Abstract
Mixed infections with genetically distinct Mycobacterium tuberculosis (MTB) strains within a single host have been documented in different settings; however, this phenomenon is rarely considered in the management and care of new and relapse tuberculosis (T.B.) cases. This study aims to establish the epidemiological and clinical features of mixed infections among culture-confirmed T.B. patients enrolled in tuberculosis care at the Florida Department of Health (FDOH) and measure its association with T.B. mortality.
We analyzed de-identified surveillance data of T.B. cases enrolled in T.B. care from April 2008 to January 2018. Mixed MTB infection was determined by the presence of more than one Copy Number Variant (CNV) in at least one locus, based on the genotype profile pattern of at least one isolate using 24-locus Mycobacterial Interspersed Repetitive Unit-Variable Number Tandem Repeat (MIRU-VNTR).
The prevalence of mixed MTB infections among the 4,944 culture-confirmed TB cases included in this analysis was 2.6% (129). Increased odds of mixed infections were observed among middle-aged patients, 45–64 years (AOR = 2.38; 95% CI: 0.99, 5.69; p = 0.0513), older adults 65 years and above (AOR = 3.95; 95% CI: 1.63, 9.58; p = 0.0023) and patients with diabetes (OR = 1.77; 95% CI: 1.12, 2.80; p = 0.0150). There was no significant association between mixed infections and death.
Our study provides insight into the epidemiological and clinical characteristics of patients with mixed MTB infections, which is essential in the management of T.B. patients.
Keywords: Mycobacterium tuberculosis, mixed-infections, MIRU-VNTR, T.B. mortality
1. Introduction
Tuberculosis (T.B.) is an ancient disease that has plagued humankind for thousands of years and continues to be a global public health threat (World Health Organisation, 2019). Traditionally, T.B. is believed to be caused by infection from a single Mycobacterium tuberculosis (MTB) strain with recurrence assumed to be due to reactivation of the initial strain that resulted in the first episode (Rie et al., 1998; Shamputa et al., 2004). However, this concept was disputed when more than one MTB strain was isolated from a single patient using phage typing (Mallard et al., 2010). Subsequently, evidence from various molecular studies using different genotyping methods have been able to distinguish infection from multiple distinct MTB strains within a single host (Baldeviano-Vidalon et al., 2005; Nardell, 2004; Cohen et al., 2011). Mixed MTB infection may occur as a result of either within-host strain evolution (microevolution) after a single infection event or infection by sequential or simultaneous exposure to more than one strain (Cohen et al., 2011; Muwonge et al., 2013). Based on the fact that an individual can be infected at one point by different strains with different phenotypic characteristics like growth rates and drug resistance thresholds, within-host competition between different strains may adversely affect the clinical outcomes of patients with these infections (Fang et al., 2008; Cohen et al., 2016). Evidence from various studies has demonstrated the occurrence of mixed infections in different geographical settings and among T.B. patients co-infected with the human immunodeficiency virus (HIV) and HIV negative T.B. patients (Rie et al., 1998; Cohen et al., 2011; Muwonge et al., 2013; Cohen et al., 2012; Dickman et al., 2010). The prevalence of mixed MTB infections depends mainly on the diversity of strains within an individual’s environment, the transmission pressure of the pathogen, and the overall prevalence of T.B. within the given population (Muwonge et al., 2013). Host immunity counts as a significant contributing factor to the occurrence of within-host genetic diversity of the MTB strain at the individual level, making the phenomenon more likely among TB/HIV co-infected patients (Cohen et al., 2012; Viedma et al., 2005; Brites and Gagneux, 2012)
Mixed infections involving both drug-susceptible and drug-resistance strains within a single host have the potential of rendering standard treatment regimens ineffective, complicating laboratory diagnosis, and affecting the transmission dynamics of the disease (Rie et al., 1998; Dickman et al., 2010; Chindelevitch et al., 2016). Drug susceptibility testing (DST) in such instances may fail to detect the minority drug-resistant strains, which may result in adverse treatment outcome due to the presence of underlying resistance as well as an increased risk of acquired resistance with the standard treatment regimen (Pang et al., 2015; Shin et al., 2018). Mixed MTB infection may alter the immune responses of patients with ongoing T.B. infections and render them more susceptible to reinfection (Warren et al., 2004). This occurrence has implications for patients’ responses to clinical therapy and optimal antimicrobial dosing strategies (Chindelevitch et al., 2016; Thwaites et al., 2008).
