Published ahead of print on August 2, 2007, doi:10.1164/rccm.200702-187OC
© 2007 American Thoracic Society doi: 10.1164/rccm.200702-187OC
Impact of Immigration on the Molecular Epidemiology of Mycobacterium tuberculosis in a Low-Incidence Country1 Norwegian Institute of Public Health, Oslo, Norway Correspondence and requests for reprints should be addressed to Ulf R. Dahle, Ph.D., Division of Infectious Disease Control, Norwegian Institute of Public Health, P.O. Box 4404 Nydalen, N-0403 Oslo, Norway. E-mail: ulf.dahle{at}fhi.no
Rationale: Programs to prevent the incidence rate of tuberculosis (TB) from increasing in many low-incidence countries are challenged by international travel and immigration from high-burden countries. Objectives: The current study aimed to determine the effect of such immigration on the genetic diversity of Mycobacterium tuberculosis isolates in an entire nation's population during 1994–2005. Methods: A total of 3,131 patients were notified with TB during the 12-year period. Of these, 2,284 (73%) had TB verified by culture, and isolates from 2,173 (96%) of these were analyzed by IS6110 restriction fragment length polymorphism. Measurements and Main Results: Only 31% of the included strains were isolated from nonimmigrants, the remaining 69% were isolated from immigrants. Although the incidence increased throughout the period, the genetic diversity remained high. A total of 135 clusters were identified; the percentage of recent disease was reduced among nonimmigrants, and remained stable among the immigrants during the study period. Although 69% of the isolates originated from immigrants from high-incidence countries, the established TB control program in the receiving country was adequate for the prevention of disease transmission. On average per year, only 2 nonimmigrants and 13 immigrants developed disease as a result of infection within the country by imported M. tuberculosis. Conclusions: Twelve years of M. tuberculosis importation as a result of immigration from high-incidence countries had little influence on the transmission of this pathogen in the receiving low-incidence country. To prevent future increase of transmission of TB, the current control strategies of low-incidence countries are adequate but must be maintained.
Key Words: tuberculosis • epidemiology • fingerprinting • immigrants
In the last decade, the incidence rates of tuberculosis (TB) have fallen in wealthy countries but risen in many developing countries. Programs to keep TB from rising in many low-incidence countries are perceived to be challenged by international travel and immigration from high-burden countries (1–8). Therefore, TB has been the subject of much concern in recent years. Surveillance programs have been questioned as have the policies for bacillus Calmette-Guérin immunization and screening of immigrants (3, 6, 9, 10). Tackling TB depends on effective surveillance, prevention, and control, and well-established systems are already in place in low-incidence countries throughout the world (6, 11, 12). The Norwegian TB control program is based on a national regulation from the Ministry of Health and Care Services (13). It requires all counties to establish and implement TB control programs to ensure case-finding and case-holding activities. Early diagnosis and adequate treatment under direct supervision remain the main strategy, in addition to increased use of treatment of latent infection. Asylum seekers and immigrants from high-incidence countries are screened on arrival by use of chest radiograph and tuberculin skin test. In addition, contact tracing is done in cases of infectious pulmonary disease. Identification of isolates is confirmed, and restriction fragment length polymorphism (RFLP) characterization is performed on all positive Mycobacterium tuberculosis cultures at the National Reference Laboratory for Mycobacteria, the Department of Bacteriology and Immunology, Division of Infectious Disease Control, Norwegian Institute of Public Health (NIPH). A national registry compares notifications of diagnosis of TB made by clinicians, antituberculosis drug prescriptions handled by pharmacies, and positive M. tuberculosis cultures reported by laboratory staff. Because of this triple reporting system, notifications are quite complete. The facilitated network of county TB coordinators further ensures the follow-up of individual cases. Mycobacterium bovis bacillus Calmette-Guérin immunization is offered to all teenagers, in addition to newly born infants of parents from high-incidence countries. Health services are free of charge for all, and sick leave is financially compensated by the government. Patients with TB are not deported while still receiving treatment. Increasing number of immigrants arriving from high-incidence countries may justify changes or adaptations to TB control programs (10). One important potential impact of imported TB cases from high-incidence countries to low-incidence countries would be an increase in the rate of transmission in the recipient country. The impact of such import is difficult to measure or forecast, due to incomplete strain collections, low levels of transmission, and the chronic nature of TB. However, the comparison of complete M. tuberculosis collections is feasible in some small countries, due to the manageable number of M. tuberculosis strains present. In addition, the fact that patients may stay latently infected for many years before they develop symptoms of active TB means extended observation periods are required. In the current study, we describe the molecular epidemiology of the complete M. tuberculosis population of a low-incidence country over 12 years. The data presented may serve as a model for the TB epidemics of current and future low-incidence countries that draw immigrants from various high-incidence parts of the world.
