Polymyositis and dermatomyositis, together with inclusion-body myositis (IBM), represent nosological forms of the rare idiopathic inflammatory myopathy (IIM) disease group, characterised by chronic, acquired skeletal muscle inflammation.¹

Cytokines such as tumour necrosis factor (TNF)α, TNFβ, interleukin (IL)1α, IL1β, IL2 and interferon (IFN)γ were raised in muscle biopsy specimens from patients with dermatomyositis or polymyositis and may contribute to the inflammatory cascade that leads to capillary and myofibril damage.²

Abnormally high levels of TNFα (and β) may be toxic to existing myofibrils, while simultaneously preventing the formation of new ones.³ An association of dermatomyositis with a −308A TNFα polymorphism has been reported.⁴ Serum levels of soluble TNF receptors 1 and 2 were raised in patients with active dermatomyositis or polymyositis when compared with those in controls⁵ or in patients with inactive disease.⁶ Increased TNFα mRNA expression in muscle biopsy specimens was reported in some,^7,8 but not all, studies.⁹

An in vitro study showed that the p75 TNFα soluble receptor alone, or in combination with the type II, IL1β soluble receptor, attenuated IL6 production and class I major histocompatibility complex expression on the surface of myoblasts stimulated with TNFα or IL1β.¹⁰

Therefore, TNFα may be an attractive therapeutic target, especially in myositis resistant to conventional treatments. Published studies are limited to sporadic case reports and small series.^11–13 We report our experience with eight patients, the largest series to date.

PATIENTS AND METHODS

Patient identification

This retrospective study was based on a review of medical records from patients with dermatomyositis or polymyositis treated in our tertiary referral centre (Hospital for Special Surgery, New York, USA) between 1998 and 2004. Patients gave informed consent. Eight patients refractory to corticosteroids and disease-modifying antirheumatic drugs (DMARDs) were treated with TNF blockers, and were followed up by the same rheumatologist. Six patients fulfilled the classification criteria of Peter and Bohan¹⁴ for definitive myositis and two patients (one with dermatomyositis and one with polymyositis who declined muscle biopsy) for probable myositis.

Anti-TNF treatment

Etanercept and infliximab were the anti-TNF agents used at doses similar to the ones used at the time for rheumatoid arthritis (ie, 3 mg/kg infusion of infliximab given at weeks 0, 2 and 6, and every 8 weeks thereafter, and 25 mg etanercept given subcutaneously twice weekly). The study required no minimal duration of treatment, and patients who had received at least one dose were included.

Efficacy and tolerance of anti-TNF treatment

Patients were followed up monthly with clinical assessment and laboratory tests, including creatine kinase, myoglobin, aspartate aminotransferase, alanine aminotransferase and lactate dehydrogenase levels in serum. Response was defined as improvement in fatigue (as reported by the patients), muscle strength (global assessment of the doctor: better or not better) and laboratory manifestations. Assessment in muscle strength included demonstrations of the following tasks by patients: climbing stairs, getting up from a deep seat and crossing legs while supine. Responders had to show improvement in the execution of all three tasks. No response was defined as the absence of noticeable improvement in 3 months after initiation of treatment. Partial response was defined as the persistence of one or several myositis-related clinical manifestations. In the case of a partial response, the remaining symptoms were recorded. In the case of discontinuation of treatment, the reason for discontinuation was recorded. All side effects, suspected or certain, were also noted, as well as the interventions prompted by their appearance.

RESULTS

Eight patients (five women and three men) were identified. Three were diagnosed as having dermatomyositis and five as having polymyositis. The mean age was 58.8 (SD 14.5; median 64) years. Mean disease duration was 8.5 (4.4; median 9) years. In all, 5 (62.5%) patients had undergone muscle biopsies and 3 (37.5%) patients had electromyograms confirming the diagnosis of dermatomyositis or polymyositis. All patients had pronounced weakness and had previously failed treatment with corticosteroids (oral or intravenous bolus), intravenous immunoglobulin G (IgG) and DMARDs (methotrexate, azathioprine, mycophenolate mofetil and leflunomide). The mean number of drugs used unsuccessfully before the institution of anticytokine treatment was 4.5 (1.4). In all, 6 (75%) patients were treated with etanercept, 1 (12.5%) with infliximab and 1 (12.5%) was sequentially treated with both agents. TNF blockade was an add-on treatment, and background treatment with corticosteroids, intravenous IgG and DMARDs was continued. Intravenous IgG and DMARDs had been used for at least 3 months before patients were considered to be incomplete responders and candidates for anti-TNF treatment. In all, 7 (87.5%) patients continued to receive low doses of oral prednisone and 6 (75%) continued monthly infusions of methylprednisolone, and 2 (25%) patients were continued on methotrexate and 3 (37.5%) on azathioprine. All patients were being infused monthly with intravenous IgG (2 g/kg).

