Tumour necrosis factor (TNF)-α is a key proinflammatory cytokine, which has an essential role in the host defence against intracellular pathogens. In addition to various autoimmune rheumatic diseases, TNF-α has also been implicated in the pathogenesis of HIV-1 infection promoting HIV replication in T cell lines and in lymphocytes in HIV-infected patients.1 2 Serum concentrations of TNF-α have been shown to increase as HIV-1 infection progresses,3 suggesting that TNF-α may contribute to disease progression. Thus, inhibition of TNF-α in the setting of HIV infection has been appealing, at least in theory. This appeal has been met with concern, however, due to the essential role for TNF-α in the host defence against infections. It has been clearly shown that anti-TNF therapy increases susceptibility to infections especially with Mycobacterium tuberculosis, atypical mycobacteria as well as other organisms.4

Little is known of anti-TNF-α therapy for rheumatic diseases in patients with HIV. A few case reports have been published,5^–9 and only one has provided long-term follow-up data.9 Given that patients with rheumatic diseases in the setting of HIV infection may not often respond to conventional therapy and may require additional therapy,8 we sought to analyse the safety and efficacy of treatment of various rheumatic conditions in HIV patients treated with etanercept, a soluble TNF receptor (p75):Fc fusion protein, infliximab, an anti-TNF-α chimeric monoclonal antibody, and adalimumab. We describe eight patients treated with etanercept and/or infliximab for their rheumatic diseases not responsive to conventional therapy in the setting of HIV infection.

PATIENTS AND METHODS

Eight HIV-1-positive patients with rheumatic diseases refractory to NSAIDs and disease modifying antirheumatic drugs (DMARDs) were treated with anti-TNF agents. These included five patients seen at the Thomas Street Clinic, a Harris County, Texas clinic specialising in outpatients with HIV infection seen by EJC, FMW and JDR (nos 3–6, 8), two patients were selected from the University of Texas at Houston university practice (JDR) (nos 1 and 2) and one from the Cedars-Sinai Medical Center practice (MI, MHW) (no. 7) on whom JDR consulted. Patient demographics, associated rheumatic disease, HIV duration before initiation of anti-TNF therapy, presence of highly active antiretroviral therapy (HAART) and past DMARDs are reviewed in table 1. The patients’ charts were reviewed for medication history, duration of both HIV-1 disease and their rheumatic disease, and their clinical manifestations. Laboratory data analysed consisted of absolute CD4 and HIV viral load, which were checked immediately prior to the beginning of anti-TNF treatment, at the first clinic visit afterwards, and periodically thereafter. As had been recommended for immunosuppressive agents in HIV-1-positive patients elsewhere,10 anti-TNF treatment was not begun in any patient with a CD4 count of less than 200/mm3 or an HIV-1 viral load of greater than 60 000 copies/mm3 (one patient, no. 7, who had a baseline CD4 count of 268 and previous Pneumocystis jiroveci pneumonia, received only high dose corticosteroids before anti-TNF therapy). Laboratory studies for the university practice patients were performed through the commercial laboratories contracted by the respective patient’s insurance company and through Harris County Hospital District’s chemistry and immunology laboratories for the Thomas Street Clinic patients. All patients received treatment with etanercept, 25 mg subcutaneous injections, twice weekly or 50 mg weekly initially after standard antituberculous screening and prophylaxis. Four patients were additionally treated with infliximab (5 mg/kg infusion every 6–8 weeks) and/or adalimumab (40 mg subcutaneously every other week) after not achieving a satisfactory clinical response with etanercept.

RESULTS

Eight patients with rheumatic diseases refractory to DMARDs were treated with anti-TNF agents who were followed 28.1 (SD 20.9) months (range 2.5–55). Of these, six had an excellent response with near total symptomatic remission (by physician global assessment), and one of the remaining two a sustained partial response (tables 1 and 2). One patient with AS only had a transient improvement of his back pain (patient no. 2) with a fall in his Bath Ankylosing Spondylitis Disease Activity Index score from a baseline of 4.5 to 2.8 after 1 month, which returned to 3.5 one month later and remained at this level thereafter, but his level of disease activity was low enough that continuing anti-TNF agents was not felt necessary. One other patient (no. 3) had a partial response, although she has continued anti-TNF treatment up to the present. All three patients with psoriatic arthritis (nos 5, 7 and 8) had near-complete clearing of their skin rash with anti-TNF treatment. Oral corticosteroids were used in three patients and in all but one (no. 1) were discontinued after anti-TNF treatment was begun.

