficients with loge C-reactive protein as outcomea
 MenWomen
N = 8260N = 10 138
Adjusted R2 = 0·20Adjusted R2 = 0·25
βbSEPβbSE P
  1. CRP, C-reactive protein.

  2. a

    Dependent variable is Loge CRP.

  3. b

    The multivariate adjusted β regression coefficient, when small, approximates proportion change in original serum CRP (mg/L) values per defined-unit increase in exposure.

Age (per 5 years)0·100·01< 0·0010·090·01< 0·001
Body mass index (per 4 kg/m2)0·060·030·030·220·02< 0·001
Waist (per 10 cm)0·260·02< 0·0010·200·02< 0·001
Height (per 10 cm)−0·160·02< 0·001−0·080·02< 0·001
Smoking
NeverReference < 0·001Reference 0·04
Former0·170·020·030·02
Current0·560·060·110·05
Alcohol (units per week)
0Reference 0·07Reference 0·31
1–6−0·040·04−0·030·03
7–13−0·110·04−0·050·03
14–20−0·040·05−0·100·05
≥ 21−0·070·04−0·010·07
Social class
[V] Nonskilled workerReference 0·03Reference 0·82
[IV] Semi-skilled worker−0·040·070·080·06
[III M] Skilled manual worker0·040·070·070·05
[III NM] Skilled nonmanual worker−0·030·070·080·05
[II] Manager−0·020·070·070·05
[I] Professional−0·140·080·100·06
Education level
Below O-levelReference 0·19Reference < 0·001
O-level−0·050·04−0·030·03
A-level−0·060·03−0·080·02
Degree−0·040·04−0·130·04
Season
April–SeptemberReference  Reference  
October–March0·110·02< 0·0010·090·02< 0·001
Prevalent disease
Cancer0·080·060·170·020·040·65
Diabetes0·080·060·200·150·080·06
Myocardial Infarction0·170·050·000·140·140·09
Cerebrovascular accident (stroke)0·130·080·110·010·040·10
Arthritis0·050·030·060·010·020·78
Medication use
Corticosteroids0·330·060·000·300·05< 0·001
NSAID0·040·040·260·040·030·22
Aspirin0·010·040·790·050·040·27
Lipid-lowering drug0·030·090·74−0·080·080·33
Postmenopause (vs. premenopause)   0·060·030·04
Hormone replacement therapy   0·560·03< 0·001
image

Figure 2. Multivariable-adjusted geometric mean C-reactive protein across categories of anthropometric and lifestyle factors in men and women.

Download figure to PowerPoint

Five years higher age was associated with ~10% increase in serum CRP levels in both men and women (Table 4). Geometric mean CRP (mg/L) was 1·0 for those aged 45–49 and 1·9 for those aged 75–79 in both men and women adjusted at the population average of anthropometric measures in nonsmokers and HRT nonusers.

Anthropometric measures

Among anthropometric measures, waist circumference was the strongest determinant of serum CRP. Ten centimetres (1-SD) higher waist circumference was associated with a 20% and 26% higher serum CRP in women and men, respectively (Table 4). BMI was a stronger determinant of CRP in women, and height was a stronger determinant in men. Women in the highest quintile of BMI compared with lowest (mean BMI difference ~11 kg/m2) had almost double CRP levels (Fig. 2). A similar effect was seen for waist circumference in men (Fig. 2). This magnitude of association was equivalent to HRT use in women and smoking in men, or being 30 years older in age. Hip circumference was not significantly related to CRP levels and was excluded from the models. There was no interaction (synergism) between BMI and waist circumference. The positive association between waist circumference and serum CRP persisted in the analysis restricted to men and women with a normal BMI (BMI < 25 kg/m2) (data not shown). The distribution of waist circumference in men had a similar shape (standard deviation) to that in women, but with a higher mean. The effect of relative waist circumference on serum CRP was comparable in men and women (Table 4). Although women in each sex-specific quintile of waist circumference had 11–15 cm lower waist circumferences compared with men in the corresponding quintile, CRP levels were comparable in men and women in the same sex-specific quintile of waist circumference (Fig. 2). Conversely, in men and women with an equal absolute waist circumference in centimetres, men had significantly lower CRP levels than women (Table 3). A similar pattern was observed when WHR and sex-specific WHR quintiles were entered in the model instead waist circumference (data not shown). The association between height and CRP was different in men and women in a similar manner.

