Article Text

Download PDFPDF

Concise report
Lack of effect of TNFα blockade therapy on circulating adiponectin levels in patients with autoimmune disease: results from two independent prospective studies
  1. Mike J L Peters1,
  2. Pauline Watt2,
  3. Lynne Cherry2,
  4. Paul Welsh2,
  5. Eric Henninger3,
  6. Ben A C Dijkmans1,4,
  7. Iain B McInnes2,
  8. Michael T Nurmohamed1,4,
  9. Naveed Sattar2
  1. 1VU University Medical Center, Amsterdam, The Netherlands
  2. 2University of Glasgow, Glasgow, UK
  3. 3Merck Serono SA, Geneva, Switzerland
  4. 4Jan van Breemen Institute, Amsterdam, The Netherlands
  1. Correspondence to Mike J L Peters, Department of Rheumatology, VU University Medical Center, P O Box 7057, 1007 MB Amsterdam, the Netherlands; mjl.peters{at}


Background Adiponectin is an anti-inflammatory and potentially antiatherogenic molecule. Some recent reports suggest that tumour necrosis factor α (TNFα) blockade therapy increases circulating adiponectin levels, but data are sparse and inconsistent.

Methods Data from a double-blind placebo controlled study of onercept in 126 patients with psoriatic arthritis (PsA) and from pre- and post-adalimumab treatment in 171 patients with rheumatoid arthritis (RA) were used to examine the effect of TNFα blockade therapy on adiponectin.

Results Despite expected associations of adiponectin with gender and baseline high-density lipoprotein cholesterol and triglyceride, adiponectin levels did not change over time with TNFα blockade therapy in either group. The mean±SD absolute change in adiponectin levels was −0.23±4.6 μg/ml in patients with PsA treated with combined onercept 50 mg and onercept 100 mg (vs placebo, p=0.60) and 0.28±3.23 μg/ml in patients with RA treated with adalimumab (vs baseline, p=0.66).

Conclusion These results do not support a significant effect of TNFα blockade therapy on circulating adiponectin levels in patients with autoimmune disease.

Statistics from


Inflammatory rheumatic diseases such as rheumatoid arthritis (RA) and psoriatic arthritis (PsA) are associated with increased cardiovascular (CV) mortality and morbidity.1 Inflammation promotes a more adverse CV risk profile and may therefore explain the excess CV risk in these conditions.2 Observational studies show that patients with rheumatic diseases responding to tumour necrosis factor α (TNFα) blockade have a lower CV risk and improved CV risk profile commensurate with the inflammatory hypothesis of vascular disease.3 Mechanisms whereby inflammation may interact with traditional and novel CV risk factors accelerating atherosclerosis are not clearly defined. Further research in this area is therefore warranted to determine the pathways through which effective anti-inflammatory drugs may have their antiatherogenic effects. Adiponectin, an adipocyte-derived secreted protein, is a possible candidate molecule as it has anti-inflammatory, antiatherogenic and antidiabetic properties.4 5 It also has TNF homology.6 Accordingly, TNFα blockade therapy may yield beneficial effects on adiponectin (ie, increasing levels of circulating adiponectin) which may eventually ameliorate the CV burden. Data addressing the influence of TNFα blockade therapy on circulating adiponectin levels in rheumatic diseases are sparse and inconsistent.7,,12 Moreover, relevant data from randomised placebo controlled trials are lacking. To examine the effect of TNFα blockade therapy on circulating adiponectin levels, we analysed data from a double-blind placebo controlled study of onercept in patients with PsA and sought to confirm our observations using a large group of patients with RA treated with adalimumab.

Patients and methods


Data for the present study came from two separate datasets. The first study group comprised 126 patients with active plaque psoriasis and active PsA who were randomised to one of three double-blind treatment arms (placebo (n=42), onercept 50 mg three times weekly (n=42) or onercept 100 mg three times weekly (n=42)) for a total of 12 weeks of treatment. Further details of this specific study, including patient flow through the study, are reported elsewhere.13 The second study group comprised 171 consecutive patients with RA, all treated with adalimumab 40 mg subcutaneously every 2 weeks at the rheumatology outpatient clinic of the Jan van Breemen Institute, Amsterdam. Serum samples were collected before the first injection with adalimumab at baseline and at 16 weeks of treatment.

