You are here

  • Home
  • Archive
  • Volume 69, Issue 11
  • Serum C reactive protein levels and genetic variation in the CRP gene are not associated with the prevalence, incidence or progression of osteoarthritis independent of body mass index

Article Text

Article menu

Download PDFPDF

Clinical and epidemiological research
Extended report
Serum C reactive protein levels and genetic variation in the CRP gene are not associated with the prevalence, incidence or progression of osteoarthritis independent of body mass index
  1. Hanneke J M Kerkhof1,2,
  2. Sita M A Bierma-Zeinstra3,
  3. Martha C Castano-Betancourt1,
  4. Moniek P de Maat4,
  5. Albert Hofman2,5,
  6. Huib A P Pols1,
  7. Fernando Rivadeneira1,2,
  8. Jacqueline C Witteman5,
  9. André G Uitterlinden1,2,5,
  10. Joyce B J van Meurs1,2
  1. 1Department of Internal Medicine, Erasmus Medical Center Rotterdam, The Netherlands
  2. 2The Netherlands Genomics Initiative-sponsored Netherlands Consortium for Healthy Aging (NGI-NCHA), Rotterdam, The Netherlands
  3. 3Department of General Practice, Erasmus Medical Center Rotterdam, The Netherlands
  4. 4Department of Hematology, Erasmus Medical Center Rotterdam, The Netherlands
  5. 5Department of Epidemiology, Erasmus Medical Center Rotterdam, The Netherlands
  1. Correspondence to Dr Joyce B J van Meurs, Genetic Laboratory, Department of Internal Medicine, Room Ee579, Erasmus Medical Center, MC PO Box 1738, 3000 DR Rotterdam, The Netherlands; j.vanmeurs{at}erasmusmc.nl

Abstract

Objective To study the relationship between serum C reactive protein (CRP) levels, genetic variation in the CRP gene and the prevalence, incidence and progression of radiographic osteoarthritis (ROA) in the Rotterdam Study-I (RS-I). A systematic review of studies assessing the relationship between osteoarthritis (OA) and CRP levels was also performed.

Methods The association between CRP levels and genetic variation in the CRP gene and ROA was examined in 861 patients with hand OA, 718 with knee OA, 349 with hip OA and 2806 controls in the RS-I using one-way analysis of covariance and logistic regression, respectively. PubMed was searched for articles published between January 1992 and August 2009 assessing the relationship between CRP levels and OA.

Results In RS-I the prevalence of knee OA, but not hip OA or hand OA, was associated with 14% higher serum CRP levels compared with controls (p=0.001). This association disappeared after adjustment for age and especially body mass index (BMI) (p=0.33). Genetic variation of the CRP gene was not consistently associated with the prevalence, incidence or progression of OA within RS-I. The systematic review included 18 studies (including RS-I) on serum CRP levels and the prevalence, incidence or progression of OA. Consistently higher crude CRP levels were found in cases of prevalent knee OA compared with controls. No association was observed between serum CRP levels and the prevalence of knee OA following adjustment for BMI (n=3 studies, meta-analysis p value=0.61).

Conclusion There is no evidence of association between serum CRP levels or genetic variation in the CRP gene with the prevalence, incidence or progression of OA independent of BMI.

https://doi.org/10.1136/ard.2009.125260

Statistics from Altmetric.com

    Osteoarthritis (OA) is characterised by cartilage destruction, bone remodelling and synovitis. Although OA is not a classical inflammatory arthropathy like rheumatoid arthritis, there are indications that local inflammation is present in OA. It is known from in vivo and in vitro studies that chondrocytes can respond to chemokines and cytokines in the synovial fluid and joint tissues.1 In addition, patients with OA have increased levels of catabolic enzymes and inflammatory mediators such as interleukin 1β and tumour necrosis factor α in the synovial fluid.1 For these reasons, it is believed that local inflammatory processes might have a role in the pathogenesis of OA.

    C reactive protein (CRP) is an acute phase protein which is produced by hepatocytes and adipocytes and regulated by proinflammatory cytokines.2 3 CRP is an established marker for severe systemic infections like sepsis.4 Previous studies have shown that CRP levels are associated with age and with measures of obesity and cardiovascular disease.5,–,7

    Several studies have investigated serum CRP levels in relation to OA to determine if systemic inflammation plays a role in the pathogenesis of OA. In 1975, Acheson and Collart were the first to show that elevated serum CRP levels are associated with OA.8 In 1992 an immunoassay was developed to perform high sensitive C reactive protein (hs-CRP) measurements, which is accurate in detecting low levels of CRP.9 10 The first study using hs-CRP measurements showed that patients with OA had higher serum CRP levels than age-matched controls.11 Since then, several studies (n=17) have been carried out assessing the relationship between the prevalence, incidence and/or progression of OA and CRP levels, but with conflicting results.2 12,–,27