Molecular genotyping methods such as whole-genome sequencing (WGS), mycobacterial interspersed repetitive unit-variable number tandem repeat (MIRU-VNTR) and other PCR-based genotyping techniques, such as spoligotyping have been critical tools in providing superior resolution in the mapping of genetic diversity among MTB strains including the presence of multiple strains within a single host as well as identifying genomic signature patterns associated with pathogenesis, drug resistance and disease transmission (Coscolla and Gagneux, 2014; Coscolla and Gagneux, 2010; Sharma et al., 2017). Compared to other genotyping techniques, the MIRU-VNTR method is the most widely used in the detection of mixed infections with MTB strains (Pang et al., 2015). In the United States (U.S.), the use of spoligotyping and 24-locus MIRU-VNTR for genotyping of MTB cases as part of routine T.B. surveillance has been influential in providing a deeper understanding of the molecular epidemiology of MTB as well as in the detection and control of T.B. transmission (Ghosh et al., 2012; Séraphin et al., 2016).
The MIRU-VNTR technique works as a microsatellite typing system that classifies MTB strains by the number of copies of repeats at different loci (Guide, 2005; Chindelevitch et al., 2016). These copy number variants (CNVs) are used in distinguishing MTB strains from other similarly typed strains (Chindelevitch et al., 2016). The occurrence of mixed infections is usually exhibited by the presence of two or three CNVs at a single locus resulting from either clonally heterogeneous infection where multiple CNVs are found in only one locus or mixed infection where two or more loci have multiple CNVs (Guide, 2005).
Although a patient’s clinical response to treatment may go beyond the presence or absence of in-host genetic diversity to consider factors like comorbidity and malnutrition (Nardell, 2004), a good understanding of the molecular epidemiology of mixed MTB infections and its associated features is critical for T.B. control effort. An appreciation of this mechanism is vital in the quest for a suitable T.B. vaccine, to evaluate treatment regimens and to predict disease trajectories (Shamputa et al., 2006).
This study describes the epidemiological and clinical features of mixed infections among culture-confirmed T.B. patients enrolled in tuberculosis care at the Florida Department of Health (FDOH) as well as determines the association between mixed infections and death among this cohort of patients. We hypothesize that mixed infections among T.B. patients can adversely impact disease presentation and the risk of death.
2. Materials and Methods
2.1. Study Design and Sample Selection
We deployed a cross-sectional study design to analyze de-identified surveillance data of culture-confirmed T.B. patients enrolled in the Florida Department of Health (FDOH) T.B. registry from April 2008 to January 2018. The dataset used in this analysis consisted of T.B. cases diagnosed and managed at the county health departments in Florida with their clinical and socio-demographic information. This dataset also included data on routinely determined genotype information for culture-confirmed T.B. cases using spoligotyping and 24-locus MIRU-VNTR methods as per the nationally standardized procedure (Ghosh et al., 2012).
The source population consisted of 6,393 registered T.B. cases, of which 5,044 (78.9%) were culture-positive. The study included all T.B. cases who were culture-positive at baseline and had assigned lineages using 24-locus MIRU-VNTR. We excluded 1,349 (21.1%) cases from the source population who were culture-negative, in addition to 100 (2.0%) cases who were culture-positive but had no MIRU-VNTR genotyping results. There were no significant differences between the sample population and the excluded population in terms of their socio-demographic characteristics. Overall, 4,944 cases representing 98.0% of the culture-confirmed T.B. cases were used in this analysis. The study was approved by the Institutional Review Boards (IRBs) of both the University of Florida and the Florida Department of Health (FDOH), with all data de-identified.