Patients and Cultures The study population comprised 96% of all patients in Norway from whom at least one sample positive for M. tuberculosis by culture was collected during 1994–2005. A total of 14 microbiological laboratories, servicing the entire nation, performed the isolation of M. tuberculosis from patient samples. Patient information was collected at the Department of Infectious Disease Epidemiology, NIPH. The Department of Bacteriology and Immunology, NIPH, including the National Reference Laboratory for Mycobacteria, received all M. tuberculosis isolates and performed IS6110 RFLP analysis. During this period, 3,131 patients were notified with TB, and cultures of M. tuberculosis were recovered from 2,284 of these patients. A total of 2,214 of these isolates were received, analyzed consecutively, and included in the national RFLP database. Denominators used to calculate the incidence of TB in Norway were extrapolated from Statistics Norway (http://www.ssb.no). The case detection rate was calculated as the number of cases notified divided by the number of cases estimated for that year and was expressed as a percentage. For Norway, during 1995–2005, these percentages were 57, 61, 60, 70, 70, 77, 102, 93, 127, 113, and 112%, respectively (http://www.who.int). This demonstrated that, during the latter years of the study, more cases were detected than the estimated number of cases and the case detection rate was considered to increase. Species identification of the isolates was based on a 16S-rDNA hybridization technique (AccuProbe; GenProbe, Inc., San Diego, CA) and standard microbiological tests (nitrate reduction and niacin accumulation tests). For this study, people born abroad were referred to as immigrants and people born in Norway to Norwegian or immigrant parents were called native or nonimmigrants. Throughout this article, 1 year implies the calendar period from January 1 until December 31.
Molecular Analyses of Isolates The IS6110 fingerprint patterns and spoligopatterns were compared by visual examination and computer-assisted analyses by use of BioNumerics, version 1.5, software (Applied Maths, Kortrijk, Belgium). To facilitate the comparison of the fingerprints, normalization was done using the molecular weight standards run on each gel. Similarity measures were calculated using the Dice coefficient, allowing 1.0% position tolerance. Cluster analysis was performed using the unweighted pair–group average method. Clusters were defined as groups of strains with 100% identical IS6110 RFLP patterns. This analysis was repeated, allowing clusters to also include strains with 80% identical RFLP patterns and differing in one or two bands. The patterns obtained with the spoligotyping were compared by visual examination and computer-assisted sorting of the results in an Excel sheet (Microsoft Excel 97; Microsoft Corporation, Redmond, WA). Although the latency period may vary considerably between different individuals, a secondary case was defined as an isolate with a fingerprint pattern identical to at least one previously isolated strain. The population diversity was calculated by dividing the number of different IS6110 RFLP patterns by the number of isolates fingerprinted each year. The degree of clustering was calculated continuously, and for each year, by dividing the numbers of isolates in clusters per number of fingerprinted isolates. We included linear regression analyses (SigmaPlot 9.0; Systat Software, Inc., San Jose, CA) to illustrate the trends of change in the M. tuberculosis population through the 12-year period, and 95% confidence intervals are given in Figures 1, 3, and 4.
The incidence of TB showed a steady increase throughout the study period, from 4.7 to 7.2 per 100,000 population (Figure 1). This was due to an increasing number of immigrants notified with the disease, whereas the incidence among nonimmigrants was declining (Figure 2). Repeated samples from individual patients that were found to be identical by use of IS6110 RFLP were removed from the database. The genetic diversity remained high, indicating limited transmission to both immigrants and nonimmigrants residing within the country.