    To examine the causal effect of a variable on disease (ie, CRP levels on OA), the method of Mendelian randomisation could be used.28 It is a method of testing for a causal relationship between CRP and OA. For example, if genetic variation in the CRP gene is associated with higher CRP levels and the same variation is also related to an increased risk for OA, this suggests that CRP is involved in OA causally. So far, only one study has investigated the role of common genetic variation in the CRP gene in relation to OA.12 This study showed that genetic variation in the CRP gene (haplotype AAGGA, frequency 5%) was associated with higher hs-CRP levels and with an increased risk of severe hand OA (OR 2.3, 95% CI 1.2 to 4.3, p=0.009) in 42 cases of severe hand OA and 70 controls.

    We report data from a large population-based cohort study, the Rotterdam Study-I (RS-I), assessing the relationship between serum CRP levels and radiographic osteoarthritis (ROA). We also investigated whether common genetic variation in the CRP gene, known to be associated with CRP levels,29 is associated with the prevalence, incidence and/or progression of OA. In addition, we give an overview of all studies published on the relationship between serum hs-CRP levels and the prevalence, incidence and/or progression of OA. We performed a meta-analysis on the relationship between CRP levels and OA.

    Patients and methods

    Association analyses of the Rotterdam Study-I

    The Rotterdam Study-I

    The study population comprises men and women aged ≥55 years from the Rotterdam Study-I (RS-I), which is a prospective population-based study on determinants of chronic disabling diseases. The rationale and study design have been described previously.30 At baseline there were 349 cases of hip OA and 3065 controls, 718 cases of knee OA and 2306 controls, 861 cases of hand OA and 2164 controls. For progression of hip OA there were 461 cases and 1149 controls, for progression of knee OA 208 cases and 627 controls, for incidence of hip OA 220 cases and 2519 controls and for incidence of knee OA 210 cases and 1360 controls.

    OA phenotypes in the RS-I

    Radiographs were scored for the presence of a total joint replacement (TJR) and ROA of the hip, knee and hand according to the Kellgren and Lawrence (K/L) score. Knee and hip OA were defined as a K/L score ≥2 of one or both joints or a TJR. Hand OA was defined as a K/L score ≥2 in two out of three hand joint groups (distal interphalangeals, proximal interphalangeals, first carpometacarpal/trapezioscaphoid) of each or both hands.31 The incidence of knee and/or hip OA was defined as a K/L <2 at baseline and TJR or K/L ≥2 at follow-up. The progression of hip and knee OA was determined in subjects with a K/L score ≥1 at baseline. Progression of hip OA was defined as a TJR or joint space narrowing ≥1.0 mm during follow-up.32 33 Progression of knee OA was defined as a TJR or an increase in K/L score ≥1 during follow-up. Subjects with a TJR due to fracture were excluded from all analyses. Controls were free of OA at the joint site studied but were allowed to have OA at other joint sites. For example, if knee OA was studied, controls had to be free of knee OA but were allowed to have hip or hand OA. One subject could be considered a case for multiple phenotypes. For example, one subject can be a hip OA case and a knee OA case.

    Pain assessment

    At baseline, hip, knee and hand symptoms in the previous month were assessed by interview. Hip, knee or hand pain was defined as pain in the right and/or left joint in the month preceding the interview.

    CRP measurement

    At baseline (1990–3) and follow-up (1996–9), blood was drawn by venous puncture, initially stored at –20°C and thawed and assayed for hs-CRP using Rate Near Infrared Particle Immunoassay (Immage Immunochemistry System; Beckman Coulter, Brea, California, USA). This method can accurately measure protein concentrations from 0.2 to 1440 mg/l with a within-run precision <5.0%, a total precision <7.5% and a reliability coefficient of 0.995.34

    Genotyping methods

    Genomic DNA was extracted from peripheral venous blood samples according to standard procedures using salting out and phenol extraction methods. Genotypes were determined using the Taqman allelic discrimination assay. The Assay-by-Design service (www.appliedbiosystems.com) was used to set up a Taqman allelic discrimination assay for three tagging single nucleotide polymorphisms (SNPs). The primers and probes used are available on request and conditions of the assay were as previously described.35 The selection of the three haplotype tagging SNPs using the Seattle SNPs Program for Genomic Applications (http://pga.gs.washington.edu/) has been described elsewhere in detail.34 In short, the SNPs rs1130864(C>T), rs1205(C>T) and rs3093068(C>G) were selected to estimate haplotypes on the basis of their presence in existing literature and on their proximity to the CRP gene. We used genotype data of each of the three SNPs to infer frequency of the haplotype alleles present in the population using the program PHASE Version 2.1.36 Haplotypes with an estimated probability <95% were excluded from the analyses. In total, four haplotypes were constructed and were coded as numbers 1–4 in order of descending frequency in the population: haplotype 1=CTC, 2=TCC, 3=CCC, 4=CCG.

