Article Text

Download PDFPDF

Serum levels of sex steroid hormones and matrix metalloproteinases after intra-articular glucocorticoid treatment in female patients with rheumatoid arthritis
  1. T Weitoft1,
  2. A Larsson2,
  3. L Rönnblom3
  1. 1
    Department of Research and Development, County Council of Gävleborg/Uppsala University, Section of Rheumatology, Gävle Hospital, Sweden
  2. 2
    Department of Medical Sciences, Clinical Chemistry, Uppsala University Hospital, Sweden
  3. 3
    Department of Medical Sciences, Section of Rheumatology, Uppsala University Hospital, Sweden
  1. Tomas Weitoft, MD, PhD, Section of Rheumatology, Department of Internal Medicine, Gävle Hospital, 801 87 Gävle, Sweden; tomas.weitoft{at}


Objectives: To study metalloproteinase activity and sex steroid hormone production in serum after intra-articular glucocorticoid treatment for knee synovitis.

Methods: 18 female patients with rheumatoid arthritis and synovitis of the knee with need for intra-articular glucocorticoid treatment were included in this study. Serum samples of matrix metalloproteinases (MMP-1/TIMP complex and MMP-3), dehydroepiandrosterone sulphate, testosterone, oestradiol, steroid hormone binding globulin, follicle stimulating hormone and luteinising hormone were collected before injection with 20 mg triamcinolone hexacetonide, and 24 h, 48 h, 1 week and 2 weeks after injection, respectively.

Results: Serum levels of MMP-3 were significantly decreased, but MMP-1/TIMP complex was unaffected. Dehydroepiandrosterone sulphate, testosterone and oestradiol levels all decreased and tended to return to baseline levels during the observation period. Steroid hormone binding globulin, follicle stimulating hormone and luteinising hormone levels were unchanged.

Conclusions: Intra-articular glucocorticoid treatment causes a temporary, but considerable suppression of sex steroid hormone secretion. The reduction of MMP-3 indicates an inhibition of the inflammatory, but probably also the cartilage destructive processes within the treated joint.

Statistics from

In a previous study we found that serum levels of adrenocorticotrophic hormone and cortisol were significantly reduced for about 1 week after intra-articular glucocorticoid treatment for knee synovitis in patients with rheumatoid arthritis (RA).1 The influence on other adrenal hormones, such as androgens and oestrogen after such treatment, has not been studied before. The first aim of this study was to elucidate how serum levels of sex steroid hormones change after a single intra-articular glucocorticoid injection.

We have also reported the impact of intra-articular glucocorticoid treatment on markers of cartilage degradation.2 Serum levels of cartilage oligomeric matrix protein decreased significantly after the injection suggesting influence on the cartilage destruction process in arthritis. This effect may be caused by decreased activity of proteolytic enzymes, such as matrix metalloproteinases (MMP).3

MMP-3 is produced by synovial cells4 and correlates with parameters of disease activity.5 After intravenous or oral glucocorticoid treatment the serum levels increase,6 7 but in contrast, after intra-articular administration the MMP-3 level decreases significantly.4 The reason behind this discrepancy has been a matter of debate,8 but to our knowledge, the contradictory findings has not yet been confirmed.

Low-dose oral glucocorticoid therapy does not seem to influence the levels of MMP-1, a collagenolytic enzyme,6 but the effects on serum levels after intra-articular administration have not been studied before. The tissue inhibitor of metalloproteinases (TIMP) inhibits metalloproteinases by forming high-affinity complexes.9

The second aim of this study was to elucidate how serum levels of MMP-1/TIMP complex and MMP-3 change after intra-articular glucocorticoid treatment for knee synovitis in patients with RA.


At the outpatient rheumatology departments in Gävle and Uppsala female patients with RA10 and clinical signs and symptoms of knee synovitis were invited to participate in the study. Patients in function class 4 according to Steinbrocker11 and patients treated with tumour necrosis factor-α blockers, systemic glucocorticoids in the past year, a glucocorticoid injection in the affected joint in the past 6 months or any intra-articular injection in the past 2 months were excluded. No patient had been using oral contraceptives or hormone replacement therapy.

The disease activity was evaluated with the DAS 28 index12 and disability with the Swedish version of the Health Assessment Questionnaire.13 Using a lateral injection technique the knee joint was entered with a 0.8×40 mm needle. After aspiration of synovial fluid, a dose of 20 mg triamcinolone hexacetonide (Lederspan) was injected. Serum samples were collected immediately before the injection, after 24 h, 48 h, 1 week and 2 weeks respectively. To avoid diurnal variation all injections were performed and all samples were collected at 08.00 h at every occasion.

Informed consent was obtained from all patients and the study was approved by the Committee of Ethics, Faculty of Medicine, Uppsala University.

Laboratory methods

The samples were centrifuged for 10 min at 1800 g within 1 h and stored at −70°C until all samples had been collected and laboratory processing could be performed.

Dehydroepiandrosterone sulphate (DHEAS), oestradiol, follicle stimulating hormone (FSH), luteinising hormone (LH), steroid hormone binding globulin (SHBG) and testosterone were analysed on a Modular E 170 immunoanalyser (Roche, Mannheim, Germany). The total analytic imprecision (coefficient of variation) of the methods were: DHEAS: 5.7–6.0%; oestradiol: 4.5–8.5%; FSH: 1.5–1.8%; LH: 2.2%; SHBG: 1.5%; testosterone: 6.8–7.6%.

The normal interval for serum levels of postmenopausal women are: FSH 25–140 IU/l, LH 7–60 IU/l, oestradiol 20–70 pmol/l, DHEAS 0.3–6.0 μmol/l, testosterone <3.0 mmol/l.

