Systemic Steroid Treatment of Paradoxical Upgrading Reaction in Patients with Lymph Node Tuberculosis

Sir—We read with interest the article by Hawkey et al. [1]. They report that paradoxical upgrading reactions (PUR) to antituberculous treatment occurred in 23% of non—HIV-infected patients. This is the highest percentage ever reported in an immunocompetent population, in whom PUR is usually <5% [2, 3]. A plausible explanation is that the authors' definition of PUR includes worsening of tuberculosis in patients treated for only 10 days, and the presence of a discharging sinus and persistent lymphadenopathy. We think that a minimum of 4 weeks should elapse between the start of antituberculous therapy and the development of PUR because, by definition, an initial clinical improvement should take place before worsening [2–4]. Also, because a discharging sinus was a well-known complication of benign tuberculous lymph node disease in the preantibiotic era, it is difficult to justify as a defining symptom of PUR. These considerations are important in the analysis of the effect of corticosteroids in the treatment of PUR. In contrast with previous studies [2–5], Hawkey et al. [1] did not find any evidence of a beneficial effect and questioned the effectiveness of corticosteroids to treat PUR. We think that their analysis could be biased by their definition of PUR, by the different use of steroids depending upon the severity of disease, and by the difficulties involved in the follow-up of the evolution of PUR by means of medical charts. In our experience with patients with severe PUR to antituberculous treatment after exposure to infliximab, anti-inflammatory treatment played an important role [4]. In our cohort study, surgery to control PUR was required in all patients that did not receive anti-inflammatory agents. Conversely, no patient treated with anti-inflammatory drugs required surgery. Our results, and the well-known beneficial effects of corticosteroid therapy in reducing edema in intracranial tuberculomas [5–7], would suggest that severe PUR must be treated with steroids. Dosing and duration of treatment have not been defined; our limited experience would indicate that 4–6 weeks of oral steroids is adequate. Hawkey et al. [1] made the interesting observation that high peripheral blood monocyte counts at baseline were associated with greater risk of developing PUR. Elevation of the TNF-α level, stimulated by lipoarabinommanan and other lipopolysaccharides present in the Mycobacterium tuberculosis cell wall, has been postulated as an initial step in the pathogenesis of PUR [8, 9]. TNF-α is secreted by macrophages and monocytes, and the observations of Hawkey et al. [1] give support to the hypothesis that increased production of this cytokine and its proinflammatory activity play a central role in the development of PUR. Better understanding of PUR as an inflammatory disease is important to understanding the beneficial effect of steroid treatement in these patients. In conclusion, the optimal managment for PUR is unknown. However, because the beneficial effect of steroids in patients with severe PUR has been well documented, early recognition and treatment should result in a more favorable outcome. Nevertheless, the role of anti-inflammatory therapy deserves further study. Potential conflicts of interest. C.G.V. and J.G.: no conflicts.