Article Text

Download PDFPDF

Systemic inflammation in osteoarthritis
  1. P Z Hrycaj
  1. Department of Rheumatology and Clinical Immunology, University of Medical Sciences, Poznan, Poland
  1. Correspondence to:
    Dr P Z Hrycaj
    Department of Rheumatology and Clinical Immunology, University of Medical Sciences, Winogrady 144, 61-626 Poznań, Poland;
  1. T Stürmer1,
  2. H Brenner2,
  3. W Koenig3,
  4. K P Günther4
  1. 1Division of Pharmacoepidemiology and Pharmacoeconomics, Division of Preventive Medicine, Brigham and Women’s Hospital, Harvard Medical School, Boston, USA
  2. 2Department of Epidemiology, German Centre for Research on Ageing, Heidelberg, Germany
  3. 3Department of Internal Medicine II—Cardiology, University of Ulm, Ulm, Germany
  4. 4Department of Orthopaedic Surgery, University of Dresden, Dresden, Germany

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Several studies have shown that the acute phase response may take place in osteoarthritis (OA), suggesting that low grade systemic inflammation may be present in patients with OA.1,2 I read with interest the paper by Stürmer et al on high sensitivity C reactive protein (CRP) in relation to the severity and extent of OA.3 As assessed by high sensitivity nephelometry, serum high sensitivity CRP was higher in 770 patients with advanced OA than in 567 age and sex matched healthy controls (geometric mean 2.5 mg/l v 1.7 mg/l, respectively). Moreover, severity of pain as measured by a visual analogue scale was associated with mean high sensitivity CRP. Interestingly, neither the bilateral nor the generalised extent of OA, nor any of the dimensions of the Western Ontario and McMaster Universities OA index (WOMAC) were associated with mean high sensitivity CRP concentrations. The authors concluded that the subjective severity of pain is associated with low level systemic inflammation in OA, and measurement of high sensitivity CRP may have some potential for monitoring and/or predicting the clinical course of OA.

In contrast with CRP, some acute phase proteins like α1-acid glycoprotein (AGP) or α1-antichymotrypsin (ACT) are glycoproteins and possess glycosylation sites attached by N-glycosidically bound, complex-type oligosaccharide side chains.4 Heteroglycans of acute phase proteins share the common core structure but differ in their outer chain sequences. According to the number of these oligosaccharide chains bi-, tri- and tetra-antennary heteroglycans can be distinguished. This structural diversity (termed “microheterogeneity”) results in different reactivity with the lectin concanavalin A (con A). It has been shown that biantennary side chains react strongly with con A. Thus, diverse microheterogeneous forms of acute phase glycoproteins, containing different number of biantennary heteroglycans, differ in their reactivity with con A.5 Glycosylation of acute phase proteins takes place in the liver and is controlled by cytokines.6

Affinity immunoelectrophoresis with con A is a simple technique that can be used to study the glycosylation pattern of acute phase proteins.7 Glycosylation variants of AGP/ACT can be separated during electrophoresis in a gel containing con A, and the area enclosed by the precipitates representing microheterogeneous variants of AGP and ACT can be measured by planimetry (fig 1). The results are usually expressed as the reactivity coefficients (AGP RC and ACT RC, respectively), calculated according to the formula: total area under the peaks of the con A reactive variants divided by the area enclosed by the peak representing the con A non-reactive variant.

Using affinity immunoelectrophoresis, we studied the systemic inflammatory response in 61 patients with OA classified as having clinically active (patients with rest joint pain, tenderness, joint swelling or effusion, n = 37) and clinically non-active (patients with radiological evidence of OA with no or mild clinical symptoms, n = 24) disease.8 In contrast with the study by Stürmer et al, patients with advanced OA and severe deformities were not included.

We found a significant decrease in the reactivity of AGP and ACT with con A in patients with clinically active OA in comparison with 24 patients with clinically non-active disease (p<0.001 and p<0.05 for AGP RC and ACT RC, respectively, fig 2). Concentrations of AGP, ACT, and low sensitivity CRP did not differ significantly between the groups. Serum concentrations of interleukin (IL) 1β, IL 6, and tumour necrosis factor α (TNFα) were either undetectable or low. However, in six of the seven synovial fluids available, IL6 concentrations were higher than in the respective serum samples. For TNFα the same could be shown in one case only.

