Methods for the Classification of Leprosy for Treatment Purposes

Abstract

The World Health Organization advocates 2 leprosy treatment regimens on the basis of disease classification (as multibacillary or paucibacillary) by skin lesion count. This method, which, in the Philippines, results in a high prevalence (78%) of patients with multibacillary leprosy, was directly compared with classification using standard histopathological and microbiological criteria in 264 currently untreated patients with leprosy. Of those whose leprosy was classified as paucibacillary, 38%–51% of patients had multibacillary leprosy according to classic criteria and were thus at risk of undertreatment according to World Health Organization recommendations.

Leprosy, a chronic disease caused by the acid-fast bacillus Mycobacterium leprae, presents a spectrum of clinical, bacteriological, immunological, and dermatopathological characteristics. In the 1960s, Ridley and Jopling [1] proposed a histological classification scheme for leprosy that ranged in severity, beginning with early indeterminant (I) leprosy and continuing with polar tuberculoid (TT) leprosy, borderline tuberculoid (BT) leprosy, mid-borderline (BB) leprosy, borderline lepromatous (BL) leprosy, and polar lepromatous (LL) leprosy. This scheme has the virtue of having a significant concordance between clinical, immunological, and pathological features, and its subdivisions correlate with the number of acid-fast bacilli present in the dermis—generally expressed on a logarithmic scale by the bacteriologic index (BI), for which scoring ranges from 0 to ≥6. Furthermore, the exact placement of individual patients in the spectrum of disease assists clinicians in determining the required intensity of antimicrobial therapy, the prospects for the development of reactional states, and whether a protective Th1 or Th2 response predominates.

In 1982, the World Health Organization (WHO) [2] advocated the use of 2 different regimens of multidrug therapy for the treatment of leprosy. Treatment regimens were originally assigned on the basis of the Ridley-Jopling classification [1], which defined I, TT, and BT cases of leprosy as being paucibacillary (PB), and BB, BL, and LL cases of leprosy as being multibacillary (MB). Under this method of classification, a BI value ≥2 at any skin site indicated therapy for MB leprosy and a BI value <2 indicated therapy for PB leprosy. By 1988, a positive skin-smear result at any site became sufficient to indicate treatment for MB leprosy [3]. Because of lack of availability and questions regarding the reliability of histopathological analysis and skin-smear results, as well as because of a desire to make multidrug therapy available in areas that do not have a substantial medical infrastructure and to rapidly facilitate the campaign to eliminate leprosy as a public health problem, the WHO has recently advocated the field-friendly method of counting skin lesions to determine whether patients should be treated for PB or MB leprosy (PB leprosy, ≤5 lesions; MB leprosy,>5 lesions) [4]. Previously, it was recommended that adult patients with MB leprosy undergo treatment for a minimum of 2 years or until negative skin-smear results were achieved [2]. The recommendation is now for 1 year [5] of daily unsupervised administration of dapsone (100 mg) and clofazimine (50 mg) therapy and monthly supervised administration of rifampin (600 mg) and clofazimine (300 mg) therapy. Patients with PB leprosy are treated for 6 months with daily unsupervised dapsone (100 mg) therapy and supervised monthly rifampin (600 mg) therapy.

This present study, which was performed in an area of leprosy endemicity in the Philippines that has a high prevalence of cases of MB leprosy, evaluates for the first time the concordance of clinical classification of leprosy by lesion counting (emphasizing MB leprosy) with both dermatohistopathological analysis and the 2 skin-smear criteria previously used by the WHO.

Patients and methods. The study population consisted of 284 patients with active leprosy who presented to the clinical branch of the Leonard Wood Memorial Leprosy Research Center (Cebu, Philippines) from July 2003 to July 2005. Of these patients, 281 had not previously received any treatment for leprosy, 20 had received some rifampin for a prior diagnosis of tuberculosis, and 3 claimed to have received some leprosy treatment a minimum of 7 years earlier. This study was approved by a local institutional review board licensed by the National Institutes of Health, and written, informed consent was obtained from each patient. Patients ranged in age from 5 to 70 years (mean age, 34 years). A total of 184 patients were male, and 80 were female. Patients underwent a throrough physical examination, and skin lesions were counted. A skin biopsy sample from the most active lesion and 6 skin smears were obtained from each patient. Patients were considered to have MB leprosy and were treated for 1 year with the WHO-recommended multidrug therapy regimen if disease was histologically determined to be BB, BL, or LL leprosy or if any skin-smear site had a BI value ≥2. Patients were considered to have PB leprosy and were treated accordingly if disease was histologically determined to be I, TT, or BT leprosy and if no skin lesion had a BI value ≥2.

Results of classifications by lesion counting were compared using dermatopathological classifications, as well as skin-smear positivity and a BI value ≥2 at any site. For statistical analysis, we used the χ² test to compare proportions, and we determined sensitivity and specificity using standard procedures [6].

