ABSTRACT
Importance Leprosy is a stigmatizing, chronic infection which degenerates the nervous system and often leads to incapacitation. Multi-drug therapy (MDT) which consists of dapsone, rifampicin and clofazimine has been effective to combat this disease. In Indonesia, leprosy is still a problem. Furthermore, there had been reports of Dapsone Hypersensitivity Syndrome (DHS) which also challenges leprosy elimination in certain aspects. HLA-B*13:01 has been found to be associated with DHS and prospective screening has proven its ability to prevent DHS in the Chinese population, but has not been validated in Indonesians.
Objective To validate HLA-B*13:01 as a biomarker for DHS in the Indonesian population.
Design This is a case-control study.
Setting Population-based, multi-district recruitment from primary care centers in two of the top 3 most prevalent provinces in Indonesia, Papua and West Papua.
Participants Leprosy patients who presented themselves with DHS were recruited as case subjects (34 cases) and leprosy patients without DHS were recruited as control subjects (52 controls).
Exposure Leprosy patients who had undergone multi-drug treatment for leprosy under the standard WHO guideline, consisting of rifampicin, dapsone and clofazamine.
Main Outcome and Measures The association of HLA-B*13:01 to DHS based on difference in allele frequencies between cases and controls. HLA-B alleles were typed using the gold-standard Sequence Based Typing method. Results were analyzed using logistic regression and risk assessment was carried out.
Results The results of HLA-typing showed that HLA-B*13:01 was the most significant allele associated with DHS, with odds ratio=247.6 and P-value=4.81E-9, confirming the strong association of HLA-B*13:01 to DHS in the Indonesian population. The sensitivity of this biomarker is 91.2% and specificity is 96.2%, with an area under the curve of 0.95.
Conclusions and Relevance HLA-B*13:01 is validated as a biomarker for DHS in leprosy patients in Indonesia, and can potentially be a good predictor of DHS to help prevent this condition in the future.
Question Can HLA-B*13:01 be a predictor of Dapsone Hypersensitivity Syndrome in Indonesians?
Findings In this case control validation study, HLA-B*13:01 is significantly associated with DHS in Indonesians and can predict a person at risk of DHS 95% of the time.
Meaning HLA-B*13:01 is validated as a biomarker for DHS in Indonesians and can be useful to screen and prevent further incidences of DHS in the population.
BACKGROUND
In the era of leprosy elimination, many efforts have been done by the government to combat leprosy1. Multi-drug treatment (MDT) supplied by the World Health Organization is intensively applied as the main treatment of leprosy in every region in Indonesia. Although the availability of data is scarce, many cases of Dapsone Hypersensitivity Syndrome (DHS) were reported from districts in Papua and oftentimes are fatal cases.
Dapsone is an antibiotic and anti-inflammatory agent used for the prevention and treatment of infectious and chronic inflammatory diseases. As an antibiotic, dapsone is used to treat leprosy, malaria, actinomycetoma as well as Pneumocystis jirovecii pneumonia in persons with human immunodeficiency virus (HIV), whereas as an anti-inflammatory agent, it’s used to treat dermatitis herpetiformis, linear IgA dermatosis, subcorneal pustular dermatosis and erythema elevatum dilutinum2,3. Among all these diseases, leprosy is currently the one that most prevalently uses dapsone for treatment.
About 0.5-3.6% of persons treated with dapsone have drug hypersensitivity syndrome3-5, which was first described by Lowe and Smith6 in 1949 and termed “dapsone hypersensitivity syndrome” (DHS)7. The syndrome is a severe idiosyncratic drug reaction characterized by the clinical manifestations of fever, rash, and systemic involvement (most commonly of the liver and the hematologic system), which can cause severe organ dysfunction. DHS is usually manifested 4 to 6 weeks after the initiation of therapy.
With the introduction of MDT for leprosy worldwide and with the use of dapsone in chemoprophylaxis for P. jirovecii pneumonia in HIV infection combined with trimethoprim/sulfamethoxazole, the incidence of DHS might have also increased. On the basis of a recent systematic review of the published epidemiologic studies, the estimated global prevalence of the DHS is 1.4%, and the associated mortality is 9.9%8. This review also showed that 72.5% of the global DHS incidence actually originated in Asia and 71.8% of these DHS cases happened in leprosy patients. A recent literature review for DHS in China concluded a prevalence of 1.5% and mortality rate of 9.6%, very close to the global estimate for DHS9.
