Plasma IL-6 as a marker of mycobacterial immune restoration disease in HIV-1 infection

Some patients develop inflammatory processes at the site of pre-existing or subclinical opportunistic infections during the early course of highly active antiretroviral therapy (HAART) [1]. This phenomenon of immune restoration disease (IRD) is an emerging cause of morbidity in the era of HAART. There are no specific tests to confirm the diagnosis of IRD. We present a case of IRD-associated mycobacterial abdominal adenopathy, and suggest that plasma IL-6 concentrations maybe a marker of IRD. A 24-year-old man with HIV-1 infection diagnosed 2 years previously was commenced on trizivir and efavirenz. He had no previous history of mycobacterial disease, and his nadir CD4 cell count was 150 cells/μl. He was well at the initiation of HAART with a CD4 cell count of 150 cells/μl and an HIV-1-RNA viral load of 150 000 copies/ml. His physical examination was normal at the time. Three weeks later he presented with a one-week history of constipation; his abdominal examination was unremarkable. After a further 2 weeks he presented to hospital with abdominal pain and constipation. He had no fever, cough, shortness of breath or weight loss. His CD4 cell count had risen to 294 cells/μl and his viral load had fallen to 1770 copies/ml. His white blood cell count was 7.4 × 109 cells/l and C-reactive protein level was 151 mmol/l. His chest X-ray was normal and blood cultures were negative. A computed tomography (CT) scan of his abdomen showed an enlarged mesenteric lymph node. Fine needle aspiration revealed acute-on-chronic inflammation with focal necrosis, no granulomas and no evidence of malignancy. The patient was treated with analgesia and laxatives and his symptoms resolved. A repeat abdominal CT scan 4 weeks later showed that the mass had increased in size to 10 × 10 cm. A CT-guided biopsy revealed non-necrotic epitheloid granulomas with multinucleated giant cells and no acid fast bacilli. A laparotomy was performed to obtain a tissue diagnosis, and histology showed extensive areas of caseating necrosis, epitheloid granulomas within mesenteric lymph nodes and scanty acid fast bacilli. The patient was commenced on antituberculous therapy with rifampicin, ethambutol, isoniazid and pyrazinamide and made a full recovery. The plasma IL-6 concentration was measured retrospectively using enzyme-linked immunosorbent assay kits (R&D Systems, Abingdon, UK) in the patient and two control patients who were started on HAART contemporaneously, but did not develop IRD (Fig. 1). The plasma IL-6 concentrations were similar in the patient and controls before starting HAART (Fig. 1); however, the IL-6 concentration in the patient increased from 0 to 20 pg/ml after only one week of antiretroviral therapy and fell to 0 pg/ml after antituberculous therapy, but remained constant at 0 pg/ml in the control patients.Fig. 1. Plasma IL-6 concentration in a patient with immune restoration disease and controls (n = 2).: ——□—— Patient; - - - -▿ - - - - controls. HAART, Highly active antiretroviral therapy.IRD typically occurs soon after the commencement of HAART, often within the first month [1]. IRD involving various mycobacterial infections has been described [1,2]; these patients often present with fever and lymphadenopathy. Microbiological specimens remain negative and the diagnosis is made by histological investigation. IRD has been postulated to be caused by an enhanced immune response to a previously subclinical infection to which the host was unable to mount an immune response. There is a rapid increase in CD4 cell numbers and T cell repertoire alterations. In particular, the T cell lymphoproliferative response to tuberculin is increased, with a restoration of tuberculosis-specific delayed type hypersensitivity responses [1]. Mycobacterium tuberculosis-associated IRD may thus be an exaggerated immune response to M. tuberculosis. This theory is supported by the fact that the clinical features of opportunistic infection-related IRD are different to the features of opportunistic infections occurring in immunocompromised patients not on therapy. The treatment of IRD is controversial, but involves as antimicrobial therapy for the suspected disease, with or without corticosteroids. IL-6 is a cytokine produced by monocytes and macrophages, which activates CD4 cells and induces the acute phase response. Our findings are in accordance with the results of Stone et al. [3], who showed that patients with a history of herpes virus-related IRD produced more IL-6. In our patient, within one week of starting HAART, the IL-6 concentration reached a peak that preceded the development of symptoms. Although IL-6 concentrations are elevated in patients with M. tuberculosis treated with HAART [4], the increase in IL-6 in our patient was unlikely to have been caused by M. tuberculosis in itself because the IL-6 was initially similar to those controls on HAART until just before the onset of his symptoms. In addition, the control patients increased their CD4 cell counts without an increase in the IL-6 concentration. Furthermore, the increase in the IL-6 concentration in our patient preceded the development of the abscess, and when the abscess was treated the IL-6 concentration decreased. Taken together, these data implicate IL-6 in the aetiology of tuberculosis-associated IRD, and suggest that IL-6 is produced as part of an exaggerated response to sparse quantities of mycobacteria.