MMP-3 and MMP-1/TIMP complex were measured by sandwich ELISAs (DY513 and DY1550), R&D Systems, Minneapolis, MN, USA. The total imprecision of the assays were approximately 7%.

Statistical calculations

To test for significant changes of serum hormone levels within the group Friedman’s test for multiple dependent data was used and for the pairwise comparisons with baseline the Wilcoxon Rank Sum test was used. p<0.05 was considered significant. Statistical analyses were performed using the statistical computer software program SPSS, version 15.0 for Windows NT (SPSS Inc., Chicago, IL, USA).


The original cohort of 20 patients with RA have been presented before,1 2 but in this investigation the results from a male patient were excluded and the samples from another patient were not available. Consequently, a total of 18 female patients remained and were included in the study. Their characteristics, including the subgroup of 12 postmenopausal patients, are presented in table 1.

Table 1 Patient characteristics for 18 female patients with RA and the subgroup of 12 postmenopausal patients, prior intra-articular glucocorticoid knee injection

Compared with baseline the DHEAS and testosterone levels decreased significantly after the intra-articular glucocorticoid injection (p<0.001, fig 1), whereas SHBG levels remained unchanged.

Figure 1 Serum levels of dehydroepiandrosterone sulphate (DHEAS) and testosterone in 18 female patients with rheumatoid arthritis after intra-articular glucocorticoid treatment for knee synovitis. Serum levels of oestradiol are shown for the subgroup of postmenopausal patients. P-values show difference to baseline levels.

Serum levels of oestradiol, FSH and LH were only analysed in the subgroup of postmenopausal patients. The oestradiol level decreased significantly, but FSH and LH did not change (fig 1).

Serum levels of MMP-1/TIMP complex were unaffected, but MMP-3 levels (fig 2) were significantly decreased (p<0.001) and remained low during the 2-week observation period.

Figure 2 Serum levels of matrix metalloproteinases (MMP-3 and MMP-1/TIMP complex) after intra-articular glucocorticoid treatment for knee synovitis in 18 female patients with rheumatoid arthritis. P-values show difference to baseline level.


The main finding in our study is that intra-articular glucocorticoid injections suppress serum levels of androgens and oestrogen in female patients with RA. The decreases of DHEAS, testosterone and oestradiol were considerable, but tended to be reversible. However, after 2 weeks the serum testosterone level was still reduced compared with baseline levels. Furthermore, we confirmed that serum MMP-3 decreases after intra-articular glucocorticoid administration, but levels of the MMP-1/TIMP complex remained unchanged.

The adrenal response occurred rapidly and the maximum reduction of the hormones was observed within 24 hours. In comparison, the effect on serum oestrogen levels was slower. Oestradiol in postmenopausal individuals is mainly produced by conversion from testosterone in peripheral tissues, such as fat and synovium.14 The observed delay suggests a slower feedback response on hormone production in these tissues compared with the adrenals. This may be due to strong activation of the aromatase enzyme in the inflamed synovium.14 15 Less available circulating adrenal androgens, which supply the local tissue with substrates required for the peripheral oestrogen synthesis, could also contribute.

In postmenopausal women a feedback stimulation of FSH/LH production already exists, because oestradiol production in ovaries has decreased. In the present study no influence on FSH or LH secretion of intra-articular glucocorticoid treatment could be observed indicating that the reduction is not large enough to further increase FSH/LH secretion.

Most steroid molecules are bound to SHBG in plasma. The small proportion of free steroids is the metabolic active part and may pass through cellular membranes and bind to the intracellular steroid receptor. The SHBG levels did not change after intra-articular glucocorticoid administration, suggesting that not only hormone levels, but also the endocrine activity or local hormone conversion is impaired.

Joints with synovitis are supposed to be the major source of MMP-3 in the circulation of patients with RA.4 The inflamed knee contains a large mass of synovitis and as arthritis of the knee was the main symptom in our patients, we assume that the serum MMP-3 levels in our study reflect the ongoing process of inflammation and destruction in this joint. In RA patients decreased serum MMP-3 levels have been observed for 14 days after intra-articular glucocorticoid injection.4 This finding is confirmed in our study and serum MMP-3 levels were reduced already within the first 24 h after glucocorticoid injection with a persistent reduction during the 2-week observation period. As the patients were followed for only 2 weeks, the low levels may have remained for an even longer period.

One explanation for the observed decrease in serum MMP-3 levels after treatment could be a reduced release of MMP-3 molecules from the joint after the treatment. Another possibility is the reduction of MMP-3 synthesis by synovial fibroblasts resulting in a lower proteolytic enzyme activity in the treated joint. This would also support the proposed local cartilage protective effects of intra-articular glucocorticoid treatment.

In summary, not only the hormone production of the hypothalamic–pituitary–adrenal axis, but also other androgen and oestrogen hormones are significantly reduced during a short period after intra-articular glucocorticoid treatment. Furthermore, the decreases in serum levels of MMP-3 after such treatment suggests a suppression of metalloproteinase enzyme activity in the treated joints. Together with previous observations, we therefore interpret the findings as a support for a cartilage protective effect of intra-articular treatment in arthritis.


We wish to thank Britt-Inger Nilsson, Åsa Olsson and Gunilla Andersson for helping us with sample collecting. Hans Högberg has been of great help for statistical analysis.



  • Funding: The study was supported by grants from Swedish Research Council, the Centre for Research and Development, County Council of Gävleborg/Uppsala University, the Swedish Rheumatism Foundation, the 80-year Foundation of King Gustaf V and Wyeth Lederle.

  • Competing interests: none.

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.