    Statistical analyses

    Analyses were carried out only in subjects with a CRP level <10 mg/l, thereby excluding 311 subjects with possible acute inflammatory conditions. All analyses were performed using natural log-transformed hs-CRP levels since the distribution was not normal in the population. After transformation the distribution of CRP levels in the population was normal (data not shown). The CRP levels in the Results section and the tables are shown as means. The associations between baseline characteristics of RS-I and OA phenotypes were assessed using one-way analysis of covariance (AN(C)OVA). Association analyses of CRP levels with hip, knee and hand OA (prevalence, incidence, progression) with pain in OA cases and with haplotypes were performed using one-way AN(C)OVA and, if applicable, linear regression. In addition, to detect possible threshold effects, sex-specific groups were created using tertiles and the median as cut-off for CRP levels. One-way AN(C)OVA was also used to assess the relationship between delta CRP levels (log-transformed) and the incidence or progression of knee or hip OA (mean follow-up time 6.3 years). A logistic regression was performed to assess the relationship between CRP tertiles/median and OA. This logistic regression was subsequently adjusted for age and body mass index (BMI) and sex-specific age and BMI-adjusted tertile/median groups were created to perform a logistic regression fully adjusted for gender, age and BMI. To study the relationship between CRP haplotypes and OA, a logistic regression model was applied. All analyses were performed crude and adjusted for gender, age and BMI. In addition, the association analyses for incidence/progression of OA were adjusted for follow-up time. Also, adjustment for usage of analgesics and/or anti-inflammatory drugs and/or coronary heart disease (defined as myocardial infarction, coronary artery bypass grafting or percutaneous transluminal coronary angioplasty) was performed but is only reported in the Results section if this adjustment changed the results. To adjust for multiple testing we performed a Bonferroni correction after which the threshold for statistical significance was set at 8×10−4 (0.05/60 tests). For the individual SNPs, allele frequencies were estimated by allele counting and Hardy–Weinberg equilibrium (HWE) was tested using Haploview 4.0. All analyses were performed using SPSS 15.0 for Windows.

    Literature study

    Identification of studies

    Relevant articles were identified by a systematic search using the database of PubMed with (‘CRP’ or ‘C-reactive protein’) and (‘OA’ or ‘osteoarthritis’) as keywords. The search was subsequently extended by screening reference lists of the included studies for review. The following inclusion criteria applied to this review: (1) listed in PubMed; (2) publication date between January 1992 and August 2009; (3) publication in the English language; (4) study in humans; (5) the article represents original data; (6) the disease of interest is OA; (7) subjects with and without OA are compared in the study; (8) the study reports hs-CRP measurements in serum; (9) the study investigates the relationship between OA and CRP levels presenting ORs, difference in median or mean; (10) the full-text article was available.

    Methodological quality assessment

    All studies were scored for the presence of the following criteria: (1) information on recruitment of cases; (2) information on recruitment of controls; (3) size of the study (n >100 cases or controls); (4) information for all subjects on age, gender and BMI; (5) description of CRP measurement; (6) OA definition according to American College of Rheumatology criteria, a TJR due to primary OA, the K/L score or Croft classification system; (7) clear description of statistical methods; (8) CRP levels were log-transformed if not following a normal distribution; (9) adjustments were made for age, gender and BMI in the analyses; (10) information on exclusion of subjects with a state of acute inflammation (ie, subjects with CRP >20 mg/l); (11) results are presented as either mean, tertiles or median with p values or ORs with 95% confidence limits. Positive scores were summed to indicate an overall internal validity score with a maximum of 100% if all items were scored positive. A study was considered to be of high quality if the methodological score was ≥60%.37

    Meta-analysis

    For the meta-analysis, the programme comprehensive meta-analysis by Biostat (www.meta-analysis.com) was used. If heterogeneity existed (I2 >25%) a random effects model (DerSimonian and Laird) was used for the analysis, otherwise a fixed effects model (inverse variance method) was applied.

    Results

    Results of the RS-I

    Serum CRP levels and OA

    The baseline characteristics of the RS-I cohort are shown in table 1. Overall, women had OA more frequently (3–18%) than men (with the exception of prevalent hip OA), they were 2–4 years older (with the exception of incident or progressive knee OA) and had 1–2 units higher BMI (with the exception of prevalent hip OA).

