A 27-year-old HIV-infected woman with a history of pulmonary tuberculosis (TB) treated successfully 2 years earlier was referred for antiretroviral treatment (ART) in a South African township. At baseline, she was asymptomatic, and the only abnormal clinical finding was some wasting. Blood tests revealed a CD4 cell count of 58 cells/μl, a plasma viral load of 113 375 copies/ml and mild normochromic anaemia. She commenced therapy with zidovudine, lamivudine and nevirapine and continued prophylaxis with trimethoprim–sulphamethoxazole. On review at 4 weeks, adherence to treatment was excellent, she was asymptomatic and reported improving energy levels. Moreover, her weight and blood haemoglobin concentration had both increased significantly. After 6.5 weeks of ART, however, she developed cough, fever and right-sided chest discomfort. Sputum smears were negative for acid-fast bacilli, but a presumptive radiological diagnosis of pulmonary TB was made pending sputum culture results. Rifampicin-containing anti-TB chemotherapy (retreatment regimen) was started. She presented again after 8 weeks of ART in very poor clinical condition. She was pyrexial (39.1°C) and had marked respiratory failure, clinical signs in the right anterior chest wall and bipedal oedema. Her condition continued to deteriorate rapidly in hospital where she died. Drug-susceptible Mycobacterium tuberculosis was subsequently cultured from the sputum. The principal findings at a post-mortem examination were an extensive, noncavitating pneumonic process in the upper lobe of the right lung and hepatic congestion consistent with right-sided cardiac failure secondary to respiratory failure. Histological examination of lung sections showed multiple foci of extensive caseous necrosis, a mononuclear cell infiltrate and numerous acid-fast bacilli (Fig. 1). A pattern of intra-alveolar granulation tissue with plugging of airways and alveolar oedema was visible. The histological appearance was characteristic of bronchiolitis obliterans organizing pneumonia (BOOP). In addition, bronchial lymph nodes showed recent granuloma formation but no acid-fast bacilli.Fig. 1: Histological examination of sections of diseased lung. A pattern of intraalveolar granulation tissue with plugging of airways and deposition of collagen was observed (top left, haematoxylin and eosin, ×20). This histological appearance is characteristic of bronchiolitis obliterans organising pneumonia. Ziehl Neelson staining (top right) revealed the presence of numerous acid-fast bacilli (arrows). Immunostaining showed very few CD3+ T cells with very few CD4+ or CD8+ T cells. Only two cells (arrows) stained positively for CD4 cells in this field (bottom left). However, staining for CD68 cells revealed that the overwhelming majority of mononuclear cells present were macrophages (bottom right).Immunostaining of involved lung tissue sections revealed very few CD3+ lymphocytes (either CD4+ or CD8+ cells), and CD56 staining for natural killer cells was negative. However, the overwhelming majority of inflammatory cells stained strongly with anti-CD68 antibody, indicating that they were macrophages (Fig. 1). Despite initially responding well to ART, this patient had developed pulmonary TB during the second month of treatment. Subsequent rapid clinical deterioration and death were attributable to the development of an unusually florid inflammatory process in the lungs that was directly associated with the microbiologically confirmed TB. Collectively, these observations are consistent with a diagnosis of ART-induced severe ‘unmasking’ immune reconstitution disease (IRD), a condition that is thought to result from dysregulated immunopathological host responses to subclinical active disease [1,2]. This case raises several points of interest, including the unusually severe clinical course, the histological pattern of disease and the predominant inflammatory cell type identified. We have previously reported that incident TB during the initial weeks of ART is very common in this setting [3] and have speculated that a proportion of these may be attributable to immune-mediated ‘unmasking’ disease. Our experience is that most patients presenting in this period do not manifest with unusually severe or hyperinflammatory features [1]. The present case, however, indicates that severe and ultimately life-threatening disease may occasionally occur as has also been described in a previous report from London [4]. Severe necrotizing inflammation has been observed in tissue specimens from some previous patients with mycobacterial IRD [5]. However, the finding of florid inflammation with features consistent with BOOP was striking. BOOP is a distinct histopathological entity that may be associated with a broad range of insults to the lung or may occur as an idiopathic form [6]. BOOP, however, has not been previously reported in association with either TB or IRD. An important characteristic of BOOP is that it is usually highly responsive to corticosteroid therapy. Initiation of such treatment early in this woman's clinical course may have been life saving. Although it has been observed that TB IRD is accompanied by large expansions of CD4 T cells in peripheral blood [7,8], the immunopathological mechanisms at the site of disease remain undefined. We are unaware of previous reports of TB IRD in which immunostaining has been used to identify the phenotype of the cells involved. Although it has been broadly assumed that IRD results from the restoration of cell-mediated immune function, large numbers of CD68+ macrophages, but very few T cells, were found at the site of disease. This case report therefore provides some support for the alternative hypothesis that rapid restoration of innate immune function in macrophages is important [9]. Future studies of the immunopathogenesis of TB IRD should include careful exploration of the role of macrophages at the site of disease. Acknowledgement Stephen D. Lawn is funded by the Wellcome Trust, London, UK. Catherine Orrell cared for the patient. Helen Wainwright performed the histopathology. Stephen D. Lawn wrote the paper and all revised and approved the final version.
No takes yet. Share an insight, caveat, or question.
Lawn et al. (2008) studied this question.