Can baseline ML Flow test results predict leprosy reactions? An investigation in a cohort of patients enrolled in the uniform multidrug therapy clinical trial for leprosy patients in Brazil

Background The predictive value of the serology to detection of IgM against the Mycobacterium leprae-derived phenolic glycolipid-I/PGL-I to identify leprosy patients who are at higher risk of developing reactions remains controversial. Whether baseline results of the ML Flow test can predict leprosy reactions was investigated among a cohort of patients enrolled in The Clinical Trial for Uniform Multidrug Therapy for Leprosy Patients in Brazil (U-MDT/CT-BR). Methods This was a descriptive study focusing on the main clinical manifestations of leprosy patients enrolled in the U-MDT/CT-BR from March 2007 to February 2012 at two Brazilian leprosy reference centers. For research purposes, 753 leprosy patients were categorized according to a modified Ridley-Jopling (R&J) classification and according to the development of leprosy reactions (reversal reaction/RR and erythema nodosum leprosum/ENL), and whether they had a positive or negative bacillary index/BI. Results More than half of the patients (55.5 %) reported leprosy reaction: 18.3 % (138/753) had a RR and 5.4 % (41/753) had ENL. Leprosy reactions were more frequent in the first year following diagnosis, as seen in 27 % (205/753) of patients, while 19 % (142/753) developed reactions during subsequent follow-up. Similar frequencies of leprosy reactions and other clinical manifestations were observed in paucibacillary (PB) and multibacillary (MB) leprosy patients treated with U-MDT and regular MDT (R-MDT) (P = 0.43 and P = 0.61, respectively). Compared with PB patients, leprosy reactions were significantly more frequent in MB patients with a high BI, and more patients developed RR than ENL. However, RR and neuritis were also reported in patients with a negative BI. At baseline, the highest rate of ML Flow positivity was observed in patients with a positive BI, especially those who developed ENL, followed by patients who had neuritis and RR. Among reaction-free patients, 81.9 % were ML Flow positive, however, the differences were not statistically significant compared to reactional patients (P = 0.45). Conclusions MB and PB patients treated with R-MDT and U-MDT showed similar frequencies of RR and other clinical manifestations. Positive ML Flow tests were associated with MB leprosy and BI positivity. However, ML Flow test results at baseline showed limited sensitivity and specificity for predicting the development of leprosy reactions. Electronic supplementary material The online version of this article (doi:10.1186/s40249-016-0203-0) contains supplementary material, which is available to authorized users.


