Philip Supply is a consultant for Genoscreen; Caroline Allix-Béguec is an employee of the same company. All other authors: no conflict. This does not alter the authors' adherence to all the PLOS ONE policies on sharing data and materials.
Conceived and designed the experiments: MF-D. Performed the experiments: KS CA-B. Analyzed the data: KS CA-B GG MW VM MF-D. Contributed reagents/materials/analysis tools: KS CA-B GG MW DB VM MF-D. Wrote the paper: KS CA-B GG MW DB VM PS MF-D
.
Emergence of extensively drug-resistant tuberculosis (XDR-TB) represents an enormous challenge to Public Health globally.
Progression towards XDR-TB was investigated in Belgium, a country with a typically low TB incidence, by analyzing the magnitude, characteristics, and treatment success of multidrug-resistant tuberculosis (MDR-TB) through a population-based study from 1994 to 2008.
Among the 174 MDR-TB patients, 81% were foreign-born, 48% of these being asylum seekers. Although the number of MDR-TB patients remained stable through the study period at around 15 new cases annually, frequencies of resistance of the patients’ first MDR-TB isolate to second-line drugs increased, as well as the total number of antibiotics it was resistant to (p<0.001). XDR-TB cases were detected from 2002 onwards. For 24 patients, additional resistance to several second-line drugs was acquired during treatment. Molecular-guided investigations indicated little to no contribution of in-country clonal spread or exogenous re-infection. The increase of pre-XDR and XDR cases could be attributed to rising proportions of patients from Asia and Central and Eastern Europe (p<0.001) and an increase in the isolation of Beijing strains in these groups (p<0.001). Despite augmented resistance, the treatment success rate improved from 63.0% to 75.8% (p = 0.080) after implementation in 2005 of improved surveillance measures and therapeutic access.
Increasing severity in drug resistance patterns leading to more XDR- and “panresistant” TB cases in a country with a low TB incidence like Belgium represents a strong alert on worsening situations in other world regions and requires intense public health measures.
A major concern in the control of tuberculosis is the spread of multidrug-resistant tuberculosis (MDR-TB), caused by
The latest estimates of the World Health Organisation (WHO) arrive at 650 000 prevalent MDR-TB cases among the 12 million tuberculosis cases worldwide
A few longitudinal studies have reported the contribution to XDR-TB of additional resistance acquisition or transmission during treatment of MDR-TB, the upward trends over time of pre-XDR-TB in some MDR-TB patient populations, or the consequences of drug resistance severity on treatment outcomes of MDR-TB patients
Here, we performed a retrospective study collectively looking at these parameters on the MDR-TB population over 15 years (sampling date between 1994 and 2008) in Belgium. This country has a TB situation representative of that of a number of other low TB incidence countries. In Belgium, the case notification rate decreased from 14.9 cases per 100 000 inhabitants in 1994 to an all-time low of 9.4/100 000 in 2008. Within this period the number of MDR-TB patients remained more or less stable
Demographic and clinical information was obtained from the National Tuberculosis Register, while information regarding patient management and treatment outcome was obtained through BELTA-TBnet, a project of the Belgian Lung and Tuberculosis Association (BELTA) set up by the federal government in 2005 to provide the best possible care to MDR-TB patients, ensuring free access to all necessary anti-TB treatment. All data accessed in the context of the present study had not been collected for research purposes but as part of the routine data collection for epidemiological surveillance, as stated in the Public Register dated 25/04/1997 in accordance with article 18 of the law of 08/12/1992 of the Belgian Government regarding the protection of the privacy of the individual when dealing with personal data. The aforementioned registration in accordance with the Belgian Privacy Commission stipulates in its §9 that no written informed consent from the patients is required for the collection and analysis of epidemiological data and treatment success.
Patients from Central and Eastern Europe were grouped for calculations.
Isolates were collected through the surveillance network of MDR-TB implemented in Belgium since 1994 and analyzed in the Belgian National Reference Centre of Tuberculosis and Mycobacteria (Scientific Institute of Public Health). For each of the 174 patients retained as true multidrug-resistant cases during the study period (1994–2008), only the first MDR isolate was included in the calculations of antibiotic resistance and cluster analysis.
DST was performed by the proportion method of Canetti
Mutations associated with resistance to rifampicin were identified by INNO-LiPA® RIF-TB (Innogenetics, Belgium) or by sequencing the 81 bp core region of the
Three different genotyping methods were performed: IS
A strain cluster was defined as two or more clinical isolates with identical RFLP, spoligotype and 24-locus MIRU-VNTR patterns, and identical mutations in the
Statistical analyses were carried out in Stata/MP 11.2
During the 1994–2008 study period, a total of 174 patients (1.8% of all registered TB patients with a DST result) were retained as true MDR-TB cases, after exclusion of 19 persons considered to be false positive as the result of laboratory cross-contamination from high bacillary load smear positive pulmonary samples (11 occurrences each involving 1 to 5 samples in 7 different laboratories). It occurred particularly during the first years of the study and a revision of the clinical samples decontamination procedure was suggested to the concerned laboratories.