    View this table:
    Table 1

    Baseline characteristics of patients and controls in the Rotterdam Study-I

    Table 2 shows the results of the association of CRP levels with the prevalence, incidence and progression of OA in RS-I. Subjects with prevalent knee OA had 14% higher CRP levels than control subjects (p=0.001). After additional adjustment for age and gender, the association remained significant (p=0.01). After adjustment for BMI, the association disappeared (p=0.33). Of all association analyses performed on CRP levels and the incidence or progression of hip or knee OA and on the prevalence of hip or hand OA, only the incidence of hip OA was associated with 13% higher CRP levels (p=0.04 adjusted for age, gender, BMI and follow-up time), but after adjustment for multiple testing this result did not remain significant (p<8×10−4 was considered statistically significant after Bonferroni correction). In addition, when the prevalence, incidence and progression of OA were studied across CRP tertiles and median to reveal a possible threshold effect, we did not observe any differences (data not shown).

    View this table:
    Table 2

    Association between C reactive protein (CRP) levels and knee, hip and hand osteoarthritis (OA) in the Rotterdam Study-I

    Longitudinal change in serum CRP and OA

    Change in serum CRP levels between baseline measurements and 5 years of follow-up was not associated with the incidence or progression of knee or hip OA (see table 1 in online supplement). Additional adjustment for usage of anti-inflammatory drugs or pain killers did not change the results (data not shown).

    Serum CRP levels and pain in OA cases

    Subjects with ROA with pain did not have higher or lower CRP levels than subjects with ROA without pain. These results did not change after adjustment for age, gender, BMI, usage of pain killers and/or anti-inflammatory drugs. Subjects with hip OA had 4% higher CRP levels if they experienced pain than subjects without pain (p=0.84); for knee OA the percentage was 8% (p=0.25) and for hand OA 3% (p=0.69) (figure 1).

    Figure 1
    Figure 1

    Association between high sensitive C reactive protein (hs-CRP) levels and pain in subjects with radiographic osteoarthritis (OA) adjusted for gender, age and body mass index.

    Genetic variation in the CRP gene and OA

    Genotype distributions of the three haplotype tagging SNPs were in HWE. Haplotype alleles were present with the following frequencies: haplotype 1 (CTC) in 32.8%; haplotype 2 (TCC) in 31.7%; haplotype 3 (CCC) in 29.5% and haplotype 4 (CCG) in 5.9%.

    Genetic variation of the CRP gene was strongly and consistently associated with serum CRP levels,34 but no consistent associations were observed with the prevalence, incidence or progression of OA (data not shown).

    Results of the literature study

    A total of 225 studies were identified in the database, of which 203 were excluded on the basis of the title and/or abstract. Of the 22 studies retrieved for full examination, 5 were excluded because they did not fulfil the inclusion criteria after reading the complete manuscript. Finally, 18 studies (17 from the literature search + results from RS-I) were included in the review. The flow chart of the selection process is shown in figure 2.

    Figure 2
    Figure 2

    Flow chart of the selection process. hs-CRP, high sensitive C reactive protein; OA, osteoarthritis.

    Methodological quality assessment

    The 11 aspects of quality of the studies mentioned in the Methods section are shown in table 2 in the online supplement. There were two studies without complete information on recruitment of both cases and controls.21 22 For example, the criteria were given for subjects to be considered a case, but the source population of recruitment and the method of recruitment were not given. Many studies (n=10) were small and included fewer than 100 cases or controls. All studies provided detailed information on the statistical methods used within their study. Five of the 18 studies performed adjustments for age, gender and BMI (RS-I and 4 published studies15 19 20 24). A study was considered of high quality if the methodological score was ≥60%, and this applied to 13 of the 18 studies (72%). Five studies were of less quality with a methodological score ranging from 36% to 55%.2 18 21,–,23

    Association between CRP and prevalence of OA

    The characteristics of the 14 studies assessing the relationship between serum CRP levels and the prevalence of OA are shown in table 3 in the online supplement (including RS-I).2 12,–,14 17,–,19 21,–,23 25,–,27 Combining all cases and controls, there were a total of 3990 OA cases and 6566 controls. In table 3 the results showing the relationship between CRP levels and prevalent OA of these 14 studies are shown. Mean CRP levels in OA cases range from 1.6 to 9.8 mg/l and in controls from 0.9 to 7.7 mg/l. Statistically significant results (p<0.05) were seen in 10 of the 14 studies and showed that subjects with OA (hip, knee or hand) have higher levels of CRP than controls (RS-I and 9 published studies13 14 18 19 22 23 25,–,27). Only a few studies were performed for hip and hand OA, which prevented us from drawing firm conclusions. For knee OA, 11 studies were performed. The results indicate a positive relation between crude serum CRP levels and knee OA, since only three small studies did not show a significant association (p>0.05) while all the other studies (including the two largest studies: RS-I and one published study19) showed a positive association between CRP levels and knee OA unadjusted for BMI. A meta-analysis could not be performed on the relationship between crude CRP levels and knee, hip or hand OA since 9 of the 14 studies reported inexact p values or geometric means were not provided.

