Letters

Predicting relapse after treatment for American cutaneous leishmaniasis

Editor – The paper by Passos et al. about prediction of relapse in patients with American cutaneous leishmaniasis identified a negative skin test to leishmanin as a significant prognostic factor in a cohort study that enrolled patients with cutaneous and mucosal disease (1). The study is important because of the scarcity of research on prognostic issues in cutaneous leishmaniasis, and because it constitutes probably the biggest cohort of American cutaneous leishmaniasis ever followed in Brazil. However, some points should be borne in mind regarding the case definition, performance of the leishmanin skin test, the potential role of Leishmania (Leishmania) amazonensis as an etiolological agent of cases in the cohort, and the absence of exact data about the total antimony dose administered.

The case definition was based on clinical evaluation and any of three diagnostic tests, none of which identified the parasite; this would be of crucial importance because of the possibility that long-term prognosis could be linked to parasite species (2). According to data from a previous report, the best estimate for the relative frequency of L. (L.) amazonensis in the clinical setting where the present cohort was followed was 7.5% (4/53 evaluated patients), but could be as high as 18.2% considering the 95% confidence interval (CI) calculated for those numbers (3).

The skin test was not as sensitive as expected for a population of patients mainly infected with L. (Viannia) braziliensis (4). The study showed that, as a best scenario, skin test sensitivity was 79% (95% CI: 74–83) if all patients with a positive skin test had a positive result in the other two tests in the inclusion criteria. The modest performance of leishmanin could be related to the potential role of L. (L.) amazonensis as an important etiological agent in the sample. Silveira et al. showed that the skin test has lower sensitivity in cutaneous leishmaniasis patients infected with L. (L.) amazonensis compared with those infected with L. (V.) braziliensis (5). These findings give rise to the hypothesis that L. (L.) amazonensis infected individuals have a higher chance of being classified as negative by skin test and that the predictive factor would be the parasite species. Although this hypothesis does not invalidate the predictive potential of the skin test, the external validity of the data could be sacrificed and the results applied just to samples with similar proportions of cases caused by the two cited species.

The procedure used to confirm the predictive capacity of the skin test did not include data of actual treatment duration or total antimony dose administered, only that doses were "at least" the standard therapy recommended by the Ministry of Health of Brazil. The relevance of that factor must be emphasized because the observed cure rate was high if we consider a sample mainly infected by L. (V.) braziliensis. Traditionally, L. (V.) braziliensis infection has been considered difficult to cure, specially the mucosal form of the disease (6), and the absence of data on treatment duration precluded the appropriate interpretation of the modelling to identify relapse predictors. Patients with higher relapse rates could be those treated with a significantly lower total dose of antimonial. Curiously, patients with mucosal disease showed a lower relapse rate compared with patients with cutaneous disease. The estimated relapse rate was 5.5% (95% CI: 0.14–27.3) and 10.3% (95% CI: 7.2–14.5) for mucosal and cutaneous disease respectively. Although the statistical treatment of those proportions does not show significant difference (mainly because of the small sample of patients with mucosal disease) the numbers observed justify the hypothesis that if patients with mucosal involvement were effectively treated with a higher total antimony dose it would explain the difference in the relapse rate, supporting the need to include an estimator of the magnitude of exposure to treatment in the model.

Gustavo Adolfo Sierra Romero

Núcleo de Medicina Tropical
Universidade de Brasilia
Caixa Postal 04517
Brasilia, DF 70919- 970, Brazil
(email: gromero@unb.br)

Conflicts of interest: none declared.

1. Passos VMA et al. American cutaneous leishmaniasis: use of a skin test as a predictor of relapse after treatment. Bulletin of the World Health Organization, 2000, 78: 968–974.

2. Romero GAS. A espécie de Leishmania como principal determinante do prognóstico na leishmaniose tegumentar americana [The species of Leishmania as the principal determinant in the prognosis of American cutaneous leishmaniasis]. Revista da Sociedade Brasileira de Medicina Tropical, 1999, 32 (suppl II): 56–57.

3. Passos VMA et al. Leishmania (Viannia) braziliensis is the predominant species infecting patients with American cutaneous leishmaniasis in the State of Minas Gerais, southeast Brazil. Acta Tropica, 1999, 72: 251–258.

4. Cuba CAC et al. The use of different concentrations of leishmanial antigen in skin testing to evaluate delayed hypersensitivity in American cutaneous leishmaniasis. Revista da Sociedade Brasileira de Medicina Tropical, 1985, 18: 231–236.

5. Silveira FT et al. Cutaneous leishmaniasis due to Leishmania (Leishmania) amazonensis in Amazonian Brazil, and the significance of a negative Montenegro skin-test in human infections. Transactions of the Royal Society of Tropical Medicine and Hygiene, 1999, 85: 735–738.

6. Marsden PD. Mucosal Leishmaniasis ("espundia" Escomel 1911). Transactions of the Royal Society of Tropical Medicine and Hygiene, 1986, 80: 859–876.