2.2. Measures
2.2.1. Exposure variable: Occurrence of Mixed infections
For each culture-confirmed T.B. case, the 24-locus MIRU-VNTR profile pattern based on genotyped results of at least one isolate was linked to the T.B. registry data by a unique patient identifier. The primary exposure variable was the presence of mixed infections based on the MIRU-VNTR pattern, which was defined as having more than one copy number variant (CNV) in at least one locus with all other profile patterns defined as simple infections. Mixed infection was measured as a dichotomous variable indicating the presence or absence of mixed strain infections.
2.2.2. Outcome variable: Death
The occurrence of death due to T.B. infection was the outcome of interest. This was defined as the occurrence of death either at diagnosis or during treatment and was analyzed as a categorical variable, dichotomized as Yes/ No.
2.3. Statistical Analysis
We computed the descriptive statistics of the sample using either Chi-square or Fisher’s exact test where appropriate. We assessed confounding effects by determining the association between the potential confounder and both the exposure variable, mixed infections, and the outcome variable, death. Covariates missing more than 5% of observations were evaluated to assess the effect of the missing values on both the exposure and outcome of interest using Little’s test (Li, 2013). We used logistic regression to obtain the crude and adjusted odds ratios (AORs) for the association between mixed infections and death. The full multivariate model was adjusted for the following demographic variables: age at diagnosis, sex, race, ethnicity, and country of birth. We also controlled for TB-related environmental risk factors like past 12 months history of homelessness, history of incarceration, history of alcohol and substance use, and other TB-related clinical risk factors like the presenting disease site (pulmonary/extrapulmonary), evidence of cavitation, miliary disease, history of previous T.B. disease, multi-drug resistant T.B. (resistant to at least isoniazid and rifampin), comorbidity with diabetes, and HIV status. The backwards elimination technique was used in the selection of the final model (Bursac et al., 2008). Test for the ultimate model fit was done using changes in Akaike Information Criteria (AIC) (Bozdogan, 1987). We tested the effect of MTB lineage as a modifier in the association between mixed infection and death by stratifying based on the different MTB lineages (Corraini et al., 2017). All data analysis was done using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA) with significance determined at a 0.05 alpha level and 95% confidence intervals (C.I.s).
3. 0. Results
3.1. Sample Characteristics
Table 1 provides the descriptive statistics of the epidemiological and TB-related clinical features of the sample population by both infection type and death. Prevalence of mixed MTB infections among the 4,944 culture-positive cases was 2.6% (129), while the prevalence of death was 15.5% (766). The study population was predominantly male (65.1%), white (49.8%), of non-Hispanic ethnicity (73.3%), and non-US born (54.3%). Most of the cases (39.2%) were within the 45-64-year group. The study sample had a fewer proportion of patients reporting a past 12 months history of homelessness (9.2%), incarceration (3.4%), alcohol use (18.2%), and substance use (9.2%).
Table 1.