Background
Leprosy, a complex chronic infectious disease caused by Mycobacterium leprae, primarily affects the skin and peripheral nerves and has a great potential to cause disability and irreversible deformities. Leprosy presents as a spectrum of manifestations: lepromatous leprosy (LL) is characterized by a high bacillary index, weak M. lepraespecific cell-mediated immunity (CMI), and high antibody titers. At the other extreme, tuberculoid (TT) patients have a low bacillary index, strong M. leprae-specific CMI, and weak antibody production. Immunologically unstable borderline tuberculoid (BT), borderline lepromatous (BL), and borderline-borderline (BB) forms lie in the middle of the spectrum combining features of both LL and TT forms [1].
During the chronic phase of leprosy, before diagnosis, during or after multidrug therapy (MDT), acute immune-inflammatory episodes, known as type 1 reactions or reversal reaction (RR) and type 2 reactions represented mainly by erythema nodosum leprosum (ENL), can lead to irreversible nerve damage [2,3]. Moreover, neuritis characterized by intense spontaneous nerve pain is often associated with leprosy reactions, but can also occur without any cutaneous involvement [4]. Inappropriate therapeutic management of neuritis may result in permanent nerve function impairment [5,6].
The World Health Organization (WHO) has been recommending MDT since 1981 [7]. Later, in 1988 a clinical classification based on the number of skin lesions was used to define two different MDT regimens [8] for either paucibacillary (PB) patients presenting up to five skin lesions or for multibacillary (MB) patients presenting more than five skin lesions. Multibacillary leprosy patients are prescribed 12 supervised monthly doses of rifampicin, dapsone, and clofazimine, plus selfadministered daily doses of dapsone and clofazimine. Meanwhile, PB leprosy patients are treated with six supervised monthly doses of rifampicin and dapsone, plus self-administered daily doses of dapsone [7]. Up until now, the total person-time observed is 780 930 person-days, i.e. 2 139.5 person-years, with a maximum of 6.66 years of follow-up [9].
Serology to detect immunoglobulin M (IgM) antibodies against the M. leprae-derived phenolic glycolipid-I (PGL-I) has been widely evaluated for leprosy classification, including its simple and rapid immunochromatographic ML Flow test which detects IgM to PGL-I in 10 min [10]. Studies have shown that the results of anti-PGL-I serology reflect the bacillary index (BI) of leprosy patients, with MB patients showing a high positivity rate (70 %-90 %) and PB patients showing low seropositivity (15 %-40 %) [11][12][13]. Conflicting data have been reported about the predictive value of anti-PGL-I serology for identifying patients who are at a higher risk of developing leprosy reactions. High levels of anti-PGL-I antibodies at diagnosis or after treatment have been associated with a higher risk of developing leprosy reactions [14][15][16][17], however, other studies have shown similar anti-PGL-I levels among reactional and reaction-free patients [18][19][20]. Therefore, whether anti PGL-I serology can be used to identify patients at risk of developing leprosy reactions remains unknown.
It is important to identify the risk factors for developing leprosy reactions in order to establish preventive strategies for reducing irreversible nerve damage and sequelae [21]. The present study describes the main clinical manifestations including leprosy reactions in a cohort of patients enrolled in the U-MDT/CT-BR. In addition, the baseline results of the ML Flow test for predicting the development of leprosy reactions observed during follow-up were evaluated.

Study population
This was a descriptive study focusing on the main clinical manifestations of leprosy patients enrolled in the U-MDT/CT-BR from March 2007 to February 2012 at two Brazilian leprosy reference centers: Dona Libânia (Fortaleza, Ceará state, northeast region) and Alfredo da Matta (Manaus, Amazonas state, north region), according to previously described rationale and study design [22].
Briefly, for the U-MDT/CT-BR study, following the WHO operational classification [8], patients were randomized into four groups: PB patients treated with U-MDT, PB patients treated with R-MDT, MB patients treated with U-MDT, and MB patients treated with R-MDT. In the R-MDT cohort, MB patients with six or more lesions were treated with 12 supervised monthly doses of rifampicin, dapsone, and clofazimine, plus daily self-administered doses of dapsone and clofazimine, while PB patients presenting five or fewer lesions received six supervised monthly doses of rifampicin and dapsone, plus daily self-administered doses of dapsone [23]. In the U-MDT cohort, six supervised monthly doses of rifampicin, dapsone, and clofazimine, plus daily self-administered doses of dapsone and clofazimine were prescribed for both PB and MB patients. For research purposes, patients were also categorized according to a modified Ridley-Jopling (R&J) classification system taking into account clinical features, histopathological results of skin biopsies, and the BI on slit skin smear; Mitsuda tests and BI of the skin biopsy were not performed. This modified R&J classification taking into account the development of reactions was used to write this paper.
Our study group included 753 out of 859 patients enrolled in the U-MDT/CT-BR; 12.3 % (106/859) of patients were excluded due to: incomplete treatment (n = 9), histopathological diagnosis not compatible with leprosy (n = 3), patients requested to leave the study (n = 8), death (n = 19), transfer of patients to other sites (n = 9), patients lost to follow-up (n = 12), adverse drug reactions (n = 25), other concomitant illnesses (n = 5), leprosy relapse (n = 1), or other reasons (n = 15). The patients' clinical information was obtained from the case report form, at baseline, and during follow-up (from March 2007 to September 2013). This included a monthly medical visit in the first year and a yearly medical visit in the following 6 years; clinical monitoring is still taking place. As part of protocol, at enrollment, all patients were advised about the signs and symptoms of leprosy reactions and nerve function impairments, and advised to immediately return to the clinic in case any abnormality occurred.