Previous study on the genetic determinant of DHS in 2,141 Chinese leprosy samples has discovered the risk factor for DHS10. They found that a single DNA variant, located in an immunity-associated gene called HLA-B, was significantly more common in those who developed DHS. Individuals with one copy of HLA-B*13:01 risk variant are 34 times more likely to develop DHS than individuals without. Two copies of HLA-B*13:01 makes individuals 101 times more susceptible to the syndrome. The results have been independently validated in various Asian populations including Thais11, Koreans12, Indians13 and Taiwanese14. Screening for this allele in new leprosy patients before the initiation of MDT has proven beneficial to reduce incidence rate of DHS in China to zero15.
In this study we aim to validate HLA-B*13:01 as a biomarker for DHS in the Indonesian population, report the allele frequency of HLA-B*13:01 in Indonesia, and confirm the sensitivity of HLA-B*13:01 as a predictor of DHS in this population.
METHODS
Study Design
This is a case-control genetic association study that collected retrospective leprosy patients who had survived DHS (‘cases’) and those whom did not exhibit DHS but has completed at least 6 months of multi-drug treatment for leprosy (‘controls’).
Study population
The study was carried out in main pockets district of Papua (Jayapura and Biak) and West Papua (Bintuni, Manokwari and Sorong) provinces in Indonesia.
Ethical Clearance
Ethical approval was obtained from National Institute of Health Research and Development with reference number LB. 02.01/5.2/KE.065/2016.
Subject Recruitment
The patients were recruited by inviting them to the point of care or visiting their homes. The informed consent was explained and given to the patient to read. The information sheet attached to the consent form contained the study aim, participant’s eligibility, risks and benefits of participation, as well as confidentiality of data. The patients who agreed to participate in the study signed the inform consent in front of a researcher, a health worker and one family member as witnesses.
Subject Recruitment Criteria
Leprosy patients treated with MDT who developed DHS symptoms were designated as ‘cases’. These patients exhibit symptoms as described by Richardus and Smith16 which include: (1) presence of at least two symptoms: fever, skin eruption, lymphadenopathy and liver abnormalities (elevated transaminases by at least 2 times); (2) symptoms appearing between the second and eighth week after the commencement of dapsone and disappearing upon discontinuation of the drug; (3) symptoms not ascribed to any other drug given simultaneously; (4) symptoms not attributable to leprosy reactions; (5) no other diseases liable to cause similar symptoms. Leprosy patients treated with MDT who did not develop any hypersensitivity reactions after 6-24 months of MDT treatment was designated as ‘controls’. Each case has at least one control obtained from the same area. The clinical information of the patients as well as the demographic and geographic origin were obtained. We took only one patient in each family based on their family name to ensure that all the samples were independent of each other and the genetic results we see are not due to familial segregation. The final number of cases and control that were collected are 34 cases and 55 controls, i.e. 14 cases and 28 controls from Papua province and 20 cases and 24 controls from West Papua province.
DNA extraction and Quantification
5 ml of EDTA anticoagulated venous blood was collected from all participants. The blood tubes were stored at -20°C and transferred in chilled (4°C) condition to Institute of Research and Development for Biomedicine Papua (IRDBP). At IRDBP, genomic DNA was extracted from the blood lymphocytes by a standard procedure using QIAGEN QIAamp DNA blood maxi kit (Catalog No. 51194). We modified the procedure from the manual to get the optimum yield, i.e. prolonged the incubation period on 70°C for 2 hours instead of 10 minutes. The DNA yield was quantified using Nanodrop™ Spectrophotometry.
HLA Sequence Based Typing (SBT)
All of the samples were typed using a research-use-only (RUO) version of an approved (CE-IVD) HLA-typing kit commercially available from TBG Diagnostics Limited, which is called HLAssure SE B Locus SBT Kit (Catalog No. 50210). This kit is a direct sequence-based typing approach, which is done by Sanger sequencing and considered as gold-standard for HLA typing. The kit is designed to detect all known polymorphic positions in exons 2 and 3 of HLA-B, which are necessary for accurate allelic determination. We followed the manufacturer’s instructions for the HLA-typing experiment protocol. Results from the sequencing is processed through a licensed software that comes with the kit, called AccuType™ Software. Analysis procedure and interpretation followed the software’s instruction manual.