Epidemiological and Clinical TB-Related Features of the Study Population by Infection Type and Death
| Characteristic | Sample (%) | Simple Infections (%) | Mixed Infections (%) | P-value |
|---|---|---|---|---|
| Total | 4944 (100) | 4815 (97.4) | 129 (2.6) | |
| Sex | ||||
| Male | 3219 (65.1) | 3125 (64.9) | 94 (72.9) | 0.0610 |
| Female | 1752 (34.9) | 1690 (35.1) | 35 (27.1) | |
| Age (years) | ||||
| ≤24 | 512 (10.4) | 506 (10.5) | 6 (4.7) | 0.0002 |
| 25-44 | 1520 (30.7) | 1491 (31.0) | 29 (22.5) | |
| 45-64 | 1937 (39.2) | 1886 (39.2) | 51 (39.5) | |
| ≥65 | 975 (19.7) | 932 (19.4) | 43 (33.3) | |
| Ethnicity | ||||
| Hispanic | 1322 (26.7) | 1280 (26.6) | 42 (32.6) | 0.1303 |
| Non-Hispanic | 3622 (73.3) | 3535 (73.4) | 87 (67.4) | |
| Race | ||||
| White | 2462 (49.8) | 2390 (49.6) | 72 (55.8) | 0.0026 |
| Black | 1831 (37.0) | 1800 (37.4) | 31 (24.0) | |
| Asian/Other | 651 (13.2) | 625 (13.0) | 26 (20.2) | |
| Country of Origin | ||||
| US-Born | 2262 (45.7) | 2213 (46.0) | 49 (38.0) | 0.0727 |
| Non-US Born | 2682 (54.3) | 2602 (54.0) | 80 (62.0) | |
| History of homelessness | ||||
| Yes | 455 (9.2) | 444 (9.2) | 11 (8.5) | 0.4564 |
| No | 4431 (89.6) | 4316 (89.6) | 115 (89.2) | |
| Unknown | 58 (1.2) | 55 (1.1) | 3 (2.3) | |
| Resident Correctional Facility | ||||
| Yes | 170 (3.4) | 166 (3.5) | 4 (3.1) | 1.0000 |
| No | 4774 (96.6) | 4649 (96.5) | 125 (96.9) | |
| History of Alcohol Use | ||||
| Yes | 902 (18.2) | 875 (18.2) | 27 (20.9) | 0.4235 |
| No | 4042 (81.8) | 3940 (81.8) | 102 (79.1) | |
| History of Substance Use | ||||
| Yes | 454 (9.2) | 443 (9.2) | 11 (8.5) | 0.7938 |
| No | 4490 (90.8) | 4372 (90.8) | 118 (91.5) | |
| Death | ||||
| Yes | 766 (15.5) | 739 (15.4) | 27 (20.9) | 0.0838 |
| No | 4178 (84.5) | 4076 (84.6) | 102 (79.1) | |
| Previous TB Disease | ||||
| Yes | 140 (2.8) | 136 (2.8) | 4 (3.1) | 0.8519 |
| No | 4804 (97.2) | 4679 (97.2) | 125 (96.9) | |
| Evidence of Cavitation | ||||
| Yes | 1291 (26.1) | 1250 (26.0) | 41 (31.8) | 0.0329 |
| No | 1100 (22.3) | 1064 (22.1) | 36 (27.9) | |
| Not Available | 2553 (51.6) | 2501 (51.9) | 52 (40.3) | |
| Miliary TB Disease | ||||
| Yes | 157 (6.8) | 153 (6.8) | 4 (5.4) | 0.8154 |
| No | 2152 (93.2) | 2082 (93.2) | 70 (94.6) | |
| MTB lineages | ||||
| Euro-American | 3714 (80.6) | 3651 (80.7) | 63 (73.3) | 0.2253 |
| Indo Oceanic | 403 (8.7) | 391 (8.6) | 12 (13.9) | |
| East Asian | 373 (8.1) | 364 (8.0) | 9 (10.5) | |
| Other | 121 (2.6) | 119 (2.6) | 2 (2.3) | |
| HIV Status | ||||
| Negative | 3725 (85.6) | 3619 (85.5) | 106 (91.4) | 0.0728 |
| Positive | 626 (14.4) | 616 (14.5) | 10 (8.6) | |
| Diabetes | ||||
| Yes | 549 (11.1) | 526 (10.9) | 23 (17.8) | 0.0138 |
| No | 4395 (88.9) | 4289 (89.1) | 106 (82.2) | |
| MDR – TB | ||||
| Yes | 74 (1.5) | 73 (1.5) | 1 (0.8) | 0.0015 |
| No | 4661 (95.0) | 4546 (95.2) | 115 (89.1) | |
| Not Available | 170 (3.5) | 157 (3.3) | 13 (10.1) | |
| Disease Site | ||||
| Pulmonary | 4175 (84.5) | 4056 (84.2) | 119 (92.3) | 0.0132 |
| Extra Pulmonary | 769 (15.5) | 759 (15.8) | 10 (7.7) | |
Most of the patients in the study sample had the pulmonary form of T.B. (84.5%) and were infected with the Euro-American strain (80.6%). The proportion of patients with multi-drug resistant T.B. was 1.5%, while those having comorbidity with diabetes and HIV infection were 11.1% and 14.4%, respectively.