Case definitions for leprosy reactions and neuritis
A RR was defined as an acute inflammation of preexisting cutaneous lesions with or without the appearance of new lesions. The clinical diagnosis of ENL was based on the acute appearance of erythematous nodular skin lesions, tender to touch or painful in the absence of external stimuli, accompanied by fever with or without peripheral nerve pain and dysfunction. Neuritis was defined as spontaneous nerve pain or by the presence of thickened nerve trunks associated with loss of sensation and motor and autonomic deficits. Study groups included reaction-free patients, patients developing a RR, patients with ENL, and a group of patients with unusual clinical manifestations such as necrotizing ENL, mixed RR-ENL reactions, polymorphic erythema, arthritis, lymphadenopathy, orchitis, iritis, iridocyclitis, and reactional hands and feet. Dermatologists with vast expertise in leprosy diagnosis, treatment, and patients' management were in charge of the clinical diagnosis and followup of leprosy reactions. Patients who developed leprosy reactions or who had impaired nerve function received appropriate treatment according to the guidelines of the National Leprosy Control Program/Brazilian Ministry of Health and remained in the study; RR, ENL, and associated clinical manifestations were registered in the case report form (CRF).

ML Flow serological tests
The ML Flow test was performed, as previously described elsewhere [10]. Briefly, 5 μl of whole blood and running buffer were added to the sample well and visual readings were performed 10 min later. Results were recorded according to the scale of positivity: 0 and 0.5 = negative; 1 +, 2 +, 3 +, and 4 + = positive [10].

Statistical analyses
Leprosy patients were stratified into subgroups, as either patients who had a reactional episode or reaction-free patients. Episodes of RR and ENL were analyzed individually or associated with neuritis. Statistical significance was assessed using the chi-square (χ2) test to compare the reactional and reaction-free groups. Results were considered statistically significant when P-values were <0.05. Descriptive analyses of the clinical, epidemiological, and laboratory variables and box plots were performed using SPSS version 21 (IBM, New York, USA). Receiver operating characteristic (ROC) curves were determined with 95 % confidence intervals (CIs) (using GraphPad Prism Software version 5, CA, USA).

Ethical considerations
The UMDT/CT-BR was performed under international (Helsinki) and Brazilian research regulations regarding human beings and was approved by regional ethics committees from all states involved in the study, as well as by the National Commission for Ethics in Research of the National Health Council/Ministry of Health (February 17, 2006, protocol number 001/06). Written informed consent was obtained from all patients prior to their inclusion in the study. For patients aged below 18 years, written parental consent was obtained. Data confidentiality was strictly guaranteed and all patients knew they were free to leave the study and opt for the R-MDT regimen outside of the study at any time (ClinicalTrials.gov identifier: NCT00669643).  Table 1). Most MB patients were male, and MB patients who developed ENL had the lowest median age in MB group (33 years).
In this cohort, 55.5 % (418/753) of patients had at least one episode of a leprosy reaction during follow-up. The distribution of the first leprosy episode according to R&J groups is presented in  Table 4

