Interventions for American Cutaneous and Mucocutaneous Leishmaniasis: A Systematic Review Update

Introduction Leishmaniasis is an important public health problem in the Americas. A Cochrane review published in 2009 analyzed 38 randomized controlled trials (RCT). We conducted a systematic review to evaluate the effects of therapeutic interventions for American cutaneous and mucocutaneous leishmaniasis. Methods All studies were extracted from PubMed, Embase, Lilacs (2009 to July, 2012 respectively), the Cochrane Central Register of Controlled Trials (6-2012) and references of identified publications. RCTs’ risk of bias was assessed. Results We identified 1865 references of interest; we finally included 10 new RCTs. The risk of bias scored low or unclear for most domains. Miltefosine was not significantly different from meglumine antimoniate in the complete cure rate at 6 months (4 RCT; 584 participants; ITT; RR: 1.12; 95%CI: 0.85 to 1.47; I2 78%). However a significant difference in the rate of complete cure favoring miltefosine at 6 months was found in L. panamensis and L. guyanensis (2 RCTs, 206 participants; ITT; RR: 1.22; 95%CI: 1.02 to 1.46; I2 0%). One RCT found that meglumine antimoniate was superior to pentamidine in the rate of complete cure for L. braziliensis (80 participants, ITT; RR: 2.21; 95%CI: 1.41 to 3.49), while another RCT assessing L. guyanensis did not find any significant difference. Although meta-analysis of three studies found a significant difference in the rate of complete cure at 3 months favoring imiquimod versus placebo (134 participants; ITT; RR: 1.45; 95%CI: 1.12 to 1.88; I2 0%), no significant differences were found at 6 and 12 months. Thermotherapy and nitric oxide were not superior to meglumine antimoniate. Conclusion Therapeutic interventions for American cutaneous and mucocutaneous leishmaniasis are varied and should be decided according to the context. Since mucosal disease is the more neglected form of leishmaniasis a multicentric trial should be urgently considered.


Introduction
Leishmaniasis is an important public health problem in 98 endemic countries of the world, with more than 350 million people at risk. WHO estimated an incidence of 2 million new cases per year (0.5 million of visceral leishmaniasis (VL) and l.5 million of cutaneous leishmaniasis (CL). VL causes more than 50, 000 deaths annually, a rate surpassed among parasitic diseases only by malaria, and 2, 357, 000 disability-adjusted life years lost, placing leishmaniasis ninth in a global analysis of infectious diseases. In addition, most patients have very poor access to the health system resulting in important underreporting of cases [1][2][3][4][5][6].
In the Americas, leishmaniases endemic areas extend from Mexico to Argentina. Approximately 67,000 clinical cases are reported every year and 40,840,000 people are at risk of developing the disease in over 21 countries, with estimated of 187,200 to 307,800 CL cases [1,6,7], and 4,500 to 6,800 VL cases [6]. While more than 90% of the VL cases occur in India, Bangladesh, Sudan, Ethiopia and Brazil, approximately 70% of CL cases occur in Afghanistan, Algeria, Colombia, Brazil, Iran, Syria, Sudan, Ethiopia, Nicaragua and Peru [5,6].
The leishmaniases are diseases caused by different species of parasites of the genus Leishmania and transmitted by vectors family Psychodidae infected from different reservoirs; it is also characterized by a large clinical polymorphism. Fifteen Leishmania species were identified as pathogenic to humans being causing cutaneous, mucosal (ML) and visceral leishmaniasis. The cutaneous and mucosal forms have a broad clinical spectrum that range from single or multiple localized skin lesions to severe diffuse and mucosal lesions [5,8].
The epidemiology of the leishmaniases is dynamic and the circumstances of transmission are continually changing in relation to environment, demography, human behavior, socioeconomic status, and other factors such as immunogenic profile of affected human populations [9][10][11][12][13][14][15].
In the New World, cutaneous leishmaniasis is caused by a variety of species belonging to the subgenera Leishmania and Viannia producing different clinical manifestations; however, part of the population have subclinical infections. Although the most frequent clinical form of cutaneous leishmaniasis presents as single or multiple lesions, disseminated lesions can also be observed. The lesions may occur anywhere in the body but commonly originate at the site of inoculation where initially a macular lesion forms, followed by a papule and then by a nodule that progressively increases in size and becomes ulcerated. These lesions can develop in weeks, months or years after infection [5,10,16].