3.2. Association between Mixed MTB Infections and TB-Related Epidemiological Characteristics
Compared to those with simple infections, cases with mixed MTB infections were more likely to be male (72.9% vs 64.9%), above 65 years of age (33.3% vs 19.4%), Hispanic (32.6% vs 26.6%), white (55.8% vs 49.6%) and non-US born (62.0% vs 54.0%). A higher proportion of cases with mixed MTB infections compared to those with simple infection were more likely to report a history of alcohol use in the past 12 months (20.9% vs 18.2%). Patients with a history of homelessness (8.5% vs 9.2%), substance use (8.5% vs 9.2%), and those previously incarcerated in a correctional facility (3.1% vs 3.5%) were less likely to have mixed MTB infections.
In the unadjusted analysis (Table 2), age of 65 years and above (OR = 3.89; 95% CI: 1.65, 9.20; p = 0.002) was associated with increased risk of mixed MTB infections while being black (OR= 0.52; 95% CI: 0.37, 0.88; p= 0.01) was a significant protective predictor of mixed MTB infections. Demographic and environmental risk factors like sex, ethnicity, country of birth, past 12 months history of alcohol use, history of substance use, history homelessness, and history of correctional facility use, were not associated with the risk of mixed MTB infections.
Table 2.
Unadjusted and adjusted association between mixed infection and demographic and clinical TB-related characteristics of the study population
| Characteristics | Unadjusted OR (95% CI) | Unadjusted P- Value |
Adjusted OR (95% CI) | Adjusted P-Value |
|---|---|---|---|---|
| Sex | ||||
| Male | 1.45 (0.98, 2.15) | 0.0626 | ||
| Female | 1.00 | |||
| Age (years) | ||||
| ≤24 | 1.00 | 1.00 | ||
| 25-44 | 1.64 (0.68, 3.97) | 0.2733 | 1.59 (0.65, 3.89) | 0.3109 |
| 45-64 | 2.28 (0.97, 5.34) | 0.0579 | 2.38 (0.99, 5.69) | 0.0513 |
| ≥65 | 3.89 (1.65, 9.20) | 0.0020 | 3.95 (1.63, 9.58) | 0.0023 |
| Ethnicity | ||||
| Non-Hispanic | 1.00 | 1.00 | ||
| Hispanic | 1.33 (0.92, 1.94) | 0.1312 | 1.30 (0.83, 2.06) | 0.2551 |
| Race | ||||
| White | 1.00 | |||
| Black | 0.52 (0.37, 0.88) | 0.0100 | ||
| Asian/Other | 1.38 (0.87, 2.18) | 0.1663 | ||
| Country of Origin | ||||
| US-Born | 1.00 | 1.00 | ||
| Non-US Born | 1.39 (0.97, 1.99) | 0.0740 | 1.39 (0.88, 2.20) | 0.1564 |
| History of homelessness | ||||
| No | 1.00 | |||
| Yes | 0.93 (0.50, 1.74) | 0.8198 | ||
| Missing | 2.05 (0.63, 6.64) | 0.2323 | ||
| Resident Correctional Facility | ||||
| No | 1.00 | |||
| Yes | 0.90 (0.33, 2.45) | 0.8312 | ||
| History of Alcohol Use | ||||
| No | 1.00 | 1.00 | ||
| Yes | 1.19 (0.78, 1.83) | 0.4239 | 1.22 (0.75, 1.99) | 0.4069 |