Our findings suggest that there are changes in the microheterogeneity of acute phase glycoproteins in OA similar to those seen in rheumatoid arthritis and other chronic inflammatory conditions.9 As glycosylation of acute phase proteins does not depend on the expression of genes encoding the polypeptide chains of these proteins,9 glycosylation of acute phase proteins may possibly be more sensitive to the cytokine control than their synthesis. Thus, even small fluctuations in the serum profile of cytokines might eventually lead to alterations in the microheterogeneity of acute phase glycoproteins, while having no apparent influence on their serum concentration. Our data suggest that determination of microheterogeneity of acute phase glycoproteins may help to determine systemic inflammatory activity in OA and may possibly be more sensitive than measurement of serum concentration of acute phase proteins, including high sensitivity CRP.

Figure 1

Glycosylation pattern of AGP (A) and ACT (B) in the sera of patients with clinically non-active (left) and clinically active OA (right). Note relative increase in the concentration of con A non-reactive variant 0 combined with decrease in the concentration of the con A reactive variants 1–3 in the serum of a patient with clinically active disease compared with the patient with clinically non-active OA.

Figure 2

Notched box and whiskers plots showing statistical summaries (medians, 95% confidence limits, square root of the number of observations, range, and extreme values) of AGP RC and ACT RC. (A) AGP RC in patients with clinically active OA; (B) AGP RC in patients with clinically non-active OA; (C) ACT RC in patients with clinically active OA; (D) ACT RC in patients with clinically non-active OA.


Authors’ reply

We thank Dr Pawel Hrycaj for his comment on our manuscript on determinants of low grade systemic inflammation as assessed by high sensitivity C reactive protein (CRP) in patients with advanced osteoarthritis (OA).1 We agree with Dr Hrycaj that there is a need to elucidate further low grade systemic inflammation in patients with OA and that research based on biochemical and pathophysiological concepts is promising.

In his letter, Dr Hrycaj compared a variety of markers of inflammatory response in 37 patients with clinically active OA, who were not further characterised, with those in 24 patients with non-active disease (again not further characterised) without presenting or taking into account information on possible determinants of these markers. He found that microheterogeneity of acute phase glycoproteins but not the serum concentrations of acute phase proteins, including high sensitivity CRP, were associated with the clinical severity of disease and concluded that determination of microheterogeneity may possibly be a more sensitive measure of the systemic inflammatory activity of OA than high sensitivity CRP.

Our study,1 based on the concepts and methods of clinical epidemiology, was very different in its aim, design, and use of analytic techniques. We focused on high sensitivity CRP as marker of subclinical systemic inflammation because this marker has well established epidemiological and clinical determinants,2,3 little diurnal variation,4 and varies only moderately within a person, allowing long term prediction of disease.5 We then assessed independent determinants of this marker in a well described population of 770 patients who were recruited in four clinical centres using a standardised protocol and interview. In our analyses, known and suspected determinants of serum levels were taken into account using multivariable regression methods.1 We found that severity of pain was a predictor of serum levels of high sensitivity CRP independent of age, sex, body mass index, smoking, alcohol consumption, and comorbidity (hypertension, coronary artery disease, congestive heart failure, diabetes).

The results presented by Dr Hrycaj are difficult to interpret owing to a lack of information on the selection of patients, on basic characteristics of the two groups compared, and determinants of variability of the proposed microheterogeneity, including diurnal and day to day within-person variability. Furthermore, the data would be much more convincing if they had been analysed with possible differences in the characteristics of the two groups other than the activity of OA taken into account.

Nevertheless, the results presented by Dr Hrycaj appear to support our conclusion that severity of pain may be associated with levels of low grade systemic inflammation in patients with OA, and we hope that they will encourage further research in this area.

As in other areas of medical research, an interdisciplinary approach combining the areas of expertise of clinicians, basic scientists, and epidemiologists seems to be most promising.