Lifestyle factors

In terms of lifestyle factors, individuals with a higher education attainment had lower geometric mean CRP (Fig. 2). This effect was only statistically significant in women (Table 4). Although social class status was significantly related to loge CRP in men (Table 4), geometric mean CRP did not vary considerably across categories of social class after adjusting for other factors (Fig. 2). Alcohol intake was not significantly related to serum CRP (Fig. 2, Table 4). Total energy intake assessed by food frequency questionnaire was not associated with serum CRP and was excluded from the final model.

Geometric mean CRP was higher in the colder months of the year (October–March) compared with the warmer months (April–September) (geometric mean CRP 1·50 vs. 1·34 mg/L in men and 1·33 vs. 1·22 mg/L in women, respectively). The same pattern was seen after exclusion of individuals with CRP levels above 10 mg/L (data not shown).

Prevalent disease and medication use

Loge CRP followed a normal distribution in those with a history of cancer, diabetes, MI, stroke and arthritis. However, mean loge CRP in these individuals was higher than that in participants free of prevalent disease. In analyses involving men and women combined, geometric mean CRP (95% CI) adjusted at the population average of other variables, nonsmoking and HRT nonuse, was 1·44 (1·40–1·47) in the absence of prevalent disease and medication use. Geometric mean CRP (95% CI) was significantly higher in participants with a history of MI [1·69 (1·55–1·85)] and diabetes [1·58 (1·43–1·74)], but was not significantly higher in participants with a history of stroke [1·58 (1·40–1·79)], arthritis [1·48 (1·42–1·53)] or cancer [1·49 (1·40–1·59)]. Geometric mean (95% CI) CRP was significantly higher in men and women taking corticosteroid medication [1·96 (1·81–2·13)], but was not significantly affected by NSAIDs [1·49 (1·42–1·57)], Aspirin [1·46 (1·37–1·55)] or lipid-lowering [1·38 (1·23–1·56)] medication.

Intra-individual reproducibility of serum C-reactive protein

C-reactive protein was remeasured in 6087 individuals (2715 men and 3372 women) on average 13 years later. Pearson's correlation coefficient in this group of participants was 0·41 for CRP. In comparison, the corresponding Pearson's correlation coefficient was 0·39 for systolic blood pressure, 0·30 for diastolic blood pressure, 0·26 for LDL, 0·74 for HDL cholesterol and 0·51 for triglycerides (all P-values were < 0·001). Serum CRP measured at the third health examination had a stronger association with concurrent smoking, HRT use and waist circumference than with those at baseline (Table S1).

Repeating the analyses restricted to individuals with CRP levels below 10 mg/L, or analyses restricted to nonsmokers and HRT nonusers did not change the results.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Address
  9. References
  10. Supporting Information

The present study provides an estimate of the distribution and determinants of high-sensitivity CRP levels in a general population sample in the UK. Median CRP was 1·6 mg/L in the general population above the age of 40, and 95% had levels below 9·8 mg/L. The present study confirmed findings of previous studies [16, 17] that within-person variability of serum CRP (measured on average 13 years apart) was similar to systolic and diastolic blood pressure, LDL cholesterol and triglycerides.

Previous reports on the distribution of CRP in large general population samples are scarce. Reported median CRP was higher in the American population [21-24] and similar to our study in the European populations [25]. The median CRP was 2·7 mg/L in women and 1·6 mg/L in men in 21 000 participants over the age of 20 in the NHANES study [21-23]. In 24 500 participants of the women's health study [24], median CRP was 2·0 mg/L in white participants and 1·5 mg/L in HRT nonusers. In a general population sample from West Germany, France and Scotland, after exclusion of women taking oral contraceptives or HRT, median CRP ranged from 1·4 to 1·7 mg/L in men and women above 45 years of age in various cities [25]. In China and Japan, the reported median serum CRP was considerably lower; that is, 0·16 mg/L in men and 0·09 mg/L in women in 6000 participants above the age of 30 in Japan [26], and 0·61 mg/L in men and 0·51 mg/L in women among 3000 participants above 18 in Shanghai, China [27]. Differences in the distribution of CRP in these populations may be due to differences in the sampling method, the assay used to measure serum CRP, the age, prevalence of smoking or other determinants of CRP in the population. Due to the small number of non-Caucasian participants (0·4%) in the present study, we were unable to assess the effect of ethnicity on serum CRP levels.