    View this table:
    Table 3

    Results of the included studies on CRP levels and prevalence of OA

    A meta-analysis assessing the relationship between CRP levels adjusted for BMI and hand or hip OA could not be performed since as this information was only available in one study (RS-I). We did perform a meta-analysis on knee OA and CRP levels adjusted for BMI, which included three studies with a total of 1171 cases and 2649 controls (RS-I and two published studies17 19). The meta-analysis did not provide evidence of an association between serum CRP levels and knee OA independent of BMI (random-effects p=0.61, I2=54%; see figure 1 in online supplement).

    Association between CRP and incidence or progression of OA

    The review included seven studies which investigated the relation between CRP levels and incidence and/or progression of OA including the results of RS-I (see table 4 in online supplement).15 16 20 24,–,26 The results of the studies on CRP levels and incidence or progression of OA are shown in table 4. A meta-analysis could not be performed for these studies since the results of the studies are shown in different ways (difference in mean or ORs using CRP levels as continuous variable or with a threshold of 3 mg/l). In addition, different OA outcomes were assessed (progression of hip or knee OA and incidence of hip or knee OA). Five of the seven studies reported statistically significant differences in crude CRP levels in relation to the incidence or progression of OA (RS-I and four published studies15 24,–,26), of which only two studies adjusted for age and BMI (RS-I).24 The study by Sharif et al on CRP levels and progression of knee OA showed higher CRP levels in subjects with progression of knee OA compared with controls (OR 1.90, 95% CI 1.01 to 3.28, p=0.04).24 This was not supported by the six times larger dataset of RS-I which showed slightly lower levels of CRP in subjects with progression of knee OA compared with controls after adjustment for gender, age, BMI and follow-up time (p=0.61). In RS-I, subjects with hip OA had 13% higher serum CRP levels than controls (p=0.04 after adjustment for age, gender, BMI and follow-up time), but this was not supported by the study by Engström et al which showed an OR of 1.1 (95% CI 0.6 to 1.9) for subjects with a CRP level ≥3 mg/l compared with subjects with CRP <1 mg/l.16

    View this table:
    Table 4

    Results of the included studies on CRP levels and incidence/progression of OA

    Discussion

    In the RS-I, the largest study published to date on serum CRP levels and OA, we showed an association between knee OA and serum CRP levels similar to previously published studies, but this association was fully driven by BMI.2 18 19 22 23 25,–,27 In addition, a meta-analysis including three studies on serum CRP levels adjusted for BMI and knee OA did not show evidence of an association (p=0.61). Although we could not perform a meta-analysis on the crude CRP levels in relation to knee OA, a consistently positive association between CRP levels and knee OA was seen in 8 out of 11 studies. However, this association is mediated through BMI and highlights the complex interaction between BMI and/or body fat, CRP and OA, as is also discussed in a paper by Engström et al.16

    Within RS-I an association between CRP levels and the incidence of hip OA (p=0.04) was observed, which remained significant after adjustment for age, gender, BMI and follow-up time. However, this association did not remain significant after Bonferroni correction and was also not observed in the study by Engström et al.16 We therefore believe that this result is a false positive due to multiple testing. In addition, no association was observed between CRP levels and pain or between change in CRP levels and incidence or progression of knee or hip OA during 5 years of follow-up. This result provides further evidence that serum hs-CRP is not a marker for acute inflammation in OA and that acute systemic inflammation is not present in OA.

    One of the limitations of this study is the fact that only three studies on knee OA and serum CRP levels adjusted for age and BMI were included in the meta-analysis. Although these three studies did not provide evidence of an association between serum CRP levels and knee OA independent of BMI, a larger meta-analysis might strengthen this conclusion even more.

    The presence of pain is assessed in the RS-I by a questionnaire. More precise measurement for pain with, for example, the Western Ontario and McMasters Universities Arthritis Index (WOMAC) pain score or a visual analogue scale was not performed. The conclusions concerning the analysis of CRP levels and pain in cases of ROA should therefore be taken with caution.

    Genetic analysis in RS-I showed that genetic variation in the CRP gene was not consistently associated with CRP levels. One can identify a causal relationship in observational studies using the method of Mendelian randomisation.28 If common genetic variation in a gene is associated with the intermediate phenotype—which, in this study, is CRP levels—and genetic variation is also associated with OA in the same direction as the association between the genetic variation and CRP levels, it can be assumed that there is a causal relationship between CRP levels and OA. In this study we did not observe evidence of such a causal relationship between CRP levels and OA.