Although lesions caused by L. mexicana may heal spontaneously in an average period of 4 months, this species and other such as L. amazonensis, L. venezuelensis, and L. pifanoi can cause diffuse cutaneous leishmaniasis, considered an anergic, severe, and chronic form of the disease. The response to the first therapeutic scheme is frequently unsatisfactory, due to changes in immunological conditions, physiological or nutritional characteristics of patients or to specific pharmacokinetics factors of drugs used [5,[17][18][19].
An atypical form of CL has been described at the same geographical area of VL presenting circumscribed and nonulcerated lesions; it mainly affects older children and young adults, while visceral leishmaniasis presents predominantly in children less than 5 years. This clinical form is caused by L. infantum (syn. L. chagasi) that can evolve into a visceral form, in patients with deprived immunological conditions [20,21].
Some species of the subgenus Viannia such as L. braziliensis, L. panamensis, and L. guyanensis might disseminate (metastasis) from the primary lesion to a distant mucosal site, leading to destructive secondary lesions especially in the nasopharyngeal areas. More rarely the musosal lesion might result by contiguity, for instance, skin lesion near the nasal or oral mucosa. This form does not evolve spontaneously to clinical cure, and if left untreated, develops to mutilation or destruction, affecting the quality of life of patients. In general, treatment failures and relapses are common in this clinical form [18,22,23].
In recent years, the relative proportion of mucosal leishmaniasis cases reported in the Americas is 3.1% among all the cutaneous leishmaniasis cases, however, depending on the species involved, genetic and immunological aspects of the hosts as well as the availability of diagnosis and treatment, in some countries that percentage is more than 5% as occurs in Bolivia (12-14.5%), Peru (5.3%), Ecuador (6.9-7.7%) and Brazil (5.7%) [24][25][26][27].
The diagnosis of CL is based on a combination of the epidemiological history (exposure), the clinical signs, symptoms, and the laboratory diagnosis which can be done either by the observation of amastigotes on Giemsa stained direct smears from the lesion or by histopathological examination of a skin biopsy. However, the sensitivity of the direct smear varies according to the duration of the lesion (sensitivity decreases as the duration of the lesion increases). Cultures and detection of parasite DNA through the polymerase chain reaction (PCR) can also be done but they are costly and their use is limited to reference or research centers. The diagnosis of mucosal leishmaniasis is based on the presence of a scar of a previous cutaneous lesion, which might have occurred several years before, and on the signs and symptoms. A positive Montenegro Skin Test (MST) and/or positive serological tests such as the immunofluorescent antibody test (IFAT) allow for indirect confirmation of diagnosis. Parasitological confirmation of mucosal leishmaniasis is difficult because the parasites are scarce and rarely found in tissue samples. Thus, histopathology not only is invasive but also demonstrates low sensitivity. This has led to the development of PCR techniques [28] which, though sensitive and specific, are still limited to research and reference laboratories.
Although pentavalent antimonial drugs are the most prescribed treatment for CL and ML, diverse other interventions have been used with varying success [29]. These include parenteral treatments with drugs such as pentamidine, amphotericin B, aminosidine and pentoxifylline, oral treatments with miltefosine, and topical treatments with paromomycin (aminosidine) and aminoglycosides. Other treatments such as immunotherapy and thermotherapy have also been tested.
The limited number of drugs available, the high levels of side effects of most of them, and the need of parenteral use, which may require hospitalization, and the fact that the use of local and oral treatment might increase patients' compliance, highlight the need of reviewing the current evidence on efficacy and adverse events of the available treatments for American cutaneous and mucocutaneous leishmaniasis.
To identify and include new evidence on the topic, we decided to update the Cochrane review published in 2009, which identified and assessed 38 randomized controlled trials also found a number of ongoing trials evaluating diverse interventions such as miltefosine, thermotherapy and imiquimod [29]. The objective of this paper is to present a systematic review which evaluates the effects of therapeutic interventions for American CL and ML.

Literature Search
We carried out a literature search to identify studies assessing the effects of therapeutic interventions for American CL and ML. Searched were planned to update findings of the Cochrane systematic review published in 2009 [29]. Structured searches were conducted in PubMed (January 2009 to July 2012), the Cochrane Library (number 6, 2012), and LILACS (January 2009 to July 2012) using a comprehensive list of key terms that were adapted to each database (Supporting Information S1. Search strategies). We searched the International Clinical Trials Registry Platform search portal of WHO (ICTRP) to identify past and ongoing trials using the key word ''leishma*. The references of both included and excluded material were examined in effort to find further relevant papers. We also completed a search in Scirus (limits: medicine, article title; July, 2012) to identify studies published in other databases. We reached out to authors and relevant key stakeholders to identify unpublished studies and related additional data from manuscripts. No language restrictions were applied.