| History of Substance Use | ||||
| No | 1.00 | |||
| Yes | 0.92 (0.49, 1.72) | 0.7939 | ||
| Death | ||||
| No | 1.00 | 1.00 | ||
| Yes | 1.46 (0.95, 2.25) | 0.0855 | 1.24 (0.74, 2.08) | 0.4104 |
| Previous TB Disease | ||||
| No | 1.00 | 1.00 | ||
| Yes | 1.10 (0.40, 3.02) | 0.8520 | 0.529 (0.13, 2.19) | 0.3794 |
| Evidence of Cavitation | ||||
| No | 1.00 | |||
| Yes | 0.97 (0.62, 1.53) | 0.8936 | ||
| Missing | 0.62 (0.40, 0.95) | 0.0268 | ||
| Miliary TB Disease | ||||
| No | 1.00 | |||
| Yes | 0.78 (0.28, 2.16) | 0.6295 | ||
| MTB lineages | ||||
| Euro-American | 1.03 (0.25, 4.25) | 0.9710 | ||
| Indo Oceanic | 1.83 (0.40, 8.27) | 0.4347 | ||
| East Asian | 1.47 (0.31, 6.90) | 0.6246 | ||
| Other | 1.00 | |||
| HIV Status | ||||
| Negative | 1.00 | 1.00 | ||
| Positive | 0.55 (0.29, 1.07) | 0.0770 | 0.65 (0.33, 1.28) | 0.2180 |
| Diabetes | ||||
| No | 1.00 | |||
| Yes | 1.77 (1.12, 2.80) | 0.0150 | ||
| MDR - TB | ||||
| No | 1.00 | 1.00 | ||
| Yes | 0.54 (0.08, 3.93) | 0.5441 | 0.54 (0.07, 3.92) | 0.5378 |
| Not Available | 3.27 (1.81, 5.94) | <.0001 | 3.96 (2.09, 7.51) | <.0001 |
| Disease Site | ||||
| Extra Pulmonary | 1.00 | 1.00 | ||
| Pulmonary | 2.23 (1.16, 4.27) | 0.0158 | 1.92 (0.98, 3.73) | 0.0557 |
The adjusted analysis showed older age groups; 45 – 64 years (AOR = 2.38; 95% CI: 0.99, 5.69; p= 0.0513), and ≥ 65 years (AOR = 3.95; 95% CI: 1.63, 9.58; p= 0.0023) as being associated with mixed MTB infections, as compared with the younger age group (≤ 24 years).
3.3. Association between Mixed MTB Infections and Clinical TB-Related Characteristics
A higher proportion of patients with mixed MTB infections were more likely to experience death (20.9% vs 15.4%), have evidence of disease cavitation (31.8% vs 26.0), have comorbidity with diabetes (17.8 vs 10.9), present as HIV negative (91.4% vs 85.5%), and have the pulmonary form of disease presentation (92.3% vs 84.2%). Cases with mixed MTB infections were more likely to be infected with either the Indo Oceanic (13.9% vs 8.6%) or the East Asian strain (10.5% vs 8.0%).
In the unadjusted analysis of the association between mixed MTB infections and clinical TB-related risk factors shown in Table 2, Patients with diabetes had a 77% increased odds of mixed MTB infections as compared to those with only T.B. infection (OR = 1.77; 95% CI: 1.12, 2.80; p= 0.015) while co-infection with HIV was not significantly associated with mixed MTB infections. Higher odds of mixed infections were associated with death (OR = 1.46 95% CI: 0.95, 2.25) although not statistically significant.