    There was a large difference in mean CRP levels measured in the different studies. In control subjects the mean CRP levels ranged from 0.9 to 7.7 mg/l and in cases from 1.6 to 9.8 mg/l. This could be due to different techniques used to measure CRP or the fact that, in some studies, subjects with CRP levels above a certain level were excluded as this might be an indication of an acute inflammatory disease while other studies did not make such exclusions.

    Although CRP is often studied in relation to OA, we cannot provide evidence of a robust association independent of BMI. A recent paper showed that, for most study designs and settings, it is more likely for a research claim to be false than true.38 It was also argued that studies are more likely to be false when the studies conducted are smaller, when effect sizes are smaller and when there is great flexibility in definitions and analytical methods. For all studies performed on CRP and OA, at least some of these characteristics apply and, in our opinion, the message of this paper could also be applied to the studies conducted on serum CRP levels and OA.

    In conclusion, the results of RS-I, the largest cohort study to date to study the relationship between serum CRP levels and genetic variation in the CRP gene in relation to the prevalence, incidence and progression of OA, do not provide evidence of an association between CRP levels and the prevalence, incidence or progression of OA independent of BMI. These results are corroborated by the systematic review in which no evidence was found to support an independent effect of CRP levels on OA. In our opinion, future studies should focus more on local inflammation processes involved in OA than on systemic inflammation.

    Acknowledgments

    The authors thank all the participants in the Rotterdam study and the general practitioners, pharmacists and many field workers at the research centre in Ommoord, Rotterdam, The Netherlands. The authors are very grateful to Dr E Odding, Dr A P Bergink, Dr M Reijman and Dr S Dahaghin for scoring the knee, hip and hand radiographs.