Infliximab and etanercept were both well tolerated, with few reported side effects. One patient on etanercept had recurrent oesophageal candidiasis, which was treated with fluconazole. Another patient on etanercept had frequent headaches and reappearance of Raynaud’s phenomenon. A third patient developed sinus congestion, epistaxis and light-headedness. Sinus congestion and increase in cardiac irregularity in a patient with baseline atrial fibrillation prompted the discontinuation of etanercept in a fourth patient. A splenic tumour, potentially associated with the underlying polymyositis or the use of anti-TNF treatment, prompted the discontinuation of etanercept after 20 months of use. A patient who had prior radiographic evidence of mild interstitial fibrosis experienced no difference after anti-TNF treatment.

Of the eight patients, 6 (75%) had a favourable response, with improved motor strength and decreased fatigue. The mean duration of follow-up was 15.2 (SD 6.5; median 15) months. Of the six responders, five were treated with etanercept and one was treated with both agents. A favourable treatment effect was discernible in the first 2–3 months and consisted of both clinical improvement and a decrease in serum levels of muscle enzymes. Responders showed a dramatic drop in serum levels of creatine kinase. Their highly raised pre-treatment serum concentrations of creatine kinase, with a mean of 4463.5 (SD 4036.4) U/l, quickly dropped by a mean of 54.4% (27.7%) after institution of anti-TNF. Only one patient achieved normalisation of serum creatine kinase levels after treatment, a development that was paralleled with complete clinical remission (table 1).

In all, two patients did not respond and anti-TNF agents were stopped after a mean duration of 4 (SD 1.4; median 4) months. The first non-responder was a 73-year-old man with dermatomyositis, treated with four infusions of infliximab (in conjunction with oral prednisone, pulse methylprednisolone and methotrexate), with no clinical benefit. The second non-responder was a 68-year-old man with polymyositis, on etanercept for 5 months (along with pulse methylprednisolone, azathioprine and intravenous IgG), who showed no improvement and eventually discontinued the drug because of side effects. These two non-responders had normal pre-treatment serum concentrations of creatine kinase (mean 118.5 (SD 19.1) U/l) in common, which was in sharp contrast with the responders’ highly abnormal pre-treatment serum concentrations of creatine kinase (mean 4463.5 (4036.4) U/l). The mean drop in percentage for the non-responders, however, was 56.1% (20.4%), which was similar to that for the responders 54.4% (27.7%).

DISCUSSION

These findings suggest that the addition of TNFα inhibitor to standard treatment may be useful in some cases of refractory polymyositis and dermatomyositis. Both etanercept and infliximab seemed to be efficacious as adjunctive treatments in our patient group. The patients who benefited the most from anti-TNF adjunctive treatment had high serum levels of creatine kinase at baseline. Clinical improvement was paralleled by considerable reduction in the serum levels of creatine kinase during TNF inhibition.

Clinicians may consider the use of TNFα inhibitors in patients with treatment-refractory dermatomyositis or polymyositis, where standard treatment with corticosteroids, DMARDs or intravenous IgG has failed to induce and sustain remission. Serum levels of creatine kinase may guide the selection of patients for anti-TNF treatment, as we showed that patients with high creatine kinase levels, possibly indicating a more inflammatory type of disease, were more likely to benefit. Monitoring the levels of creatine kinase may also help in assessing response to treatment.

These results are consistent with those of previous reports on favourable outcomes with the use of TNF blockers in smaller series of patients with IIM. Hengstman et al¹² reported the results of two patients with newly diagnosed IIM (one with dermatomyositis and one with polymyositis) who were treated with infliximab (10 mg/kg every 2 weeks) alone. The increase in muscle strength correlated with improvements in serum levels of creatine kinase, electromyograms and the histological picture in post-treatment muscle biopsy specimens. In a follow-up paper, Hengstman et al¹³ suggested that the TNF blockade may be more efficacious as rapid induction treatment, as the benefit in these two patients waned over time without the concomitant use of DMARDs. Sprott et al¹¹ reported a case of polymyositis resistant to prednisone, methotrexate and intravenous IgG and intolerant to azathioprine, whose treatment with etanercept (25 mg subcutaneously twice weekly) led to improved strength and the discontinuation of prednisone.¹¹ In another report, etanercept treatment of four refractory patients with dermatomyositis or polymyositis resulted in decreased levels of creatine kinase, prednisone tapering and increased muscle strength in two of the four patients.¹⁵

Our study has several limitations because of its retrospective design and lack of standardised outcome measures. Muscle biopsy specimens before and after treatment were not obtained. It is possible that, in those three patients who did not undergo a muscle biopsy, the diagnosis of myositis was made erroneously, as some rare genetic myopathies can occasionally mimic myositis. A longer follow-up would be necessary to show a disease-modifying versus a remission–induction effect.

Despite these limitations, many of which are inherent in retrospective research, we believe that our real-life experience suggests a possible therapeutic role for TNF inhibitors, which can be proved only with randomised controlled studies using standardised efficacy measures.¹⁶ Until then, TNF blockade may be judiciously used as an adjunct treatment in refractory cases in which other conventional treatments have failed.