Risk factor measurements

Details about how clinical and CV risk factors were measured in both studies have been given elsewhere.13 14 Total plasma adiponectin was analysed using a commercially available kit (R&D Systems, Oxon, UK) with coefficients of variation of <8%.

Statistical analysis

For each parameter the values were summarised by treatment group using descriptive statistics at baseline and either week 12 for patients with PsA or week 16 for patients with RA. The change from baseline was summarised in the same way, expressed as absolute change, and was compared between placebo and each dose of onercept in PsA patient groups using non-parametric analysis of covariance, with ranked values of the change from baseline as the outcome and adjusting for ranks of the baseline value. All patients with PsA who were randomised and who received treatment were included in the analyses in an intent-to-treat manner. The Wilcoxon signed rank test was used to observe differences in adiponectin before (baseline) and after 16 weeks of adalimumab administration. The correlation between concentrations of adiponectin and CV risk factors were examined using Spearman rank correlation coefficients; p values <0.01 were considered significant without applying corrections for multiplicity.


Demographic data

Table 1 shows the baseline characteristics of the patients with PsA and RA evaluated in the present study. The comparator PsA patient groups were well balanced for demographic characteristics, inflammatory disease characteristics and CV risk factors.

Table 1

Baseline characteristics

Adiponectin and its relation to other CV risk factors at baseline

Adiponectin levels correlated positively with high-density lipoprotein (HDL) cholesterol (r=0.501, p<0.001 in PsA; r=0.553, p<0.001 in RA), apolipoprotein A-1 (r=0.422, p<0.001 in PsA) and sex hormone binding globulin (r=0.387, p<0.001 in PsA), and negatively with triglycerides (r=−0.356, p<0.001 in PsA; r=−0.373, p<0.001 in RA). No correlation was observed between adiponectin and C reactive protein (r=−0.117, p=0.191 in PsA; r=−0.052, p=0.499 in RA), intercellular adhesion molecule-1 (r=−0.106, p=0.235 in PsA), total cholesterol (r=0.123, p=0.167 in PsA; r=0.192, p=0.030 in RA), apolipoprotein B (r=−0.082, p=0.357 in PsA) and lipoprotein(a) (r=−0.024, p=0.792 in PsA). As anticipated, women had significantly higher adiponectin levels in both patient groups, but correlations between adiponectin and CV risk factors were almost identical in both genders (data not shown).

Changes in disease activity

The clinical outcomes of the patients with PsA will be reported separately. However, it is important to note that, according to the modified PsA response criteria, treatment with onercept 50 mg (p<0.05) and onercept 100 mg (p<0.01) resulted in a significantly better response relative to placebo. Body weight was stable (mean and median changes ≤1 kg) in all treatment groups. In patients with RA, treatment with adalimumab led to a significantly improved Disease Activity Index-28 score (and its individual components) at 16 weeks compared with baseline (p<0.001).

Changes in adiponectin levels

Mean±SD baseline levels of adiponectin were similar in all PsA groups (16.72±8.41 μg/ml in the placebo group, 17.45±8.79 μg/ml in the onercept 50 mg group and 17.46±9.03 μg/ml in the onercept 100 mg group). After 12 weeks of treatment adiponectin levels did not change significantly in any of the PsA treatment groups (absolute change in adiponectin levels −1.09±5.95 μg/ml in the placebo group; −1.05±5.00 μg/ml in the onercept 50 mg group (vs placebo, p=0.85); 0.59±4.07 μg/ml in the onercept 100 mg group (vs placebo, p=0.27); and −0.23±4.6 μg/ml in the onercept 50 mg + onercept 100 mg combined group (vs placebo, p=0.60); figure 1). Circulating adiponectin levels also did not change in patients with RA after 16 weeks of adalimumab administration (absolute change in adiponectin levels 0.28±3.23 μg/ml (vs baseline, p=0.66); figure 1). No significant correlations were found between changes in adiponectin levels and changes in C reactive protein in any of the groups.