    References

    1. ↵
      1. Goldring MB,
      2. Goldring SR
      . Osteoarthritis. J Cell Physiol 2007;213:626–34.
      OpenUrlCrossRefPubMedWeb of Science
    2. ↵
      1. Pearle AD,
      2. Scanzello CR,
      3. George S,
      4. et al
      . Elevated high-sensitivity C-reactive protein levels are associated with local inflammatory findings in patients with osteoarthritis. Osteoarthr Cartil 2007;15:516–23.
      OpenUrlCrossRefPubMedWeb of Science
    3. ↵
      1. Calabro P,
      2. Chang DW,
      3. Willerson JT,
      4. et al
      . Release of C-reactive protein in response to inflammatory cytokines by human adipocytes: linking obesity to vascular inflammation. J Am Coll Cardiol 2005;46:1112–13.
      OpenUrlCrossRefPubMedWeb of Science
    4. ↵
      1. Keshet R,
      2. Boursi B,
      3. Maoz R,
      4. et al
      . Diagnostic and prognostic significance of serum C-reactive protein levels in patients admitted to the department of medicine. Am J Med Sci 2009;337:248–55.
      OpenUrlCrossRefPubMedWeb of Science
    5. ↵
      1. Hak AE,
      2. Stehouwer CD,
      3. Bots ML,
      4. et al
      . Associations of C-reactive protein with measures of obesity, insulin resistance, and subclinical atherosclerosis in healthy, middle-aged women. Arterioscler Thromb Vasc Biol 1999;19:1986–91.
      OpenUrlAbstract/FREE Full Text
    6. ↵
      1. Van Der Meer IM,
      2. De Maat MP,
      3. Hak AE,
      4. et al
      . C-reactive protein predicts progression of atherosclerosis measured at various sites in the arterial tree: the Rotterdam Study. Stroke 2002;33:2750–5.
      OpenUrlAbstract/FREE Full Text
    7. ↵
      1. Visser M,
      2. Bouter LM,
      3. McQuillan GM,
      4. et al
      . Elevated C-reactive protein levels in overweight and obese adults. JAMA 1999;282:2131–5.
      OpenUrlCrossRefPubMedWeb of Science
    8. ↵
      1. Acheson RM,
      2. Collart AB
      . New Haven survey of joint diseases. XVII. Relationship between some systemic characteristics and osteoarthrosis in a general population. Ann Rheum Dis 1975;34:379–87.
      OpenUrlAbstract/FREE Full Text
    9. ↵
      1. Montagne P,
      2. Laroche P,
      3. Cuillière ML,
      4. et al
      . Microparticle-enhanced nephelometric immunoassay for human C-reactive protein. J Clin Lab Anal 1992;6:24–9.
      OpenUrlCrossRefPubMedWeb of Science
    10. ↵
      1. Montagne P,
      2. Laroche P,
      3. Bessou T,
      4. et al
      . Measurement of eleven serum proteins by microparticle-enhanced nephelometric immunoassay. Eur J Clin Chem Clin Biochem 1992;30:217–22.
      OpenUrlPubMedWeb of Science
    11. ↵
      1. Loose LD,
      2. de Oliviera R,
      3. Sipe JD,
      4. et al
      . A possible systemic component of osteoarthritis: elevated concentrations (by ELISA) of C reactive protein in serum of OA patients and modulation by tenidap [abstract]. Arthritis Rheum 1996;39:S166.
      OpenUrl
    12. ↵
      1. Bos SD,
      2. Suchiman HE,
      3. Kloppenburg M,
      4. et al
      . Allelic variation at the C-reactive protein gene associates to both hand osteoarthritis severity and serum high sensitive C-reactive protein levels in the GARP study. Ann Rheum Dis 2008;67:877–9.
      OpenUrlAbstract/FREE Full Text
    13. ↵
      1. Chen HC,
      2. Shah S,
      3. Stabler TV,
      4. et al
      . Biomarkers associated with clinical phenotypes of hand osteoarthritis in a large multigenerational family: the CARRIAGE family study. Osteoarthr Cartil 2008;16:1054–9.
      OpenUrlCrossRefPubMedWeb of Science
    14. ↵
      1. Conrozier T,
      2. Carlier MC,
      3. Mathieu P,
      4. et al
      . Serum levels of YKL-40 and C reactive protein in patients with hip osteoarthritis and healthy subjects: a cross sectional study. Ann Rheum Dis 2000;59:828–31.
      OpenUrlAbstract/FREE Full Text
    15. ↵
      1. Conrozier T,
      2. Chappuis-Cellier C,
      3. Richard M,
      4. et al
      . Increased serum C-reactive protein levels by immunonephelometry in patients with rapidly destructive hip osteoarthritis. Rev Rhum Engl Ed 1998;65:759–65.
      OpenUrlPubMed
    16. ↵
      1. Engström G,
      2. Gerhardsson de Verdier M,
      3. Rollof J,
      4. et al
      . C-reactive protein, metabolic syndrome and incidence of severe hip and knee osteoarthritis. A population-based cohort study. Osteoarthr Cartil 2009;17:168–73.
      OpenUrlCrossRefPubMedWeb of Science
    17. ↵
      1. Garnero P,
      2. Piperno M,
      3. Gineyts E,
      4. et al
      . Cross sectional evaluation of biochemical markers of bone, cartilage, and synovial tissue metabolism in patients with knee osteoarthritis: relations with disease activity and joint damage. Ann Rheum Dis 2001;60:619–26.
      OpenUrlAbstract/FREE Full Text
    18. ↵
      1. Hulejová H,
      2. Baresová V,
      3. Klézl Z,
      4. et al
      . Increased level of cytokines and matrix metalloproteinases in osteoarthritic subchondral bone. Cytokine 2007;38:151–6.
      OpenUrlCrossRefPubMedWeb of Science
    19. ↵
      1. Kraus VB,
      2. Stabler TV,
      3. Luta G,
      4. et al
      . Interpretation of serum C-reactive protein (CRP) levels for cardiovascular disease risk is complicated by race, pulmonary disease, body mass index, gender, and osteoarthritis. Osteoarthr Cartil 2007;15:966–71.
      OpenUrlCrossRefPubMedWeb of Science
    20. ↵
      1. Mazières B,
