Tuberculosis Caused by Mycobacterium africanum, United States, 2004–2013

Routine reporting of TB caused by this organism does not appear warranted at this time.

T uberculosis (TB) is an infectious disease caused by a group of highly-related organisms comprising the Mycobacterium tuberculosis complex (MTBC), which includes M. tuberculosis, M. africanum, and M. bovis. Although all members of MTBC might cause disease in humans, M. tuberculosis and M. africanum are the primary cause of disease in humans globally, whereas M. bovis primarily causes disease in cattle (1,2). Like M. tuberculosis, M. africanum is spread by aerosol transmission (3).
Several studies have explored whether there are clinical differences between TB caused by M. africanum and TB caused by M. tuberculosis. These studies demonstrated variable findings with regard to associations of M. africanum with HIV status and findings on chest radiography (8,(28)(29)(30). Contacts of persons with TB caused by M. africanum appeared to have a lower rate of progression to active TB compared with contacts of persons with TB caused by M. tuberculosis, and a lower rate of genotype clustering has been described for M. africanum than for M. tuberculosis in relatively small studies from West Africa (14,29).
Although bacterial strains causing TB from all over the world can be found among cases of TB in the United States, analysis of routinely collected genotyping data for [2005][2006][2007][2008][2009] showed that 179 (0.5%) of 36,458 TB cases reported nationally were caused by M. africanum (31). We sought to further expand knowledge of M. africanum in the United States by reviewing all cases of TB reported nationally during 2004-2013. The objectives of this study were to ascertain the proportion of TB cases caused by M. africanum in the United States; compare clinical and epidemiologic characteristics between M. africanum and M. tuberculosis; and determine the extent to which M. africanum strains in the United States might be related by transmission on the basis of genotype clustering.

Methods
Genotype data from the Centers for Disease Control and Prevention (CDC; Atlanta, GA, USA) National TB Genotyping Service for 2004 through 2013 were linked to routine demographic and clinical data from all culture-confirmed cases in the CDC National TB Surveillance System from all 50 US states and the District of Columbia (32). As described previously (33), phylogenetic lineage (M. africanum and M. tuberculosis) for TB cases was assigned on the basis of spoligotype by using a set of rules correlating spoligotype to lineages defined by large sequence polymorphisms; for cases that did not meet a full rule for assignment on the basis of spoligotype, 12-locus mycobacterial interspersed repetitive unit variable number tandem repeats (MIRU-VNTRs) was used in addition to spoligotype to assign lineage. Cases reported during 2004-2008 only had 12-locus MIRU-VNTR data available, and cases reported during 2009-2013 had 24-locus MIRU-VNTR data available. To identify cases that could be caused by ongoing transmission in the United States, clusters of cases were defined as >2 cases with the same spoligotype and 24-locus MIRU-VNTR pattern in a given county. Cases that were caused by organisms other than M. africanum or M. tuberculosis were excluded from analysis.
All analyses were conducted by using R statistical software version 3.0.1 (R Core Group, Vienna, Austria). Statistical test results were considered significant at p<0.05. We examined patient attributes, genotype clustering, clinical characteristics (e.g., disease site), and social risk factors (e.g., homelessness) associated with M. africanum and M. tuberculosis. Odd ratios (ORs) and 95% CIs were calculated. Differences in proportions of cases were detected by using Fisher exact and Pearson χ 2 tests.
Factors identified as statistically significant by bivariable analysis at p<0.05 were entered into a multivariable logistic regression model to assess whether these factors were independently associated with M. africanum and M. tuberculosis. Tolerance <0.10 was used to detect colinearity, and the likelihood ratio test was used to test for interaction. To address collinearity between race/ethnicity and origin of birth, variables for race/ethnicity, country of origin, and West African origin were combined into a single variable and included in selection of the multivariable regression model. West African origin was defined as having been born in any of the following countries in West Africa: Nigeria, Liberia, Sierra Leone, Guinea, The Gambia, Ghana, Mali, Senegal, Côte d'Ivoire, Togo, Cameroon, Mauritania, Niger, and Guinea-Bissau.

Ethics Statement
Data for this study were collected as part of routine TB surveillance by CDC. Thus, this study was not considered research involving human subjects, and institutional review board approval was not required.