Study and Information Selection
The titles, abstracts, and studies identified in the literature search were assessed by two reviewers. We included randomized clinical trials (RCT) which assessed the effects of interventions for treating CL and ML. Subjects having CL and/or ML or VL by clinical presentation and confirmed by histopathology, polymerase chain reaction (PCR) analysis or culture of lesions were included. We considered any intervention compared with no intervention, placebo, or other treatment regimens. Studies in which the intervention group included vaccines were excluded. All studies matching the inclusion criteria were reviewed by the authors and disagreement on inclusion was settled through discussion.

Data Extraction and Outcomes
At least two reviewers (ANM-E and LR) independently extracted the relevant data using a predesigned data extraction form; disagreements between reviewers were resolved by referring to a third author. Taking into account that a Cochrane review assessed and extracted data from previously published trials, we focused our assessment on updating provided evidence. Therefore, we designed a data collection form to systemically extract data from RCTs published later than previous the Cochrane review. The authors examined retrieved papers, identified, and recorded the main characteristics of the study including: qualitative aspects (such as date of publication, study design, geographical location and setting, population description, selection criteria, patient samplings, and funding source), characteristics of participants (age, sex, ethnicity, socioeconomic status), species of causative Leishmania, interventions (i.e. type, duration, method used to measure) and outcomes (type of outcome, outcome assessment method, type of statistical analysis, adjustment variables) and the risk of bias.
Clinical and/or parasitological cure at least three months after the end of treatment were the main outcomes considered in the review regardless of the microbiological method used to diagnose leishmaniasis. We defined cured as disappearance of all inflammatory signs (either skin edema or hardening, or both), and the occurrence of scarring or epithelialization of in ulcerative lesions [29]. We also extracted data on recurrence; the degree of functional and aesthetic impairment and/or prevention of scarring; emergence of resistance; and mortality. We also included those adverse events reported in RCTs and did not search for additional adverse event studies or records. Findings are presented according to categories that were pre-specified by the trial.
We performed an evaluation on the risk of bias for each new identified trial following the Cochrane Collaboration tool for the assessment of these variables [30]. We also extracted information on inclusion and exclusion criteria; sample size calculation; and baseline comparability of age, gender, relevant clinical characteristics, and diagnoses. We registered data in the studies' table (Table 1). When necessary, authors were contacted to obtain additional information about their studies.

Statistical Analysis
We present a summary of main findings from the Cochrane review as well as an update of the evidence provided by new identified trials. We used the RevMan 5.1 software from the Cochrane Collaboration to perform the statistical analysis. For dichotomous primary outcomes the results, expressed as relative risk (RR) and 95% confidence intervals (CI), were calculated using the Mantel-Haenszel random effects model. For the pooled analysis we calculated the I square (I 2 ) statistic that describes the percentage of total variation across studies attributed to heterogeneity [30]; low, moderate, and high levels of heterogeneity are roughly estimated as I 2 values of 25%, 50%, and 75%, respectively. PRISMA checklist is included as supplementary file (Supporting Information S2).

Assessment of Risk of Bias
Overall the quality of the reporting and design of the RCTs was moderate to good (Table 3). Nine out of ten RCTs were judged as having low risk of bias for sequence generation; only one was considered having unclear risk of bias [77]. Five RCTs had low risk of bias for allocation concealment [70,71,75,76,81]. Two studies were placebo controlled trials The majority of trials provided a sample size framework and a scientific rationale for the sample size determination [70][71][72][73][74][75][76].
Anthelminthic therapy versus placebo (pentavalent antimony in both arms). One study [72] found no significant differences in overall time to cure and clinical failure at 3 months between groups. Overall, adverse events (only grade 1 and 2 events were observed) were reported in 60% of patients in both groups.
Meglumine antimoniate vs pentamidine. We included one study that evaluated intravenous meglumine antimony compared with intramuscular pentamidine in Brazil [69]. The Cochrane systematic review identified two additional RCTs [32,40]. Meta-analysis of two RCTs found no significant differences between groups in the rate of complete cure after 6 months of follow-up; however, statistical heterogeneity was very high (I 2 :90%). One RCT [32] found that meglumine antimoniate was superior to pentamidine in the rate of complete cure in the treatment of L. braziliensis (80 particpants, ITT RR 2.21 95% CI: 1.41-3.49), while another RCT [69] assessing L. guyanensis did not find any significant difference. Another RCT [40] also did not found any significant difference in the rate of failure between    Table 3. Cont.