Power calculation
We assessed power using QUANTO17 for the validation of the HLA-B*13:01, assuming a similar odds ratio (OR) with the initial discovery in Chinese10 applies to the Indonesians (Figure 1). Assuming the frequency of HLA-B*13:01 in Indonesians is around 2% (similar to the Chinese), the sample size to analyze in this proposed study will have a good power (> 0.8) to capture the associations of HLA-B*13:01 at a type-1 error rate of 5% if we have at least 15 cases and 15 controls. Therefore, our collection of 34 cases and 55 controls are well powered to detect a positive association.
Statistical Analysis
Data analysis was done using R, which is a dedicated statistical and data management tool for genetic analysis. Association test of HLA-B*13:01 and DHS was done by performing logistic regression.
Interpretation of Results
P-value was interpreted as the statistical significance of the test results as compared to by chance alone. A small p-value (<0.05) indicates evidence against the null hypothesis of no association, hence it will be rejected, whereas a large p-value (>0.05) indicates weak evidence against the null hypothesis, hence we fail to reject it. Odds ratio (OR) is a relative measure of effect comparing the DHS cases relative to the non-DHS cases. If the OR equals to 1, it implies no difference between the two groups.
RESULTS
In this study we recruited 34 leprosy patients with DHS as cases and 55 leprosy patients without DHS as controls. One control failed to have a good quality DNA; yellow extract with a 260/280 OD ratio of 1.31. Excluding this sample, we had 34 cases and 54 controls typed on the HLAssure B Locus kit. Unfortunately, two other controls were not able to be sequenced successfully as only 25% of the reads were unable to be mapped on the AccuType™ Software, and it failed to distinguish which HLA-B alleles these sample carry. Finally, excluding all the failed samples, 34 cases and 52 controls with successful HLA-B typing results were brought for further analyses.
HLA-B*13:01 is the most associated allele with DHS in Papuans
Based on the result of the SBT, we can see a spectrum of HLA-B allele frequency in the leprosy patients from Papua (Table 1). We performed a log additive test (logistic regression) to observe the association of all the identified HLA-B alleles to DHS, but we discovered that only HLA-B*13:01 allele showed association with DHS. HLA-B*13:01 is present in 50% of the cases, 1.9% of the controls and the difference between the two groups is significant (P=4.81×10−9; odds ratio=247.65), surpassing the genome-wide significance threshold. When the region where the samples came from was considered as a potential confounding factor (Papua or West Papua province), the adjusted P-value and odds-ratio of HLA-B*13:01 is considered unaffected (adjusted P=7.85×10−9; odds ratio=255.39), indicating that the evidence of association was not driven by the effect of population stratification.
Allele frequency spectrum of Papua Population
HLA-B*13:01 as a risk predictor for DHS in Papuans
There were 31 carriers and 3 non-carriers of HLA-B*13:01 among the DHS cases; whereas among the controls, there were 2 carriers and 50 non-carriers (Table 2). In other words, HLA-B*13:01 was present in 91.2% of case patients and 3.8% of the controls. This suggests that HLA-B*13:01 had a sensitivity of 91.2% and specificity of 96.2% as a predictor of DHS. However, we also observed a significant deviation of the genotype frequency of HLA-B*13:01 in the case patients (P=0.00039), unlike the controls (P=1).
The allele Frequency and Hardy Weinberg P-Value
Using HLA-B*13:01 as a predictor of DHS risk in Papuans, we observed that the area under the curve (AUC) = 0.95, which means that HLA-B*13:01 can correctly classify people with and without risk of developing DHS 95% of the time (Figure 2). Consequently, the risk carried by a person who has at least one copy of HLA-B*13:01 allele (heterozygous odds ratio) is then 233.3 times higher than the risk of those who do not carrying any copy of HLA-B*13:01 allele. Assuming that the prevalence of DHS in Papua is similar to the estimated global prevalence of 1.4%8, HLA-B*13:01 would have a positive predictive value (PPV) of 25.4% and a negative predictive value (NPV) of 99.9% in identifying people at risk to DHS. To prevent one case of DHS, we would theoretically need to screen 79 leprosy patients and hence we could reduce DHS risk by 10-fold (from 1.4% to 0.13%). Our data from 2016 in Papua has shown that the incidence of DHS among leprosy patients is 10-times higher than the global estimate (∼11%). This translates to a PPV of 74.8%, NPV of 98.9%, and 10 people needed to screen to prevent one case of DHS.
Receiver-operating curve for an additive prediction model of DHS using HLA-B*13:01 allele as predictor in Papuans. AUC, area under the curve. Overall sensitivity is 91.2% and specificity is 96.2%.