ML Flow test results at baseline
At baseline, higher seropositivity in the ML Flow was recorded among patients with a positive BI as compared with patients with a negative BI (86 %; 366/426 versus 29.1 %; 95/327) (P <0.001). The baseline serologic profile of patients evaluated by the ML Flow test showed higher positivity in patients who developed reactions (77 %; 322/418) as compared with reaction-free patients (41.2 %; 138/335) (P <0.001) (see Fig. 1) (see also Additional file 2: Table S1).
At baseline, the highest rate of ML Flow positivity and the highest colour intensity of readings were recorded, as expected, among patients with a positive BI from the following categories: patients who developed ENL or ENL associated with neuritis (92 %, 44/48; mean ML Flow reading score: 2.7), followed by patients who presented with neuritis (86.2 %, 81/94; mean ML Flow reading score: 2.6) and patients who developed a RR (84.7 %, 111/131; mean ML Flow reading score: 2.4). Reaction-free patients had the lowest positivity (81.9 %, 86/105; mean ML Flow reading score: 2.4), however, differences among the groups were not statistically significant (P = 0.45). Among the patients with a negative BI, those who developed a RR or neuritis had a higher ML Flow positivity rate and higher colour intensity in test readings compared with reaction-free patients (RR: 45.7 %, 16/35, mean ML Flow reading score: 0.9; neuritis: 42.6 %, 23/54, mean ML Flow reading score: 0.8; reaction-free: 22.7 %, 52/229, mean ML Flow reading score: 0.4) (P < 0.001) (see Figs. 1 and 2). Among the patients who had a positive BI, the strength of the ML Flow signal was similar in patients who had a reaction (ENL, RR, or neuritis) and reaction-free patients (P > 0.05), while among the patients who had a negative BI, a higher signal strength was observed in patients who developed RR as compared with the reaction-free patients (P = 0.003) (see Fig. 2).
The accuracy of the ML Flow test at baseline to predict whether a patient will develop leprosy reactions was analyzed using the ROC curve. For patients who had a negative BI and developed a RR during follow-up or who did not have a reaction, the area under the curve was <0.7. Establishing a specificity of 90 % (95 % CI: 85-93 %), the sensitivity was 28 % (95 % CI: 14-46 %) for a reading score of ≥2. Similar results were observed in patients who had a positive BI and developed ENL or RR (see Additional file 3: Figure S1).