The 174 patients represent 80.6% of the culture positive MDR-TB patients reported in the Belgian TB register during the study period. For the remaining 19.4%, the culture was not sent to the Reference Laboratory or was too contaminated by other organisms to be further analysed.
Sixty-six percent of the patients were male, the mean age was 36 years (67.2% of the MDR-TB patients were in the 15–39 years age group) and 81% were foreign-born. Among the latter, 87.9% (123/140) came from highly endemic TB countries (defined as countries with an estimated prevalence of >100/100 000 inhabitants according to WHO
Previous treatment history was known for 139 patients. Among the patients whose strain showed a pre-XDR or XDR resistance pattern, 80.0% (20/25) had been treated before, versus 44.7% (51/114) of the patients harbouring a strain resistant to first-line drugs only.
The first MDR-TB isolate obtained from each patient showed resistance to ethambutol, pyrazinamide, rifabutin, amikacin and ofloxacin in respectively 66.1%, 50.0%, 73.8%, 11.8% and 8.8% of the tested cases. An increase of resistance over time was observed for each of these drugs, and reached statistical significance (p<0.05) for ethambutol, amikacin and ofloxacin. Moreover, among the 170 MDR isolates for which at least ofloxacin and amikacin susceptibility results were available, the number of antibiotics each isolate was resistant to significantly increased over time (p<0.001). A total of 25/170 (14.7%) pre-XDR cases and 5/170 (2.9%) XDR cases were observed, and their proportion increased over time as well, to reach 36.8% for pre-XDR and 10.5% for XDR in 2008 (
Five genetic groups comprised 85.6% (149/174) of all clinical isolates: Beijing (28.2%), LAM (20.7%), Haarlem (10.3%), and generic Euro-American T1 (16.1%) and T2 (10.9%) groups. A statistically significant increase over time of Beijing profiles was observed (p<0.001), linked to their association with patients from Asia (15/24, p<0.001) and Central and Eastern Europe (30/60, p<0.001), the increase of MDR-TB patients from these two world regions (p = 0.009 and p = 0.038, respectively) and the increased proportions of Beijing strains isolated in these population groups (p = 0.064 and p = 0.051, respectively) through the study period.
Mutations associated with resistance to rifampicin were found in 97.1% (168/173) of the MDR-TB strains; 58.3% (98/168) of these mutations corresponded to a Ser531Leu change in the
In contrast, a much more diverse array of mutations with no specific time trend was found in
Serial isolates were obtained from 37 patients including 4 patients whose first isolate was fully susceptible to first-line drugs (
For these 24 MDR-TB patients out of 37 with multiple isolates, resistance to additional drugs was observed during treatment. Legend: ▪ = Susceptible, ▪ = Resistant, −Test not performed, I: Isoniazid, R: Rifampicin, E: Ethambutol, Z: Pyrazinamid, Rb: Rifabutin, Ofl: Ofloxacin, Amk: Amikacin, Thio: Thioamide, Cap: Capreomycin NA = Not Applicable as less than 2 isolates were FLP profiled, so no conclusion possible.
In all cases, the genotypes of the successive isolates remained completely identical over the different typing methods tested (except for one patient where a single RFLP-band change was observed), indicating that additional resistance was not due to exogenous re-infection.
Combining all genotyping techniques performed on the first MDR-TB isolate of the 174 studied patients, as well as interrogating mutations in the
Legend: Y = yes; N = no, U = unknown; − = not applicable.
The patients coming from Asia were less involved in clusters than the patients from the other regions (p = 0.019). However, no significant relation could be determined between the genetic lineage of the isolates and their clustering,
An epidemiological link could be established for 6 clusters, while 3 partial links could be documented in 2 other clusters (
Treatment duration averaged 529 days,
Legend: astart of BELTA-TBnet in 2005 bincluding 3 isolates with unknown amikacin susceptibility cincluding 1 isolate with unknown amikacin susceptibility MDR: first-line drug resistance only Pre-XDR: MDR with additional resistance to either a fluoroquinolone or amikacin XDR: MDR with additional resistance to a fluoroquinolone and amikacin.