DISCUSSION
In this study, we replicated the association of with dapsone hypersensitivity syndrome in the Indonesian population, where Papua and West Papua provinces were selected as they are considered top three largest leprosy pockets in Indonesia with a prevalence of 4,06 per 10,000 population in Papua and 11,48 per 10,000 population in West Papua. These two areas are also considered as focus community areas in the National Program for Leprosy Elimination. In terms of DHS, according to an unpublished annual local report, Papua has an incidence between 2-10% per year. Besides, the neighboring country sharing the same island as Papua has shown to have the highest reported allele frequency in the world (28%), unlike other major ethnicities in Indonesia, such as Javanese and Sundanese (1.5%) and Riau population (5.4%)18.
Initially discovered as a risk factor and predictor of DHS in the Chinese population10, HLA-B*13:01 has been shown in multiple other studies in Asia as being a strong predictor to the same condition11-14,19. It has also been shown to be associated with dapsone hypersensitivity syndrome occurring in non-leprosy patients in Thailand11. Compared to the studies in Chinese and Thais, the odds ratio we discovered in Papuans are 12-times what was reported in Chinese10 and 4.6-times what was reported in Thais11. Presumably, the abundance of HLA-B*13:01 in Papuans contribute to the high odds ratio we see in this population. In Chinese, the frequency of HLA-B*13:01 ranges between 3-8% across North and South China, but for the Melanesians and Australian Aboriginals who are evolutionary related to Papuans, the allele frequency may reach up to 28%18.
Similar to the publication in the Chinese population10, we found a significant deviation of HLA-B*13:01 genotype frequency from Hardy Weinberg equilibrium in only the case patients (P=3.9×10−4). We also think the reason for this is due to the potential oversampling as the effect of the highly penetrant HLA-B*13:01 as a susceptibility factor for DHS and the bias towards sampling only the surviving DHS patients. Although not very well documented, the mortality rate for DHS in Papua is estimated to be quite high. Many of the patients come from rural areas and oftentimes they were treated after active case findings into their home villages. Upon drug delivery, any side effect that occur to them were often not reported. Patients even think these side effects were curses instead of a medical condition. Hence, with not much understanding of the disease and potential side effects, many of the fatal cases would have been missed and being under-reported.
Until now, there is no preventive action to reduce DHS incidence that been implemented in hospitals as well as in point of care clinics to prevent DHS in Indonesia. The DHS treatment is performed with intensive care at the time DHS symptoms appear, and it may be due to the patient being warded for several days up to weeks. Besides the lack of knowledge and public awareness of DHS, Indonesia’s geographic landscape adds as a challenge for patients who need immediate treatment upon a severe adverse reaction. There have been reported deaths in DHS patients due to delays in handling and some community perceptions that stigmatizes DHS as a curse instead of a form of an adverse drug effect. The incidence of DHS case in the family or community also causes trauma as well as rejection to undergo MDT. This issue affects leprosy elimination program especially in endemic areas of leprosy in Indonesia. Screening of HLA-B*13:01 among leprosy patients before treatment can be valuable in many aspects: clinically to prevent DHS and its related mortality, socially to prevent trauma of losing family members and experiencing severe hypersensitivity, and in public health perspective to increase compliance to treatment and reduce transmission for reaching the goal of national elimination program.
LIMITATION
As a case-control population-based study, participants were recruited based on pass history of DHS diagnosis as written in the medical record. The limitation to this is that misdiagnosis of DHS may have happened and there will be no way to verify this real-time. However, to alleviate this issue, during sample collection we visited the subject and conducted in-depth interview accompanied by the attending physician and/or medical doctor who diagnosed and treated the subject when DHS had manifested.
CONCLUSION
HLA-B*13:01 is validated as a biomarker for DHS in leprosy patients in Indonesia. Screening for HLA-B*13:01 may potentially reduce DHS incidence significantly and benefit many aspects of the leprosy control program in Indonesia.
Data Availability
All data generated and analysed in this study has been included. Any additional data not in the paper can be provided by authors upon request.
ACKNOWLEDGEMENTS
We thank all the individuals who had participated in this project. This work was funded by grants from the National Key Research and Development Program of China (2016YFE0201500), Genome Institute of Singapore, Agency for Science Technology and Research, and National Institute of Health Research and Development, Ministry of Health Republic of Indonesia.
Abbreviations
- DHS
- dapsone hypersensitivity syndrome
- DNA
- deoxyribonucleic acid
- HLA
- Human Leukocyte Antigen
- MDT
- multi-drug treatment
- OR
- odds ratio
REFERENCES