It is most noteworthy that, of the 58 patients who had ≤5 skin lesions, 22 (38%) had leprosy that was histologically determined to be BB (1 patient), BL (16 patients), or LL (5 patients) (table 2). The actual number of skin lesions and the average BI value of 6 skin-smear sites in those patients with BL and LL leprosy who had ≤5 skin lesions are presented in table 3. The 1 patient with BB leprosy who had <6 skin lesions had 4 lesions and a BI value of 0. Also, 33 (57%) of the patients with ≤5 lesions were skin-smear positive at ≥1 site, and 18 (31%) had a BI value ≥2 at ≥1 site. The criteria of skin-smear positivity at any site resulted in a significantly greater frequency (57%) of patients whose leprosy was classified as PB by lesion counting who were considered to have MB leprosy by Ridley-Jopling classification (38%; P = .04) or by a BI value ≥2 (31%; P = .005). In these patients, there was no significant discrepancy in disease classification found between histological analysis and a BI value ≥2 at any site.

Table 2

Classification of multibacillary (MB) and paucibacillary (PB) leprosy by lesion count and dermatopathological analysis.

Table 3

Number of skin lesions and average bacteriologic index (BI) of patients with histologically borderline lepromatous (BL) and polar lepromatous (LL) leprosy with paucibacillary (PB) lesion counts.

The specificity of counting lesions for the purpose of classification of MB leprosy was low: it was 53% for histopathological analysis, 52% for skin-smear positivity, and 50% for skin-smear BI values ≥2. Sensitivity for this method was better: it was 89% for histopathological analysis and 85% for both skin-smear criteria (table 1).

Table 1

Sensitivity and specificity of classifying leprosy cases as multibacillary (MB) on the basis of lesion count versus histopathological analysis and and 2 skin-smear criteria.

For those who were classified as having MB leprosy by lesion count, a smaller but statistically significant percentage were found to be discordantly classified by Ridley-Jopling criteria as having TT or BT leprosy (32 [15%] of 206 patients), being skin-smear negative (23 [11%] of patients), and having no skin-smear BI value ≥2 (40 [19%] of 206 patients). Of these 3 criteria, only the proportion of patients who had a skin-smear BI value ≥2 was determined to be statistically significantly different (greater) that the proportion of those who were skin-smear positive (P = .02).

Discussion. The current study is the first to be performed in a location with a high prevalence of patients with MB leprosy to compare disease classification by the WHO's currently recommended method of lesion counting with the dermatopathological classification scheme of Ridley- Jopling and 2 previously used WHO classification criteria (a positive BI value and a BI value ≥2 at any site). This study has demonstrated that classification of leprosy through the counting of lesions in a geographical region that has a preponderance of patients with BB, BL, and LL leprosy and a high bacterial burden results in the assignment of a high percentage (31%–58%) of patients to PB leprosy treatment regimens, even though they had been considered to have MB leprosy by previously implemented criteria.

Our findings are not entirely surprising because, although patients with BB, BL, and LL leprosy and patients with a high BI value generally have multiple skin lesions, such patients with 1 or a few skin lesions have been recognized as having the PB form of the disease [7]. Fortunately, in the Philippines, we still classify patients as having MB leprosy if disease is dermatopathologically determined to be BB, BL, or LL or if the BI value of any skin-smear site is ≥2.

The present study has some precedent. Oskam [8] reviewed previous studies [9–14], mostly prior to 1993, that compared the number of skin lesions (although not always with>5 as the cutoff) and body zones involved with skin smears or biopsies with a BI value>0–1 (table 4). The first 3 studies resulted in a low specificity (39%–42%) for a clinical diagnosis of MB leprosy in locations where MB leprosy is more common [9–11] but not in areas where it is less common (81%–88%) [12, 13]. Misclassifying patients with PB leprosy as having MB leprosy may result in overtreatment, whereas misclassifying patients with MB leprosy as having PB leprosy may result in undertreatment. The former places a greater burden on health services, whereas the latter results in a risk of undertreatment and treatment failure.

Table 4

Sensitivity and specificity of classifying leprosy patients by certain clinical criteria compared with bacteriological methods.

There is evidence from 3 studies [15–17] that double-digit relapse rates occur in patients who have MB leprosy as defined by histological analysis and a BI value ≥2, and who are treated with multidrug therapy for 2 years; these rates of relapse are almost entirely confined to patients with BL or LL leprosy who have a high BI value [15–18]. One study [16] found that prolonging therapy until smear results are negative (a mean of 5 years) decreases the relapse rate from 17% to 4%. Furthermore, the WHO currently recommends only 1 year of multidrug therapy for patients with MB leprosy. A greater concern arising from the present study is that the current practice of counting lesions for the purpose of assigning treatment regimens would have resulted in many patients being assigned to PB leprosy treatment regimens who, by previous methods, would be treated with longer and more intensive MB leprosy treatment regimens. Misclassification, such as we found as a result of lesion counting, also has the potential to result in drug resistance. This is of considerable concern, because, in Thailand [13], the risk of relapse has been observed to be highest in patients with MB leprosy who are wrongly classified as having PB leprosy and are, therefore, undertreated. Thus, it is recommended that, particularly in areas with a high frequency of patients with MB leprosy, skin-smear sample testing and biopsy sample analysis be reinstituted and used to classify patients for treatment purposes.

Acknowledgments

Most of the patients included in this study were recruited under the study “Endocrine Changes as Risk Factors for Leprosy Reactions.”

Financial support. The Leonard Wood Memorial and the American Leprosy Missions.

Potential conflicts of interest. All authors: no conflicts.