4.0. Discussions
Many studies have associated high prevalence of mixed MTB infections with settings with high T.B. incidence and high transmission potential (Cohen et al., 2012; Shamputa et al., 2006; Chaves et al., 1999; Fitzpatrick et al., 2002). This is evident in the study by Cohen et al. (2011) that reported a 21.1% prevalence of heterogeneous infections among T.B. case in KwaZulu-Natal, South Africa and that by Shamputa et al. (2006) that found a prevalence of 13.1% among male inmates in a penitentiary hospital in Georgia, USA. Considering the low incidence of T.B. in Florida (3.2/100 000 population), a 2.6% prevalence of mixed MTB infection is consistent with this hypothesis. Our findings, however, are contradictory to results from studies in India and China that found a prevalence of <0.4% and 3.5%, respectively, in settings with high T.B. incidence and high population density (Das et al., 2004; Pang et al., 2015). This suggests that the occurrence of mixed MTB infection in moderate and low-incidence T.B. settings may not be entirely due to clinical and socio-epidemiological factors but may be influenced by bacterial factors like the infectivity of the strain (Martín et al., 2011). Additionally, factors such as the sensitivity of the genotyping technique used, (Cohen et al., 2012) handling and processing of samples which may involve lapses in the decontamination procedures, and different culturing methods may significantly contribute to the differences in the proportion of mixed MTB infection observed in different environments (Pang et al., 2015). Variations in the prevalence of mixed MTB infections can also be attributed to differences in its detection rate, which is mainly dependent on the study design, sample size, and the method used in the genotyping of isolates (McIvor et al., 2017).
The MIRU- VNTR genotyping technique assesses heterogeneity from a restricted set of loci lowering its sensitivity and may also be limited in the differentiation of mixed MTB infection from allele evolution (Cohen et al., 2012; Cohen et al., 2011). Whole Genome Sequencing (WGS) provides the ability to distinguish between highly related but genetically distinct strains of MTB and stands as the method of choice in providing a more accurate estimation of the prevalence of mixed MTB infections.
Mixed MTB infection was not significantly associated with death in our study population. However, the existence of mixed MTB infections, as demonstrated by other investigations, have the potential to affect disease outcome adversely, especially among cases infected with multiple strains of different drug-susceptibility thresholds (Cohen et al., 2012; Tarashi et al., 2017). This may be due to the ability of the resistant strains to thrive under standard treatment regimen and the possible reactivation of the susceptible strains in a switch to MDR treatment. The presence of underlying resistance may also increase the risk of acquired-resistance under antibiotic pressure during standard treatment. The phenomenon is common in places with high T.B. burden and high prevalence of MDR-TB (Tarashi et al., 2017), which explains the lack of association observed in our study population.
The “dose-dependent-like” association between mixed T.B. infections and age, observed in our study suggest increased heterogeneity within the latent T.B. population over time with the age of latent infection. The concept is in line with theories that attribute optimal persistence levels of MTB strains in human to the latency period, with disease activation subject to the host immune system and the dose of the challenging MTB infection (Brites and Gagneux, 2015; O’Garra et al., 2013). The altered immunity associated with old age, explains the possible activation of these latent strains that might have survived over years of accumulation from numerous exposures to different MTB strains.
The presence of underlying morbidity among T.B. patients such as diabetes is associated with immunosuppression (Restrepo and Schlesinger, 2014), patients with diabetes may experience a delay in the clearance of the MTB pathogen after treatment and may have a high risk of reinfection (Baker et al., 2012). The increased odds observed in this study is consistent with findings from other studies reporting significant associations between diabetes and both MTB infection and increased risk of adverse treatment outcomes (Restrepo and Schlesinger, 2014; Martinez and Kornfeld, 2014). Patients with diabetes, like those with HIV infection, may be prone to multiple infections under the same mechanism.
Although some MTB lineages are known to induce various forms of immune responses and may create an environment to co-exist with other strains,(Pareek et al., 2017) most studies like ours, have not been able to establish a clear association between mixed infections and lineage. The East Asian (Beijing) lineage has been credited with its unique ability to generate a phenolic glycolipid that weakens the host’s innate immune responses and ability to fight infections and is known to cause a more aggressive form of the disease with a high proportion of adverse treatment outcomes (Merker et al., 2015; Kargarpour Kamakoli et al., 2020). However, other studies have failed to establish these findings (Thwaites et al., 2008; Pareek et al., 2017). Further investigation may be required to determine the association between mixed infections and MTB lineages.
The study findings are based on a cross-sectional analysis of routinely collected data and should not be interpreted as causal relationships. The lack of a prospective component to this study limited our ability to examine possible residual confounders, such as other comorbidities that patients may present with and the influence of socioeconomic status. The use of the MIRU-VNTR technique as compared to Whole Genome Sequencing (WGS) coupled with the study design limited our ability to differentiate between clonal heterogeneity and mixed infection resulting from either exogenous exposure to multiple MTB strains or superinfections.