Results
A total of 125,038 cases were reported to the National TB Surveillance System during 2004-2013 ( Figure 1). Of these cases, 95,836 (76.6%) had a culture result positive for MTBC. Of cases with positive culture results, 73,290 (76.5%) had available lineage identification on the basis of genotype data. Of the cases for which lineage identification was available, the causative agent was determined to be M. africanum for 315 (0.4%) and M. tuberculosis for 71,727 (97.9%) cases: 1,248 (1.7%) cases had an isolated organism other than M. africanum or M. tuberculosis and were excluded from further analysis ( Figure 1).
M. africanum was assigned as the causative agent of TB for isolates with a genotype-assigned lineage of L5 or L6. All isolates designated as M. africanum met the conventional spoligotype rule of the absence of spacers 8, 9, and 39 or the absence of spacers 7-9 and 39 (7). M. tuberculosis was assigned as the causative agent of TB for isolates with a genotype-assigned lineage of L1, L2, L3, L4, or L7.
Among case-patients with M. africanum as the causative agent of TB, 276 (87.6%) had country of birth other than the United States (online Technical Appendix Ta  confirmed TB that were genotyped ranged from 68.0% to 97.1%, which was comparable with the overall proportion of culture-confirmed TB cases that were genotyped nationally. On the basis of the genotype cluster definition of >2 cases in the same county with identical spoligotype and 24-locus MIRU-VNTR patterns, only 1 cluster of M. africanum cases was identified during 2009-2013. The cluster consisted of 2 case-patients with the L5 lineage: 1 foreignborn person and 1 US-born person. Among 315 cases of M. africanum TB, 183 distinct genotypes were identified (spoligotype and 12-locus MI-RU-VNTR available for cases reported during 2004-2013; online Technical Appendix M. africanum TB cases had lower odds than M. tuberculosis TB cases of being in a cluster (defined by spoligotype and 24-locus MIRU) of cases (OR 0.1, 95% CI 0.1-0.2), being in persons >65 years of age (OR 0.2, 95% CI 0.1-0.5), being in persons with an abnormal chest radiographic result and cavitation (OR 0.6, 95% CI 0.5-0.9) and in persons without cavitation (OR 0.5, 95% CI 0.4-0.7), being in a resident of a correctional facility (OR 0.2, 95% CI 0.0-0. 6 To control for possible host differences in larger analysis, we conducted a subanalysis of cases among foreign-born persons from West Africa. In this subanalysis, clustering was the only significant variable at the bivariable level, and M. africanum TB cases had lower odds of being in a cluster of cases than M. tuberculosis TB cases (OR 0.1, 95% CI 0.1-0.9). Among foreign-born persons with West African origin, we found no significant differences in clinical characteristics (e.g., HIV status, cavitary disease, sputum smear results) between TB cases caused by M. africanum versus those caused by M. tuberculosis. M. africanum TB cases with L5 and L6 lineages had similar proportions of HIV positivity (18.1% vs. 17.5%; p = 0.9) and cavitary disease by chest radiography (25.4% vs. 42.5%; p = 0.051). We found no significant differences in clinical characteristics or social risk factors for TB caused by L5 or L6 lineages.

Discussion
This study used nationally reported data on TB cases linked to genotype data to describe the epidemiology of M. africanum in the United States. The findings from this analysis indicate that M. africanum is a rare cause of TB in the United States and represents 315 (0.4%) of 73,290 cases with available genotype data reported during 2004-2013. Most cases were identified in large metropolitan areas throughout the United States. Although M. africanum is an infrequent cause of TB, most states reported >1 case of TB caused by M. africanum during the study period, which suggested that M. africanum is broadly distributed.
In this study, TB caused by M. africanum was more likely to occur in foreign-born West Africans and US-born non-Hispanic blacks and less likely in foreign-born persons originating from countries not in West Africa. These associations suggest that the epidemiology of M. africanum in the United States is driven primarily by migration of persons from West Africa. We also identified cases of M. africanum in US-born persons, primarily in non-Hispanic blacks.   occur in the United States, but the possibility of acquisition of TB during travel (e.g., to West Africa) cannot be excluded because travel history was not available in national surveillance data. In an initial report of 5 M. africanum cases in the United States, several case-patients did not report a history of travel to West Africa (34). The low proportion of TB cases attributed to M. africanum suggests decreased transmissibility in the United States. Reasons for decreased transmission of M. africanum are unknown but could include decreased infectiousness or decreased progression to disease compared with M. tuberculosis, as was previously reported (8).
Our findings support the observation that M. africanum is highly restricted to West Africa, where it has been estimated to cause up to 50% of all TB cases, although the reason for this restriction remains unclear (8). A recent study from Ghana reported an association between M. africanum and patient ethnicity, which suggests specificity of host−pathogen interaction could be 1 factor in limiting the spread of M. africanum to West Africa (35).
Most M. africanum TB cases were not part of genotype clusters, which suggested that transmission of M. africanum in the United States is not common. M. africanum TB cases were less likely to be associated with genotype clustering than M. tuberculosis TB cases by analyses of all cases reported in the United States and in a subanalysis of persons born in West Africa. This lower association of clustering is consistent with investigations from Ghana and The Gambia, which found M. africanum less likely to be in spoligotype-defined clusters (30,36).
After controlling for other factors, we found that TB cases in the United States caused by M. africanum and M. tuberculosis were similar regarding clinical presentation, social risk factors, and treatment outcomes. These findings are consistent with those of studies that compared treatment outcomes among cases of M. africanum and M. tuberculosis TB in West Africa, but contrast with studies describing differential associations with HIV and chest radiography findings (8,14,28,29). Unlike several reported studies, we could not compare specific chest radiographic findings for M. africanum versus M. tuberculosis because detailed radiographic information is not available in US surveillance data (8). Our study demonstrated similar clinical characteristics of TB caused by L5 and L6 lineages of M. africanum, which is consistent with that of a previous report (29).
Our results should be interpreted in light of the incomplete availability of genotype data. Nationwide coverage of genotyping has increased over time (37), but genotype data were not available for all culture confirmed cases. Although it is possible that our study underestimates the true burden of M. africanum, we expect that changes in system coverage do not substantially affect the main findings of the study. In addition, M. africanum and M. tuberculosis were identified by spoligotype and MIRU-VNTR, rather than by more phylogenetically robust methods, such as largesequence polymorphism analysis. Therefore, some misclassification might have occurred, but there is no reason to assume any bias was introduced. Finally, our definition of clustered cases was based solely on identical spoligotype and 24-locus MIRU-VNTR in the same county during 2009-2013 and therefore probably overestimates the extent of transmission that might be occurring at the county level. More robust methods for identifying clustered cases rely on a narrower time interval between cases and evidence of epidemiologic links between cases (38). Even with the direction of bias toward overestimation of clustering, we found only 1 cluster.
Although the annual number of reported TB cases in the United States has decreased in the past decade, the proportion of TB contributed by foreign-born persons has increased to >60% in recent years (39