Interventions for Leishmaniasis: A Review treatment groups after one year of follow-up (L. braziliensis). No significant differences between groups were found when assessing serious adverse events.
Nitric oxide releasing patch vs meglumine antimoniate. One study [81] found a significant difference favoring meglumine antimoniate in complete cure (143 participants; RR: 0.40; 95% CI: 0.29 to 0.55). No significant difference was found for serious adverse events. Patients treated with nitric oxide releasing patch presented a significant lower proportion of non-serious adverse events such as fever, headache, myalgia, and arthralgia while those treated with meglumine antimoniate had a significant lower proportion of local rash and pain. Thermotherapy versus meglumine antimoniate. A significant difference favoring meglumine antimoniate was found in the rate of complete cure at 6 months (189 participants; RR: 0.80; 95% CI: 0.68 to 0.95) [71]. However, no significant difference was found when analyzing L. panamensis (37 participants; RR: 0.81; 95% CI: 0.54 to 1.21) and L. braziliensis (65 participants; RR: 0.80; 95% CI: 0.59 to 1.10) [71]. An RCT [53] from the Cochrane review found no significant difference between groups.
A summary of the main findings from the previous Cochrane review and the updated studies can be found in Table 4.

Discussion
The present update on the treatment of American CL showed an increase in the number of papers published during the past 3 years and an improvement of the quality of the studies. The 10 RCTs included in this systematic review represent more than 25% of the 38 studies included in the Cochrane Review 2009, which covered a period of 25 years. The scope of this systematic review took into consideration the main challenges faced in this specific therapeutic field, mainly, the parasite diversity observed across the continent, the variety of therapeutic interventions currently in use and the importance of the quality of the included reports which were submitted to a stringent evaluation of risk of bias.
Miltefosine, considered as the first effective oral treatment for cutaneous leishmaniasis has been used at a dose of 2.5 mg/kg weight, with cure rates which vary both according to species and to the geographic location where the studies have been performed [85]. Adverse effects include vomiting, nausea, kinetosis and headache, and elevation of creatinine and aminotransferase levels [86]. Meglumine antimoniate has been widely used for the treatment of ACL. The currently recommended dose is 20 mg/kg of body weight/day for 20 days [5]. For these reasons, meglumine antimoniate is frequently used as the comparator in clinical trials of new treatments for ACL [86].
The attempt to summarize the effect of miltefosine compared to meglumine antimoniate included four studies and indicated no  difference between treatments [70,[73][74][75]. However, the heterogeneity measure was high and a careful observation of data showed that one study, the one with the largest sample size [75], demonstrated an inconsistent effect compared to the other three [70,74,87]. This inconsistent study concluded with statistical significance for the inferiority of miltefosine in patients infected with L. braziliensis or L. panamensis and the other three concluded for the lack of difference between both drugs. Furthermore, the two studies conducted in Brazil included patients with just one parasite species in each study, L. braziliensis [74] and L. guyanensis [73]. However, the apparent consistency between the Brazilian studies needs to be taken with caution because of the differences of  therapeutic response rate observed between L. braziliensis and L. guyanensis infected patients as previously reported in a quasiexperimental study [88].The Colombian studies are more difficult to analyze because of the inclusion of a mixture of patients infected with at least two parasite species in each study [70,75]. Just one species, L. panamensis, was common to both studies. The therapeutic response variation observed in patients from different geographical areas could be at least partially explained by the diversity of parasite species that was well documented in those four trials but further evidences are needed to conclude that parasite species was determinant of the therapeutic response and homogeneous across different geographical regions [5]. Recently, data from Peru stimulated the debate on the role of parasite species on therapeutic response to antimonials and raised other hypotheses to explain the differences observed between Peruvian and Brazilian cases infected with L. braziliensis or L. guyanensis [89].
Overall, caution needs to be applied to the summary estimates related to the comparison of miltefosine with meglumine antimoniate and these data deserve proper contextualization for each of the specific scenarios where the evidences were produced.
Anthelminthic therapy versus placebo, both associated with standard treatment with antimonials, was evaluated in just one RCT showing no significant difference [72]. The rational for this approach is based on the possible influence of helminths parasites on the modulation of the immune response against leishmaniasis. This study deserves attention and further investigation because of the small sample size and the unexpected worse response observed (although not statistically significant) in the group submitted to anthelminthic therapy.