In this general population sample, participants with a past history of diabetes and MI had a significant, and those with a history of cancer, stroke or arthritis had a nonsignificant higher level of CRP, which represented a right shift in the distribution of CRP and suggests a higher inflammatory state associated with such conditions. Although previous studies have suggested a CRP-lowering effect with statin therapy [28], we did not find such an association with lipid-lowering drugs in this study. However, similar to other studies [29], we found significantly higher CRP levels associated with corticosteroid medication. Serum CRP levels were not lower in participants taking Aspirin or NSAIDs.

One of the strengths of the present study is that this large population–based study enables direct comparisons between men and women. Previous population-based studies are not conclusive about possible differences in serum CRP level by sex [22, 25, 30, 31]. In this study, we found that the distribution and geometric mean/median serum CRP is not significantly different by sex among the older adult population. However, the relationship of anthropometric measures on serum CRP was slightly different in men and women. BMI was a stronger predictor of serum CRP in women than in men and height was a stronger predictor in men. With an equal waist circumference in centimetres, women had higher CRP levels than men, while men and women in the corresponding sex-specific quintile of waist circumference had comparable CRP levels. These findings might be explained by the magnitude of body fat mass. With an equal waist circumference or BMI, women having higher percentage body fat than men [32, 33], and another study found that men and women with similar age and fat mass had similar serum CRP levels [34]. The stronger association of CRP with BMI in women is consistent with previous studies [30, 35, 36].

Apart from absolute level of body fat, the association between distribution of body fat (central adiposity) and CRP also warrants special attention. Waist circumference, a measure of central obesity, was associated with higher levels of CRP independent of BMI and also in participants with a normal BMI. This finding is consistent with previous studies [37, 38]. Understanding the effect of total body fat and distribution of body fat on inflammation and CRP requires further investigation and is not entirely possible by anthropometric measures alone. Neither BMI is a perfect measure of total body fat (as higher BMI might indicate a higher muscular mass rather than a higher fat mass), nor waist circumference is a perfect measure of abdominal adiposity [32, 39]. As such, in the Framingham study [40], visceral adipose tissue (assessed by multidetector computed tomography) was associated with CRP after controlling for BMI and waist circumference, suggesting that BMI or waist do not completely account for visceral adiposity. Consistently, percentage body fat mass was reported to be a better predictor of high serum levels of CRP than any of the anthropometric measures in a Taiwanese population study [4]. However, it is useful to understand the association between serum CRP and anthropometric measures as they are more convenient to measure.

A major limitation of the present study is that in an observational study, it is not possible to establish causality. However, reduction in CRP levels induced by weight loss reported in interventional studies [41], underpins the likelihood of the causal association between obesity and increased CRP levels. This association can be explained by previously defined mechanisms. Adipose tissue produces cytokines that recruit monocytes and activate them to become macrophages. Activated macrophages release TNF-α and IL-6 which stimulate liver production of CRP [42, 43].

The association between serum CRP levels and height has rarely been discussed in previous studies. The mechanisms by which height may relate to serum CRP levels are not well established.

Postmenopausal HRT had a strong association with serum CRP in this study. Although causality cannot be established from this study, it has been shown in randomized controlled trials that HRT increases serum levels of CRP [44-48]. The mechanism by which HRT increases CRP is not well known. However, randomized controlled trials or observational prospective studies have shown that HRT was associated with a significant decrease in other inflammatory markers such as cytokines (i.e. IL-6 and TNF-α) [46, 48-50] and markers of coagulation [47], and a significant improvement in markers of endothelial function [47-50]. Therefore, the effect of HRT on CRP is more likely to be a metabolic effect through hepatic induction rather than an inflammatory response [49].

In line with previous studies [16, 22, 23, 30, 31, 51], we found a positive association between smoking and serum levels of CRP. Men smokers had higher average serum CRP compared with women which might be explained by men smoking higher number of cigarettes per day than women. In both men and women, former smokers had levels close to never smokers (especially in women). Moreover, baseline smoking was not a significant predictor of CRP measured on average 13 years later, which indicates the effect of smoking on CRP is likely to be short term rather than long term. Smoking is likely to induce inflammation and a subsequent increase in CRP by causing tissue injury (especially in lungs) and increased susceptibility to infection.