      2. Garnero P,
      3. Guéguen A,
      4. et al
      . Molecular markers of cartilage breakdown and synovitis at baseline as predictors of structural progression of hip osteoarthritis. The ECHODIAH Cohort. Ann Rheum Dis 2006;65:354–9.
      OpenUrlAbstract/FREE Full Text
    21. ↵
      1. Melikoglu MA,
      2. Yildirim K,
      3. Senel K
      . Relationship between radiographic grading of osteoarthritis and serum beta-2 microglobulin. Ir J Med Sci 2009;178:151–4.
      OpenUrlPubMedWeb of Science
    22. ↵
      1. Otterness IG,
      2. Swindell AC,
      3. Zimmerer RO,
      4. et al
      . An analysis of 14 molecular markers for monitoring osteoarthritis: segregation of the markers into clusters and distinguishing osteoarthritis at baseline. Osteoarthr Cartil 2000;8:180–5.
      OpenUrlCrossRefPubMedWeb of Science
    23. ↵
      1. Sharif M,
      2. Elson CJ,
      3. Dieppe PA,
      4. et al
      . Elevated serum C-reactive protein levels in osteoarthritis. Br J Rheumatol 1997;36:140–1.
      OpenUrlFREE Full Text
    24. ↵
      1. Sharif M,
      2. Shepstone L,
      3. Elson CJ,
      4. et al
      . Increased serum C reactive protein may reflect events that precede radiographic progression in osteoarthritis of the knee. Ann Rheum Dis 2000;59:71–4.
      OpenUrlAbstract/FREE Full Text
    25. ↵
      1. Sowers M,
      2. Jannausch M,
      3. Stein E,
      4. et al
      . C-reactive protein as a biomarker of emergent osteoarthritis. Osteoarthr Cartil 2002;10:595–601.
      OpenUrlCrossRefPubMedWeb of Science
    26. ↵
      1. Spector TD,
      2. Hart DJ,
      3. Nandra D,
      4. et al
      . Low-level increases in serum C-reactive protein are present in early osteoarthritis of the knee and predict progressive disease. Arthritis Rheum 1997;40:723–7.
      OpenUrlCrossRefPubMedWeb of Science
    27. ↵
      1. Stürmer T,
      2. Brenner H,
      3. Koenig W,
      4. et al
      . Severity and extent of osteoarthritis and low grade systemic inflammation as assessed by high sensitivity C reactive protein. Ann Rheum Dis 2004;63:200–5.
      OpenUrlAbstract/FREE Full Text
    28. ↵
      1. Thanassoulis G,
      2. O'Donnell CJ
      . Mendelian randomization: nature's randomized trial in the post-genome era. JAMA 2009;301:2386–8.
      OpenUrlCrossRefPubMedWeb of Science
    29. ↵
      1. Despriet DD,
      2. Klaver CC,
      3. Witteman JC,
      4. et al
      . Complement factor H polymorphism, complement activators, and risk of age-related macular degeneration. JAMA 2006;296:301–9.
      OpenUrlCrossRefPubMedWeb of Science
    30. ↵
      1. Hofman A,
      2. Breteler MM,
      3. van Duijn CM,
      4. et al
      . The Rotterdam Study: 2010 objectives and design update. Eur J Epidemiol 2009;24:553–72.
      OpenUrlCrossRefPubMedWeb of Science
    31. ↵
      1. Dahaghin S,
      2. Bierma-Zeinstra SM,
      3. Ginai AZ,
      4. et al
      . Prevalence and pattern of radiographic hand osteoarthritis and association with pain and disability (the Rotterdam study). Ann Rheum Dis 2005;64:682–7.
      OpenUrlAbstract/FREE Full Text
    32. ↵
      1. Reijman M,
      2. Hazes JM,
      3. Pols HA,
      4. et al
      . Role of radiography in predicting progression of osteoarthritis of the hip: prospective cohort study. BMJ 2005;330:1183.
      OpenUrlAbstract/FREE Full Text
    33. ↵
      1. Ornetti P,
      2. Brandt K,
      3. Hellio-Le Graverand MP,
      4. et al
      . OARSI-OMERACT definition of relevant radiological progression in hip/knee osteoarthritis. Osteoarthr Cartil 2009;17:856–63.
      OpenUrlCrossRefPubMedWeb of Science
    34. ↵
      1. Kardys I,
      2. de Maat MP,
      3. Uitterlinden AG,
      4. et al
      . C-reactive protein gene haplotypes and risk of coronary heart disease: the Rotterdam Study. Eur Heart J 2006;27:1331–7.
      OpenUrlAbstract/FREE Full Text
    35. ↵
      1. van Meurs JB,
      2. Rivadeneira F,
      3. Jhamai M,
      4. et al
      . Common genetic variation of the low-density lipoprotein receptor-related protein 5 and 6 genes determines fracture risk in elderly white men. J Bone Miner Res 2006;21:141–50.
      OpenUrlCrossRefPubMedWeb of Science
    36. ↵
      1. Stephens M,
      2. Smith NJ,
      3. Donnelly P
      . A new statistical method for haplotype reconstruction from population data. Am J Hum Genet 2001;68:978–89.
      OpenUrlCrossRefPubMedWeb of Science
    37. ↵
      1. de Klerk BM,
      2. Schiphof D,
      3. Groeneveld FP,
      4. et al
      . Limited evidence for a protective effect of unopposed oestrogen therapy for osteoarthritis of the hip: a systematic review. Rheumatology (Oxford) 2009;48:104–12.
      OpenUrlAbstract/FREE Full Text
    38. ↵
      1. Ioannidis JP
      . Why most published research findings are false: author's reply to Goodman and Greenland. PLoS Med 2007;4:e215.
      OpenUrlCrossRefPubMed

    Footnotes

    • Funding This project was funded by European Commission framework 7 programme grant 200800 TREAT-OA and the Netherlands Genomics Initiative (NGI)/Netherlands Organisation for Scientific Research (NWO) project nr. 050-060-810.

    • Competing interests None.

    • Patient consent Obtained.

    • Ethics approval This study was conducted with the approval of the Erasmus University Medical School, Rotterdam, The Netherlands.

    • Provenance and peer review Not commissioned; externally peer reviewed.