Pentamidine isethionate has been used for the treatment of cutaneous leishmaniasis at a dose of 2-4 mg/kg/day with 2 to 4 applications on alternate days. Frequent adverse effects include musculoskeletal pain, anorexia, abdominal pain, nausea, vomiting, headache, asthenia and fatigue. Pentamidine can also cause hypoglycemia, which can sometimes be severe, and insulindependent diabetes mellitus [86,90].
The attempt to meta-analyze the evidences that compared pentamidine against antimonials was troublesome because of the high heterogeneity observed which again could be due to the diversity of parasite species causing the diseases and the geographical variation in therapeutic response already mentioned above.
Imiquimod stimulates the production of nitric oxide by macropahges, which decreases the number of parasites in vitro [85]. Topical imiquimod has been used in combination with pentavalent antimonials for treatment of cutaneous leishmaniasis [35,52]. In spite of the lack of significant differences observed for the six-month and 12-month outcomes the result of the summary estimate indicating the benefit of imiquimod improving the initial response obtained at the three months with antimonials deserves more attention. Sometimes the time to obtain clinical cure is not included at least as a secondary endpoint to evaluate the usefulness of therapeutic interventions. However reducing time-to-cure period could be interesting to save costs for the health system.
The included evidence on nitric oxide is not encouraging to put more resources in such type of intervention and could be considered as a proof of principle of lack of therapeutic effect.
The only new evidence on thermotherapy demonstrated the inferiority of this intervention in patients infected with L. panamensis or L. braziliensis. Subgroup analysis was strongly affected by the small sample size, but both subgroups maintained a consistent direction of the effect with point estimates favoring antimonial treatment.
The rationale for administering local treatments, which include thermotherapy, intralesional administration of pentavalent antimonials, and topical treatments, is that the risk of developing a mucosal form is low, not necessarily prevented by systemic treatment, and localized treatments are better tolerated and have less frequent and severe adverse effects as compared to systemic treatments [91]. However, there is a need to standardize and evaluate the efficacy of localized and topical treatments for cutaneous leishmaniasis and to develop recommendations for their use.
This study also confirmed the lack of RCTs in patients with the mucosal form of the disease. This is a relevant ''negative'' result which is widely recognized as a pitfall in the health care offered to patients with leishmaniasis. This is a neglected aspect that remains waiting for the organization of a multicentric initiative to develop RCTs to prove the efficacy of the current therapeutic options.
The main limitation of the present systematic review is the lack of a larger number of studies to perform the meta-analysis taking into consideration the already known characteristics which affect the prognosis of CL. Small sample size is still a problem in RCTs for leishmaniasis. Meta-analysis could help if the scenarios were homogeneous but this is not the case as already demonstrated for different parasite species and the geographical variation in response to treatment. In addition, some RCT s had short follow-up periods in which relapses could easily be missed in a chronic condition like MCL/CL. Future studies should consider longer follow up period [92].
Data on parasite species appears to be more commonly registered in recent trials but some patients are still included and treated without this information. The need for species-specific parasitological diagnosis of patients enrolled in clinical trials needs to be stressed. Recent development of molecular tools allows species identification with some effort but fortunately, nowadays, the success in parasite isolation and culture, which is more troublesome, is no longer required [93].
Although there are no simple and cheap assays to evaluate parasite resistance in vitro and that such type of evaluation requires parasite isolation, it would be reasonable to nest into RCTs a subgroup analysis of resistance to the specific drugs which are under evaluation, mainly in those scenarios where lack of therapeutic response is high [94,95].
Finally, the lack of registry of other potential prognostic factors such as immunological status [96], co-morbidity, [72] age-related pharmacokinetics) [97] which could explain the observed differences between studies and regions deserves more attention and a minimum set of variables with prognosis potential needs to be discussed for further trials in order to enrich the comprehension of the observed variability. This could be as important as the use of standardized outcomes and time to main and secondary outcomes.
In conclusion the present updated systematic review revealed that a lot of work needs to be done to achieve a strong evidence to recommend specific treatments against cutaneous leishmaniasis. There is still a need for well conducted RCT to assess the effectiveness and safety of different anti-Leishmania alternatives drugs. As compared with the Cochrane review [29], studies included in this update had lower risk of bias and reported information in a more standardize manner. Local or regional evidences should be obtained taking into consideration parasite species diversity and other prognostic factors to make valuable evidence-based recommendations. Mucosal disease is the more neglected form of CL and a multicentric trial should be urgently considered.