Relatively lower levels of serum CRP in participants with higher education (not seen with social class), especially among women, warrants special attention. This finding was largely independent of BMI, smoking and HRT and thus might be related to diet or physical activity. This finding requires further elaboration in future studies.

We did not find any linear or nonlinear (U-shaped) association between alcohol intake and CRP. Previous cross-sectional studies had inconsistent findings and showed a positive [52], inverse [53] or U-shaped association [54, 55] between alcohol intake and serum CRP levels. However, these studies had a smaller sample size and thus were less powered compared with the present study. Two small RCTs also showed decrease in CRP levels after short-time alcohol consumption [56, 57]. Similar to our findings, the NHANES [22, 23] and the Young Finns study [51] found no association. Our findings thus do not rule out a possible noninflammatory mechanism for the protective effect of moderate alcohol drinking on cardiovascular diseases.

Higher CRP levels in the colder months of the year suggest a higher systemic inflammatory state in the colder compared with the warmer months. Prevalence of acute infections in the colder season cannot fully account for the observed pattern, because the association persisted after excluding subjects with high (> 10 mg/L) CRP levels.

Conclusions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Address
  9. References
  10. Supporting Information

About a third of the UK adult population had CRP levels above 3 mg/L. The distribution of serum CRP in the population is similar in men and women after taking into account smoking and HRT use. Anthropometric factors as well as educational status are strongly related to serum CRP. Anthropometric and lifestyle factors have a short-term rather than long-term effect on serum CRP.

Acknowledgements

We thank all study participants, general practitioners and the EPIC-Norfolk study team for their contribution. The EPIC-Norfolk study is supported by funding from the Medical Research Council and Cancer Research UK with additional support from the Stroke Association, British Heart Foundation, Research Into Ageing and the Academy of Medical Science. SA is supported by the Gates Cambridge scholarship. Funding sources did not have a role in the design and conduct of the study, the collection, management, analysis and interpretation of the data or the preparation, review, approval or decision to submit the manuscript. The authors declare that there is no conflict of interest associated with this manuscript.

SA analysed the data and wrote the manuscript with co-authors. RL performed all data management and record linkage. KTK and NJW are principal investigators in the EPIC-Norfolk population study. All authors provided detailed comments on the manuscript, reviewed and edited the manuscript and contributed to the discussion.

Address

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Address
  9. References
  10. Supporting Information

Department of Public Health and Primary Care, University of Cambridge, Strangeways Research Laboratory, Worts Causeway, Cambridge CB1 8RN, UK (S. Ahmadi-Abhari, R. N. Luben, K.-T. Khaw); MRC Epidemiology Unit, Institute of Metabolic Science, Box 285, Addenbrooke's HospitalHills Road, Cambridge CB2 0QQ, UK (N. J. Wareham).