Detection of Laboratory Cross-contamination Of the 2,214 strains, 41 (1.9%) were reported as possible cases of laboratory cross-contamination (identical RFLP patterns in strains isolated from different patients within the same laboratory during the same week). These cases of cross-contamination included 23 episodes. One included 7 false positive samples, 2 episodes each included 4, 3, and 2 false positive samples, and 16 episodes included single samples originally expected to be negative for M. tuberculosis. The cases were observed in 11 laboratories. The strains were excluded from further analyses. Eleven of the 41 patients had initiated treatment for TB by the time the false-positive results were suspected.
Patient Origin and Age During 1986–2002, the median observation period between arrival to the country and TB notification for all cases was 1.5 years (range, 0–15.4 yr). Immigrants from Asia, Latin-America and the Caribbean, Africa, and Europe (excluding the Nordic countries) had median intervals of 2.1, 1.6, 1.4, and 0.5 years, respectively (16). Few cases were diagnosed by compulsory screening on arrival (19 in 2005), thus active TB developed after immigration.
DNA Polymorphism of M. tuberculosis Isolates
The percentages of cases due to clustered isolates are given in Table 1. The rate of patients infected by TB caused by isolates carrying unique RFLP patterns increased throughout the study period. The rate of clustered isolates, however, was stable during 1996–2005 (Table 1). Regression analysis demonstrated that this rate remained around 0.8 during the study (data not shown). Thus, the increasing incidence of TB in this country represented an increase in RFLP-unique cases among immigrants. A total of 385 isolates were assigned to 135 clusters. Only 13 clusters included more than five patients and all these clusters had index cases notified before 2000. The index cases of 10 of these major outbreaks were of foreign origin and three were of nonimmigrant origin. Only 16 of the 135 clusters included strains isolated from both immigrants and natives. Although the IS6110 RFLP patterns of M. tuberculosis are fairly stable (18), a number of outbreaks and community-based studies have observed that IS6110-RFLP patterns may evolve during ongoing transmission (19). If a cluster was defined as two or more strains with 80% similarity and that the RFLP patterns differed by one or two bands, the total number of clustered isolates increased to 494 isolates assigned to 143 clusters. In most cases, this led to an increased number of strains included in the above-mentioned clusters. In some cases, however, it led to the convergence of existing clusters or introduced new ones. Due to the high stability of IS6110 (18, 19), this definition of a cluster is considered to overestimate the level of transmission. However, even with this assumption, the degree of clustering was very low and it confirmed that the analysis did not greatly underestimate the transmission level in the current setting. The number of clustered strains is expected to increase in parallel with an increasing number of RFLP-typed strains available in a database (20). We thus also calculated the percentage of patients who had a clustered isolate within a 24-month period (8). Defining clustering within a 2-year period is expected to represent strains that lead to rapid progression to disease (21). When a 24-month period was used as a clustering interval, among the nonimmigrant group the proportions of clustered cases gradually decreased during 1994–2005, whereas, among the immigrants, these proportions remained stable (Figure 4). This will also ensure a similarly sized RFLP database for all isolates. Although the overall incidence of TB increased, the percentage of clustered M. tuberculosis isolates remained stable during the last decade. The increasing incidence was due to M. tuberculosis isolates carrying unique RFLP patterns (Figure 5).
The number of RFLP patterns observed among the M. tuberculosis isolates for each year increased throughout the period, as did the number of unique RFLP patterns observed each year (Table 1). This supported the observation that importation of M. tuberculosis did not lead to increased levels of transmission and that the established TB control program could control the imported M. tuberculosis.
Presence of Beijing Lineage
Influence of Immigration A total of 250 secondary TB cases were diagnosed among the culture-positive cases during the 12-year period. Only 75 of these were nonimmigrants, and 175 were immigrants. Among the nonimmigrants, only 23 were infected with an M. tuberculosis strain believed to be imported due to its previous isolation from a patient of foreign origin. Among the immigrants, 159 were for the same reason expected to have been infected within the country by an imported strain. Individuals may, however, also have been infected by identical strains outside the country. Thus, our assumption may overestimate the number of individuals infected within the country, but it is not likely to underestimate the level of transmission. It was demonstrated that, on average, a maximum of 2 nonimmigrants and 13 immigrants developed TB disease after infection within the country each year by imported M. tuberculosis strains. Even when strains with 80% similarity and one to two different RFLP bands were clustered (data not shown), only 351 secondary cases were found. This supported the conclusion that transmission of imported M. tuberculosis within the current country was very low.
