Mucosal Kaposiʼs sarcoma following protease inhibitor therapy in an HIV-infected patient

HIV-infected patients in whom AIDS-related complications have regressed following therapy with HIV protease inhibitors have recently been described. In these circumstances in particular, cutaneous Kaposi's sarcoma (KS) lesions may completely resolve and human herpesvirus (HHV)-8 may be cleared from peripheral blood mononuclear cells [1–3]. In contrast, we describe an HIV-infected man who became dangerously symptomatic with KS shortly after starting protease inhibitor treatment. A 43-year-old man was found to be HIV-infected in 1987 during investigation of thrombocytopenia. Between 1991 and 1996, he was treated with the nucleoside analogue reverse transcriptase inhibitors zidovudine, zalcitabine and lamivudine. His CD4 cell count declined and in January 1995 he had an episode of Pneumocystis carinii pneumonia despite co-trimoxazole prophylaxis. KS was observed in his mouth in March 1995 and proven by histology. In July 1996 he had radiotherapy for KS lesions on his tongue and penis. In December 1996, his CD4 cell count was 10 × 106/l and HIV viral load was 279 547 RNA copies/ml (Roche PCR, Roche Diagnostics, Welwyn Garden City, Hertfordshire, UK). He was started on ritonavir therapy, increasing to 600 mg twice daily, and saquinavir 600 mg twice daily. Seven weeks later, his CD4 count was 160 × 106 cells/l and HIV RNA was undetectable (< 400 copies/ml). His CD8 count, which had declined along with the CD4 cell count, increased 192 × 106/l a level not recorded since 1990 (Fig. 1).Fig. 1: . Peripheral blood CD4 and CD8 counts (× 106/l) from 1990 to 1997. ZDV, Zidovudine; ddC, zalcitabine; 3TC, lamivudine.Eight weeks after starting protease inhibitors he was admitted to hospital complaining that his voice had been hoarse for 1 month and that for 2 weeks he had been coughing up sputum containing fleshy masses. He had also noticed increasing dyspnoea. Laryngoscopy showed a swollen epiglottis with discoloration suggestive of KS. There was a danger of obstruction and a tracheostomy was performed. Biopsy confirmed KS and showed a mild–moderate lymphocytic infiltrate with predominant CD8 cells (not an unusual finding in KS) [4]. He was treated with local radiotherapy with an excellent response, the tracheostomy tube was removed in April, and he remains well. It is possible that, following introduction of protease inhibitors, improvement in CD4 and CD8 cell numbers and function led to an enhanced cytotoxic response to HHV-8 in a previously unnoticed epiglottic KS lesion. Inflammation and oedema of the lesion may have precipitated upper airway obstruction. A similar phenomenon has been reported in an HIV-infected patient with chronic hepatitis B who developed acute hepatitis when given protease inhibitors [5]. We would encourage physicians to be wary of mucosal KS lesions during the introduction of protease inhibitor therapy.