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Address
  9. References
  10. Supporting Information
  • 1
    Danesh J, Wheeler JG, Hirschfield GM, Eda S, Eiriksdottir G, Rumley A et al. C-reactive protein and other circulating markers of inflammation in the prediction of coronary heart disease. N Engl J Med 2004;350:138797.
  • 2
    Agrawal A. CRP after 2004. Mol Immunol 2005;42:92730.
  • 3
    Kaptoge S, Di AE, Lowe G, Pepys MB, Thompson SG, Collins R et al. C-reactive protein concentration and risk of coronary heart disease, stroke, and mortality: an individual participant meta-analysis. Lancet 2010;375:13240.
  • 4
    Lin CC, Kardia SL, Li CI, Liu CS, Lai MM, Lin WY et al. The relationship of high sensitivity C-reactive protein to percent body fat mass, body mass index, waist-to-hip ratio, and waist circumference in a Taiwanese population. BMC Public Health 2010;10:579.
  • 5
    Wang CS, Sun CF. C-reactive protein and malignancy: clinico-pathological association and therapeutic implication. Chang Gung Med J 2009;32:47182.
  • 6
    Hussain SP, Harris CC. Inflammation and cancer: an ancient link with novel potentials. Int J Cancer 2007;121:237380.
  • 7
    Trichopoulos D, Psaltopoulou T, Orfanos P, Trichopoulou A, Boffetta P. Plasma C-reactive protein and risk of cancer: a prospective study from Greece. Cancer Epidemiol Biomarkers Prev 2006;15:3814.
  • 8
    Il'yasova D, Colbert LH, Harris TB, Newman AB, Bauer DC, Satterfield S et al. Circulating levels of inflammatory markers and cancer risk in the health aging and body composition cohort. Cancer Epidemiol Biomarkers Prev 2005;14:24138.
  • 9
    Siemes C, Visser LE, Coebergh JW, Splinter TA, Witteman JC, Uitterlinden AG et al. C-reactive protein levels, variation in the C-reactive protein gene, and cancer risk: the Rotterdam Study. J Clin Oncol 2006;24:521622.
  • 10
    Chaturvedi AK, Caporaso NE, Katki HA, Wong HL, Chatterjee N, Pine SR et al. C-reactive protein and risk of lung cancer. J Clin Oncol 2010;28:271926.
  • 11
    Allin KH, Bojesen SE, Nordestgaard BG. Baseline C-reactive protein is associated with incident cancer and survival in patients with cancer. J Clin Oncol 2009;27:221724.
  • 12
    Gunter MJ, Stolzenberg-Solomon R, Cross AJ, Leitzmann MF, Weinstein S, Wood RJ et al. A prospective study of serum C-reactive protein and colorectal cancer risk in men. Cancer Res 2006;66:24837.
  • 13
    Erlinger TP, Platz EA, Rifai N, Helzlsouer KJ. C-reactive protein and the risk of incident colorectal cancer. JAMA 2004;291:58590.
  • 14
    Aleksandrova K, Jenab M, Boeing H, Jansen E, Bueno-de-Mesquita HB, Rinaldi S et al. Circulating C-reactive protein concentrations and risks of colon and rectal cancer: a nested case-control study within the European Prospective Investigation into Cancer and Nutrition. Am J Epidemiol 2010;172:40718.
  • 15
    Windgassen EB, Funtowicz L, Lunsford TN, Harris LA, Mulvagh SL. C-reactive protein and high-sensitivity C-reactive protein: an update for clinicians. Postgrad Med 2011;123:1149.
  • 16
    Chen TH, Gona P, Sutherland PA, Benjamin EJ, Wilson PW, Larson MG et al. Long-term C-reactive protein variability and prediction of metabolic risk. Am J Med 2009;122:5361.
  • 17
    Emberson JR, Whincup PH, Morris RW, Walker M, Lowe GD, Rumley A. Extent of regression dilution for established and novel coronary risk factors: results from the British Regional Heart Study. Eur J Cardiovasc Prev Rehabil 2004;11:12534.
  • 18
    Day N, Oakes S, Luben R, Khaw KT, Bingham SWelch A, Wareham N. EPIC-Norfolk: study design and characteristics of the cohort. European Prospective Investigation of Cancer. Br J Cancer 1999;80(Suppl 1):95103.
  • 19
    Shohaimi S, Luben R, Wareham N, Day N, Bingham S, Welch A et al. Residential area deprivation predicts smoking habit independently of individual educational level and occupational social class. A cross sectional study in the Norfolk cohort of the European Investigation into Cancer (EPIC-Norfolk). J Epidemiol Community Health 2003;57:2706.
  • 20
    Simera I, Moher D, Hoey J, Schulz KF, Altman DG. A catalogue of reporting guidelines for health research. Eur J Clin Invest 2010;40:3553.
  • 21