Discussion
For many years, leprosy patients were treated with dapsone monotherapy, and the development of ENL was reported in half of LL patients and in one-quarter of BL patients [24]. The introduction of MDT in the 1980s reduced the frequency and the severity of ENL, probably as a result of the anti-inflammatory activity of clofazimine  that was included in the treatment regimen [25]. However, previous data from the U-MDT cohort showed that there was no statistical difference in the frequency of leprosy reactions among MB patients receiving R-MDT or U-MDT, indicating that six or 12 months of clofazimine treatment does not result in the prevention of leprosy reactions [26].
In the U-MDT/CT-BR, until 2014, 780 930 person-days, i.e. 2 139.5 person-years, with a maximum of 6.66 years of follow-up were evaluated [9]. In the current study, a comparable frequency of leprosy reactions and other clinical manifestations was observed in PB patients being treated with U-MDT or R-MDT. Findings from a previous report together with findings from the current investigation indicate that whether patients receive U-MDT or R-MDT does not have any impact on the incidence of leprosy reactions and other clinical manifestations such as neuritis.
In this cohort of leprosy patients, composed mainly of MB patients (~80 %), more than half developed at least one episode of a leprosy reaction. Moreover, the great majority of patients who developed reactions were MB patients with high BIs, and these variables have been previously associated with a high probability of developing leprosy reactions [2,27,28]. Our results confirm that MB patients have a higher risk of developing leprosy reactions as compared with PB patients. In addition, the study confirms that the most common reaction in both  Overall, more than half of the patients had a positive BI and reactions were more frequent in these patients as compared with patients who had a negative BI. The development of reactions was more common in the first year of follow-up compared to subsequent years. Our results are in accordance with other studies, which showed the predominance of RR over ENL and the higher incidence of reactions in MB patients with a positive BI as compared with patients with a negative BI [29,30].
In this cohort, RRs were seen mainly in BL MB patients, a finding which is consistent with other studies [2,14,31]. The majority of patients who developed ENL had LL and high BIs. Other studies have reported that ENL is more often seen in LL patients than in BL patients [32,33] and a higher BI is a known risk factor for developing ENL [34,35]. Interestingly, MB patients who developed ENL were younger than reaction-free MB patients. Accordingly, a study that evaluated risk factors for developing ENL showed that patients older than 40 years were less likely to develop ENL [34]. Another study showed that patients whose first leprosy symptom occurred during adolescence had a greater chance of developing ENL than patients whose onset of leprosy occurred after adolescence [36]. In this study, PB patients who developed RR were older than reaction free-PB patients. Other studies have shown that older age was an important risk factor for developing both RR at diagnosis and developing sequelae after treatment [37,38].
During MDT, bacillary death occurs resulting in a massive release of mycobacterial antigens favoring the formation of immune complexes, mainly among MB patients, and immune complexes are considered to play a role in the pathophysiology of leprosy reactions [39][40][41][42]. In this study, reactions were more frequent during the course of MDT, as previously reported [2,27,43]. However, patients mainly developed ENL during follow-up. The results of this study underline the importance of alerting patients about the possible development of reactions before, during, and even years after the conclusion of MDT, since reactions require immediate assistance and specific treatment to avoid irreversible nerve damage. Overall, neuritis was a common clinical manifestation that occurred in all clinical forms of leprosy, with higher frequencies observed in BT and BL forms. Similar to previously reported data, isolated neuritis with no other dermatological or clinical symptom was present in 20 % of cases, while less than half of the neuritis cases were associated with a reaction [44].
In this study, baseline positivity determined using the ML Flow test was associated with MB disease, a positive BI, and the development of reactions during follow-up, mainly ENL. Bacillary load is known to directly correlate with antibody levels and with the development of leprosy reactions [2,16,27,28]. Moreover, antibodies, which are abundant in MB patients, probably play a role in the pathophysiology of ENL [15][16][17]39]. However, little, if anything, is known about the role of antibodies in the development of RRs and other clinical manifestations such as neuritis. In this study, at baseline, patients with a negative BI who developed neuritis and RRs during follow-up had higher ML Flow positivity with a higher colour intensity compared with reaction-free patients who had a negative BI. Further studies on the immunopathogenesis of RRs are needed in order to clarify whether antibodies play a role in the development of RRs.
Our ROC analysis showed that the results of the ML Flow test at baseline had limited sensitivity and specificity to predict whether patients will develop leprosy reactions during follow-up (see Additional file 3: Figure S1). Additionally, high ML Flow seropositivity was not always associated with leprosy reactions and other clinical manifestations, as high positivity was also observed in reactionfree patients.

Conclusions
In the U-MDT/CT-BR, until 2014, 780 930 person-days, i.e. 2 139.5 person-years, with a maximum of 6.66 years of follow-up were evaluated, in this period MB and PB patients receiving R-MDT or U-MDT showed similar frequencies of RR, ENL and other clinical manifestations. Moreover, our data confirm that MB patients with a positive BI are more likely to develop leprosy reactions, and overall RRs are more common than ENL. Leprosy reactions were more likely to happen in the first year post-MDT, suggesting that special emphasis needs to be placed on clinical monitoring during this period. A significant proportion of patients presented neuritis alone or neuritis associated with reactions. ML Flow test results at baseline showed limited sensitivity and specificity for predicting whether patients will develop leprosy reactions during follow-up. Positive ML Flow tests were associated with MB disease, having a positive BI, and developing reactions, particularly ENL. As well as that, RR and neuritis were also reported in patients with a negative BI, and these patients had more positive ML Flow test results compared with the reaction-free patients who had a negative BI. Therefore ML Flow test results at baseline cannot be used as a predictive marker for leprosy reactions. Further studies to investigate other quantitative serological markers, besides anti PGL-I response, may help us to further understand the role of serology as a predictive tool for leprosy reactions.