It is well known that much of the TB in Australia, Canada, Europe, and the United States occurs in foreign-born people and that the risk of TB is great in the first years after arrival in the host country (12, 20, 23–32). However, TB can persist within immigrant communities for extended periods, and the prevalence of the disease increases with age (7, 29, 33–35). It is therefore less clear how importation of pathogens influence local populations over extended periods, such as years or decades, because detailed characterizations of complete microbial collections are difficult to obtain. Also, due to the chronic nature of TB and the persistence of substantial immigrant populations with latent TB in low-incidence countries (29), it may take many years before the full impact of imported TB can be measured. In the current study, immigrants represented the majority of TB cases, and all strains of M. tuberculosis isolated within an entire country during 12 years were analyzed by IS6110 RFLP. The number of notified TB cases within the country increased during these years. The percentage of immigrants increased in general and among patients with TB. Each year, a few additional patients developed disease due to established strains but slightly less than one significant outbreak was detected for each year of the study (10 outbreaks in 12 yr). Among the strains initially isolated from immigrants during the 12 years, only 10 were able to establish an epidemic that included more than five patients within 5 years. This inability to establish outbreaks included the Beijing lineage, members of which are commonly believed to harbor increased ability for transmission. This indicated that the importation of M. tuberculosis isolates did not represent an immediate challenge to the national TB control program. It was demonstrated that the number of clustered isolates among nonimmigrants declined during the study period, despite the import of foreign strains. The proportion of clustered isolates among immigrants was stable during the 12 years. Transposition of IS6110 apparently depends on strain-specific properties and allows changes in RFLP patterns during ongoing transmission (19, 21, 36). IS6110 RFLP studies are, however, regarded as sufficiently stable to allow study of transmission dynamics of M. tuberculosis over time (19). The public debate on immigration to wealthy countries has often focused on the negative effects immigration may have on public health. Immigrants have been accused of spreading TB. However, the current study demonstrated that importation of M. tuberculosis, over 12 years, did not generate significant negative effects on the transmission of TB in a country that was considered to be in the elimination phase of this disease (37). Although the current study did not include individuals with latent M. tuberculosis infection, the number of M. tuberculosis strains isolated from nonimmigrants declined continuously throughout the 12-year period (Table 1), and immigration did not appear to influence the TB situation among the existing residents of the country. In the current 12-year study, importation of M. tuberculosis led to an increased incidence of TB during the period, but otherwise had little effect on the TB situation in this low-incidence setting. The social situation of the immigrant population is expected to have changed during the 12-year period, as has the size of this group. Such changes may in part explain the changing TB epidemiology. However, the low number of clustered strains could not support the statement that public health in this recipient country was hampered by immigration from high-incidence countries. It appeared that the TB control efforts were not overwhelmed by the challenge of imported TB. The national control program against TB in Norway is comparable to that of most other low-incidence countries (38, 39). The strategies used mainly include active entry screening of immigrants, contact tracing of notified infectious cases, and directly observed treatment. A national registry ensures updated surveillance data, and the network of county-level TB coordinators monitors the follow-up of individual cases. If these control strategies are well maintained, elimination of indigenous transmission of TB could be achievable, despite extensive import from high-incidence countries.
The authors acknowledge the microbiological laboratories in Norway for sending in their M. tuberculosis isolates to the Norwegian Institute of Public Health (NIPH). They thank the staff at the NIPH for their skillful assistance. They are also grateful to Richard M. Anthony, Royal Tropical Institute, Amsterdam, The Netherlands, for revising the manuscript.
Originally Published in Press as DOI: 10.1164/rccm.200702-187OC on August 2, 2007 Conflict of Interest Statement: None of the authors has a financial relationship with a commercial entity that has an interest in the subject of this manuscript. Received in original form February 5, 2007; accepted in final form July 25, 2007
Related articles in AJRCCM:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
AT A GLANCE COMMENTARY