    Woloshin S, Schwartz LM. Distribution of C-reactive protein values in the United States. N Engl J Med 2005;352:16113.
  • 22
    Ford ES, Giles WH, Mokdad AH, Myers GL. Distribution and correlates of C-reactive protein concentrations among adult US women. Clin Chem 2004;50:57481.
  • 23
    Ford ES, Giles WH, Myers GL, Mannino DM. Population distribution of high-sensitivity C-reactive protein among US men: findings from National Health and Nutrition Examination Survey 1999–2000. Clin Chem 2003;49:68690.
  • 24
    Albert MA, Glynn RJ, Buring J, Ridker PM. C-reactive protein levels among women of various ethnic groups living in the United States (from the Women's Health Study). Am J Cardiol 2004;93:123842.
  • 25
    Imhof A, Frohlich M, Loewel H, Helbecque N, Woodward M, Amouyel P et al. Distributions of C-reactive protein measured by high-sensitivity assays in apparently healthy men and women from different populations in Europe. Clin Chem 2003;49:66972.
  • 26
    Yamada S, Gotoh T, Nakashima Y, Kayaba K, Ishikawa S, Nago N et al. Distribution of serum C-reactive protein and its association with atherosclerotic risk factors in a Japanese population: Jichi Medical School Cohort Study. Am J Epidemiol 2001;153:118390.
  • 27
    Zhao Y, Wang R, Ma X, Yan X, Zhang Z, He X et al. Distribution of C-reactive protein and its association with cardiovascular risk factors in a population-based sample of Chinese. Dis Markers 2010;28:33342.
  • 28
    Albert MA, Danielson E, Rifai N, Ridker PM. Effect of statin therapy on C-reactive protein levels: the pravastatin inflammation/CRP evaluation (PRINCE): a randomized trial and cohort study. JAMA 2001;286:6470.
  • 29
    Grace FM, Davies B. Raised concentrations of C reactive protein in anabolic steroid using bodybuilders. Br J Sports Med 2004;38:978.
  • 30
    Arcari A, Zito F, Di CA, De CA, Dirckx C, Arnout J et al. C reactive protein and its determinants in healthy men and women from European regions at different risk of coronary disease: the IMMIDIET Project. J Thromb Haemost 2008;6:43643.
  • 31
    Lee YJ, Lee JH, Shin YH, Kim JK, Lee HR, Lee DC. Gender difference and determinants of C-reactive protein level in Korean adults. Clin Chem Lab Med 2009;47:8639.
  • 32
    Camhi SM, Bray GA, Bouchard C, Greenway FL, Johnson WD, Newton RL et al. The relationship of waist circumference and BMI to visceral, subcutaneous, and total body fat: sex and race differences. Obesity 2011;19:4028.
  • 33
    Kuk JL, Lee S, Heymsfield SB, Ross R. Waist circumference and abdominal adipose tissue distribution: influence of age and sex. Am J Clin Nutr 2005;81:s13304.
  • 34
    Maki KC, Rains TM, Bell M, Reeves MS, Farmer MV, Yasunaga K. Fat mass, abdominal fat distribution, and C-reactive protein concentrations in overweight and obese men and women. Metab Syndr Relat Disord 2011;9:2916.
  • 35
    Thorand B, Baumert J, Doring A, Herder C, Kolb H, Rathmann W et al. Sex differences in the relation of body composition to markers of inflammation. Atherosclerosis 2006;184:21624.
  • 36
    Festa A, D'Agostino R Jr, Williams K, Karter AJ, Mayer-Davis EJ, Tracy RP et al. The relation of body fat mass and distribution to markers of chronic inflammation. Int J Obes Relat Metab Disord 2001;25:140715.
  • 37
    Lapice E, Maione S, Patti L, Cipriano P, Rivellese AA, Riccardi G et al. Abdominal adiposity is associated with elevated C-reactive protein independent of BMI in healthy nonobese people. Diabetes Care 2009;32:17346.
  • 38
    Visser M, Bouter LM, McQuillan GM, Wener MH, Harris TB. Elevated C-reactive protein levels in overweight and obese adults. JAMA 1999;282:21315.
  • 39
    Flegal KM, Shepherd JA, Looker AC, Graubard BI, Borrud LG, Ogden CL et al. Comparisons of percentage body fat, body mass index, waist circumference, and waist-stature ratio in adults. Am J Clin Nutr 2009;89:5008.
  • 40
    Pou KM, Massaro JM, Hoffmann U, Vasan RS, Maurovich-Horvat P, Larson MG et al. Visceral and subcutaneous adipose tissue volumes are cross-sectionally related to markers of inflammation and oxidative stress: the Framingham Heart Study. Circulation 2007;116:123441.
  • 41
    Selvin E, Paynter NP, Erlinger TP. The effect of weight loss on C-reactive protein: a systematic review. Arch Intern Med 2007;167:319.
  • 42