In no patient did clinical deterioration or disease progression related to HIV occur in up to 48 months of follow-up. Two patients (3 and 5) manifested signs of infliximab hypersensitivity, and one of these (patient 5) was able to continue treatment with premedication with corticosteroids. No opportunistic infections or malignancies were observed over the period of follow-up. Of note, only patient 7 received Pneumocystis jiroveci prophylaxis (in this case with dapsone) because none of the other patients experienced falls in their CD4 counts to a level where such prophylaxis is used in our practice (<200/mm3), nor did any patient require Toxoplasmosis prophylaxis. Patient 1 in our series did experience one herpetic lesion 1 week after beginning etanercept. However, he had a previous history of recurrent herpetic infections and his symptoms resolved without any treatment or complications. Patients 3 and 5 had a history of pulmonary tuberculosis and latent tuberculosis (positive tuberculin skin test), respectively, and were treated previously with isoniazid prophylaxis. Active tuberculosis was not observed in either patient during treatment with prolonged anti-TNF therapy. Patient 4 had two separate occasions of iritis while on etanercept not experienced previously, as has been observed in patients without HIV.11 12 Patient 7 had a facial abscess while on infliximab, which responded to antibiotic therapy without any further complications.

No adverse effect was observed on CD4 cell counts or HIV viral load in seven of the eight patients treated. One patient (5) had a substantial increase in his HIV viral load after initiation of infliximab therapy (up to 428 503/mm3), requiring temporary discontinuation, but this did not recur with subsequent infusions.

Patient 4 was unable to continue anti-TNF treatment after 2004 due to incarceration. Patient 6 was lost to follow-up after 2005.

During the time period analysed, patients 1, 2, 3 7 and 8 remained on HAART without any adjustments in their therapy, while patients 4, 5 and 6 did not require HAART at the time anti-TNF treatment was begun. Of the latter three, only patient 5 is still followed and at last follow-up (September 2007) still does not require antiretroviral treatment despite being on anti-TNF treatment for nearly 3 years.

DISCUSSION

This retrospective open-label case series suggests that the use of anti-TNF therapy is safe and effective in HIV-positive patients with rheumatic diseases even on long-term follow-up. It should be emphasised that anti-TNF treatment was begun due to failure of response to standard antirheumatic therapy, not as a clinical trial, although the risks of anti-TNF drugs were carefully discussed with each patient before commencing therapy. Although the rate of serious infections in this group was small (one of eight or 12.5%), this is higher than the rate reported in recent studies in the literature of HIV patients.13 However, given the fairly small number of patients in this series and the fact that a “control” HIV-positive group was not included, it is impossible to speculate on the significance of this.

At the present time, there are few case reports of the use of anti-TNF agents in HIV-positive patients with rheumatic disease with anti-TNF-α in the setting of HIV infection, two using etanercept5 6 and three infliximab.7^–9 Only one of these, presented long-term follow-up data in two patients.9 Anti-TNF treatment was effective and well tolerated—the sole exception being the report of Aboulafia et al,5 where the CD4 counts was less than 50/mm3 at the start of anti-TNF treatment and etanercept had to be discontinued because of frequent polymicrobial infections. In our patients, treatment was not begun unless the CD4 count was >200 mm3 and the HIV viral load was <60 000 copies/mm3, as recommended for immunosuppressive agents in the setting of HIV.10

Recent controlled trials of anti-TNF-α-based therapies in patients with HIV have also been reported without demonstrating adverse effects on CD4 counts, HIV viral load or on susceptibility to infection.14 15

As this was primarily a study of safety, objective clinical response measures such as the disease activity score (for 28 joints; DAS28), erythrocyte sedimentation rate and C-reactive protein were not utilised (and in fact have not been validated in the potentially confounding setting of HIV infection, where the erythrocyte sedimentation rate and C-reactive protein can be elevated for other reasons).

These agents were shown to be generally well tolerated without any adverse effect on CD4 counts or on HIV viral load in these eight patients, suggesting that anti-TNF agents may be given with a reasonable ratio of benefits to risks if the patients’ underlying HIV (and other) infections are controlled and they are not severely immunocompromised (CD4 count of >200/mm3 and HIV viral load of <60 000 mm3).