Figure 1

Changes in adiponectin with tumour necrosis factor α blocking therapy. Mean adiponectin levels at baseline (black bars) and 12 weeks (grey bars) stratified for patients with psoriatic arthritis (PsA) receiving placebo, onercept 50 mg, onercept 100 mg and combined onercept 50 mg+onercept 100 mg; p values represent the difference in change in adiponectin levels between treatment groups relative to the placebo group. The last column represents mean adiponectin levels at baseline (black bar) and after 16 weeks (grey bar) for patients with rheumatoid arthritis (RA) receiving adalimumab 40 mg every 2 weeks; the p value for the difference in pre–post treatment change is reported owing to the lack of a placebo group (p=0.66).


There is considerable interest in the potential role of adiponectin in arthritis and vascular disease but the current evidence base is sparse. The present report details the effects of anti-TNF treatment on adiponectin levels in two prospective studies, one of which was a randomised controlled trial. As a confirmation of the applicability of our study, we observed the expected associations of adiponectin with other metabolic markers (positive correlations with HDL cholesterol, apolipoprotein A-1 and sex hormone binding globulin and negative correlation with triglyceride). However, despite these expected associations with metabolic markers, we did not observe any association of adiponectin with the inflammatory marker C reactive protein at baseline. In addition, TNFα blockade therapy was not associated with significant changes in circulating adiponectin levels over time despite significant reductions in inflammatory activity in both cohorts. These findings show that circulating adiponectin levels are unlikely to be meaningfully influenced by the inflammatory state or TNFα blockade therapy. Any reduction in CV risk associated with TNFα blockade in patients with arthritis is therefore difficult to attribute to the effects of circulating adiponectin concentration.

Previous studies (all in RA) have had mixed results with three showing increases and two showing no significant changes in adiponectin levels after treatment with TNFα blockade therapy.7,,12 Our studies are the largest to date to address this topic and therefore have considerable strengths over previous studies. The relevance of our findings is uncertain, but they give some degree of caution against claims for a major interaction between adiponectin and the systemic inflammatory process in patients with rheumatic disease. Additional evidence for the lack of interaction between TNFα blockade and adiponectin comes from a study examining the in vitro application of etanercept and adalimumab to adiponectin-stimulated synovial fibrobast cultures.15 This study found a reduction in interleukin 6 and matrix metalloproteinase 1 even when adiponectin (with its strong TNF homology) was preincubated with either adalimumab or etanercept. These findings suggest that adiponectin does not bind with TNFα blocking agents and that the anti-inflammatory effects of TNFα blocking agents are not mediated by adiponectin. However, the picture is complex and there is a need for further research to examine these paradoxical observations with regard to adiponectin and to investigate the precise role of adiponectin in autoimmune diseases.

A number of limitations should be acknowledged. We did not measure differing fractions of adiponectin in this study, however current research does not indicate a substantial difference in the association of total versus high molecular weight adiponectin with regard to metabolic factors.5 The studies were short in duration at 12–16 weeks and potentially longer follow-up would be desirable, although substantial changes in inflammation markers occur rapidly (within such timescales) with TNFα blockade and other metabolic markers, as previously reported.13 Despite these limitations, our results do not support a significant effect of TNFα blockade on circulating adiponectin levels in patients with autoimmune diseases. The beneficial effects of TNFα blockade therapy on CV risk and disease activity are therefore unlikely to be mediated via adiponectin.


The authors would like to thank the contributing clinicians who entered patients into the clinical phase of this study.



  • MTN and NS are joint senior authors

  • Funding Serono International SA had a role in designing the study and data collection involving patients with PsA. Serono gave permission to submit the manuscript for publication and approved the content of the manuscript. Abbott partially supported the clinical part of the study involving patients with RA but did not have any involvement in designing the study, data collection, writing the report and the decision to submit the paper for publication. MJLP received a EULAR bursary and this research was conducted while he was an ARTICULUM Fellow.

  • Competing interests EH is an employee of Merck Serono SA, Geneva, an affiliate of Merck KGaA, Darmstadt, Germany.

  • Ethics approval Ethics approval was obtained and all patients gave written informed consent to participate in the study.

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

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.