    Brooks GC, Blaha MJ, Blumenthal RS. Relation of C-reactive protein to abdominal adiposity. Am J Cardiol 2010;106:5661.
  • 43
    Yudkin JS, Stehouwer CD, Emeis JJ, Coppack SW. C-reactive protein in healthy subjects: associations with obesity, insulin resistance, and endothelial dysfunction: a potential role for cytokines originating from adipose tissue? Arterioscler Thromb Vasc Biol 1999;19:9728.
  • 44
    Kwok S, Charlton-Menys V, Pemberton P, McElduff P, Durrington PN. Effects of dydrogesterone and norethisterone, in combination with oestradiol, on lipoproteins and inflammatory markers in postmenopausal women. Maturitas 2006;53:43946.
  • 45
    Walsh BW, Paul S, Wild RA, Dean RA, Tracy RP, Cox DA et al. The effects of hormone replacement therapy and raloxifene on C-reactive protein and homocysteine in healthy postmenopausal women: a randomized, controlled trial. J Clin Endocrinol Metab 2000;85:2148.
  • 46
    Sezer OK, Erenus M, Yoldemir T. The impact of tibolone and hormone therapy on serum C-reactive protein, tumor necrosis factor-alpha and hepatocyte growth factor in postmenopausal women. Climacteric 2009;12:6671.
  • 47
    van Baal WM, Kenemans P, Emeis JJ, Schalkwijk CG, Mijatovic V, van der Mooren MJ et al. Long-term effects of combined hormone replacement therapy on markers of endothelial function and inflammatory activity in healthy postmenopausal women. Fertil Steril 1999;71:66370.
  • 48
    Koh KK, Ahn JY, Jin DK, Yoon BK, Kim HS, Kim DS et al. Effects of continuous combined hormone replacement therapy on inflammation in hypertensive and/or overweight postmenopausal women. Arterioscler Thromb Vasc Biol 2002;22:145964.
  • 49
    Silvestri A, Gebara O, Vitale C, Wajngarten M, Leonardo F, Ramires JA et al. Increased levels of C-reactive protein after oral hormone replacement therapy may not be related to an increased inflammatory response. Circulation 2003;107:31659.
  • 50
    Eilertsen AL, Hoibraaten E, Os I, Andersen TO, Sandvik L, Sandset PM. The effects of oral and transdermal hormone replacement therapy on C-reactive protein levels and other inflammatory markers in women with high risk of thrombosis. Maturitas 2005;52:1118.
  • 51
    Raitakari M, Mansikkaniemi K, Marniemi J, Viikari JS, Raitakari OT. Distribution and determinants of serum high-sensitive C-reactive protein in a population of young adults: the Cardiovascular Risk in Young Finns Study. J Intern Med 2005;258:42834.
  • 52
    Volpato S, Pahor M, Ferrucci L, Simonsick EM, Guralnik JM, Kritchevsky SB et al. Relationship of alcohol intake with inflammatory markers and plasminogen activator inhibitor-1 in well-functioning older adults: the Health, Aging, and Body Composition study. Circulation 2004;109:60712.
  • 53
    Albert MA, Glynn RJ, Ridker PM. Alcohol consumption and plasma concentration of C-reactive protein. Circulation 2003;107:4437.
  • 54
    Imhof A, Woodward M, Doering A, Helbecque N, Loewel H, Amouyel P et al. Overall alcohol intake, beer, wine, and systemic markers of inflammation in western Europe: results from three MONICA samples (Augsburg, Glasgow, Lille). Eur Heart J 2004;25:2092100.
  • 55
    Imhof A, Froehlich M, Brenner H, Boeing H, Pepys MB, Koenig W. Effect of alcohol consumption on systemic markers of inflammation. Lancet 2001;357:7637.
  • 56
    Estruch R, Sacanella E, Badia E, Antunez E, Nicolas JM, Fernandez-Sola J et al. Different effects of red wine and gin consumption on inflammatory biomarkers of atherosclerosis: a prospective randomized crossover trial. Effects of wine on inflammatory markers. Atherosclerosis 2004;175:11723.
  • 57
    Sierksma A, van der Gaag MS, Kluft C, Hendriks HF. Moderate alcohol consumption reduces plasma C-reactive protein and fibrinogen levels; a randomized, diet-controlled intervention study. Eur J Clin Nutr 2002;56:11306.

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Address
  9. References
  10. Supporting Information
FilenameFormatSizeDescription
eci12116-sup-0001-Table.docWord document60KTable S1. Percent difference in serum C-reactive protein measured at third health examination per defined unit difference in determinants measured cross-sectionally or at baseline (13 years apart).

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.