Three of the 5 patients whose first isolate collected in Belgium was already XDR presented an unsuccessful treatment outcome (
Part I: patient characteristics and DST results. Part II: second-line therapy with treatment outcome. Legend: *New treatment regimen initiated after 19 months without anti-TB treatment **Retreatment after interruption of 5 months #Stop after 8 months ##Stop after 7 months Antibiotics: R = Rifampicin, I = Isoniazid, E = Ethambutol, Z = Pyrazinamid, Rb = Rifabutin, SM = Streptomycin, Amk = Amikacin, Ofl = Ofloxacin, Cip = Ciprofloxacin, Mox = Moxifloxacin, Cap = Capreomycin, Thio = Thioamides, Clf = Clofazimin, Clr = Clarithromycin, PAS = para-aminosalycilic acid, Lzd = Linezolid, A-Cl = amoxicillin–clavulanate, Mrp = Meropenem, Trz = Thioridazine. The first treatment failure (XDR03) was simultaneously infected by two genetically different Beijing XDR strains, as detected by double alleles in 5 MIRU-VNTR loci. The second strain, not included in this study because not isolated from the patient’s first specimen, had additional resistance to capreomycin and clofazimine compared to the first one. After 31 months of unsuccessful hospital-supervised therapy, treatment was stopped. Nineteen months later, a new 9-drug regimen (including thioridazine) was prescribed but the patient died 4 months later. The second failure case (XDR04, the “panresistant” patient MDR169 mentioned in the text) returned to his home country after 18 months therapy, and subsequently died.
Two XDR patients cured. The first one, from Chechnya, responded well to a regimen of 25 months duration including pyrazinamid, capreomycin, ethionamide and cycloserin. The second one, from Georgia, had become bacteriologically negative after 3 months of second-line therapy but abandoned treatment after 12 months. Five months later, direct smear examination and culture were still negative but treatment was nevertheless restarted for an additional 21 months, with pyrazinamid, rifabutin, cycloserin, linezolid, meropenem and amoxicillin-clavulanic acid.
To our knowledge, this study is the first longitudinal nationwide survey of MDR and XDR-TB, covering a 15-year period (1994–2008), simultaneously integrating bacteriological and molecular data on the isolates, as well as epidemiological investigation, clinical aspects of the patients and treatment outcomes. The country studied has an epidemiological situation characteristic of that of many other low TB incidence countries
In-country transmission has played little to no role in fueling these trends, as indicated by the molecular-guided epidemiological investigations of the various patient strain clusters: 9 links only, in 8 clusters, detected or confirmed events of TB transmission. None of these involved XDR cases but one was a pre-XDR nosocomial transmission. In the other clusters, it was impossible to find any link between patients except that most cases in a same cluster were foreigners of the same geographic origin. The lack of proven links could reflect the difficulty of conducting retrospective field investigations in this specific population and/or the fact that many of these patients were already infected by a strain circulating in their country before they arrived in Belgium. The latter hypothesis is supported by data from the systematic screening of all asylum seekers for TB upon their arrival, and by the large diversity of strains with a unique genotype isolated in 67% of the MDR-TB patients.
Of note, no attempt has been made to systematically correlate the results of the cluster analysis to a possible classification of the cases as primary versus acquired resistance because the available information had, in most cases, been provided by the patients themselves without any documented proof.
Along the same lines, in contrast to what can be expected for high-TB incidence settings such as Uzbekistan
The increasing frequency of pre-XDR and XDR cases in Belgium appears to be driven mainly by the increased proportion of MDR-TB patients originating from Asia (p<0.001) and Central and Eastern Europe (p<0.001). Interestingly, this trend seems to be accompanied by a shift in the bacterial strain population structure in these patient groups, as seen by the increased proportions over time of the Beijing profiles isolated from these two world regions (p = 0.064 and p = 0.051, respectively). These findings are consistent with reports indicating both an increase in MDR- and XDR-TB rates, and an apparent selective expansion of some specific MDR- and pre-XDR-TB clones of
Overall, a successful treatment outcome was obtained for 67.8% of the MDR-TB cases, above the 56.3% cure rate of MDR-TB patients observed in Europe in 2010
In conclusion, while the annual MDR-TB case notification rate in Belgium has remained low, the severity of the drug resistance patterns has been increasing year by year, leading to more pre-XDR, XDR and “panresistant” TB strains. This trend mostly reflects the degrading situation of drug-resistant TB in Asia and Central and Eastern Europe, from which originate increasing numbers of MDR-TB patients in the country. Therapeutic outcomes have nevertheless been satisfactory and improved over the study period. This can be attributed to rapid diagnosis using early molecular detection of genotypic resistance, to molecular-guided cluster analysis confirming or disproving transmission and ensuring appropriate public health actions, as well as to enlarged therapeutic options and adapted structures for the patients’ long term treatment. In the face of the still growing proportions of pre-XDR, XDR and “panresistant” cases, these efforts must be maintained and intensified.
We thank Jean-Pierre Vincke, Kristien De Greef, Brigitte Vanfleteren, Romuald Lizon, Thùy Trang Uông and Polydor Sonck for their technical expertise that contributed to this research.