The sample size used in the study was sufficiently large to give accurate results and provide a smaller margin of error with a satisfactory statistical power of one. The use of 24-locus MIRU-VNTR was an advantage that increased the sensitivity and the discriminative power in the detection of mixed infections in our study population as compared with studies (Nardell, 2004; Tarashi et al., 2017), that used less sensitive techniques like spoligotyping and RFLP.
5.0. Conclusion
The study demonstrates a low prevalence of mixed T.B. infections (2.6%) in a low-incidence T.B. setting. Mixed T.B. infection was not significantly associated with death. However, higher risk of mixed infections was associated with older individuals and patients with diabetes. The findings provide a comprehensive understanding of the epidemiological and clinical features of mixed infections in Florida, which is essential in T.B. case management and control strategies.
Future investigations should include prospective studies examining mixed infections and prevalent risk factors among T.B. patients, as well as the use of Whole Genome Sequencing (WGS), which provides a higher discriminatory power in differentiating between clonal heterogeneity due to microevolution of the MTB strain and mixed infections due to sequential or simultaneous exposure to multiple MTB strains.
Acknowledgements
We thank Dr. Sonja A. Rasmussen, Dr. Catherine L. Striley, Dr. John Glenn Morris Jr. and Dr. Awewura Kwara for their support and guidance in reviewing the manuscript. We are grateful to the team at the Florida Department of Health and the Southeastern National Tuberculosis Center for making available the de-identified surveillance data used in the analysis.
Funding Sources
This research work was supported by the University of Florida-University of Ghana Training Program in Tuberculosis and HIV Research in Ghana funded by Fogarty International Center at the National Institutes of Health [grant number TW010055 to MAB] and the NIH/NCATS Clinical and Translational Science Award to the University of Florida (KL2TR001429 to MNS). The funding agencies had no role in the design, conduct, analysis, interpretation, review, and approval of the manuscript.
Footnotes
Declaration of Interests
The authors declare no competing interests.
References
- Baker MA, Lin HH, Chang HY, Murray MB, 2012. The risk of tuberculosis disease among persons with diabetes mellitus: A prospective cohort study. Clinical Infectious Diseases 54, 818–825. 10.1093/cid/cir939 [DOI] [PubMed] [Google Scholar]
- Baldeviano-Vidalon G, Quispe-Torres N, Bonilla-Asalde C, Gastiaburu-Rodriguez D, Pro-Cuba J, Llanos-Zavalga F, 2005. Multiple infection with resistant and sensitive M . tuberculosis. Int J Tuberc Lung Dis 9, 1155–1160. [PubMed] [Google Scholar]
- Bozdogan H, 1987. Model selection and Akaike’s Information Criterion (AIC): The general theory and its analytical extensions. Psychometrika 52, 345–370. 10.1007/BF02294361 [DOI] [Google Scholar]
- Brites D, Gagneux S, 2015. Co-evolution of Mycobacterium tuberculosis and Homo sapiens. Immunological Reviews 264, 6–24. 10.1111/imr.12264 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Brites D, Gagneux S, 2012. Old and new selective pressures on Mycobacterium tuberculosis. Infection, Genetics and Evolution 12, 678–685. 10.1016/j.meegid.2011.08.010 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bursac Z, Gauss CH, Williams DK, Hosmer DW, 2008. Purposeful selection of variables in logistic regression. Source Code for Biology and Medicine 3, 1–8. 10.1186/1751-0473-3-17 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Chaves F, Dronda F, Alonso-sanz M, Noriega AR, 1999. Evidence of exogenous reinfection and mixed infection with more than one strain of Mycobacterium tuberculosis among Spanish HIV-infected inmates 615–620. [DOI] [PubMed] [Google Scholar]
- Chindelevitch L, Colijn C, Moodley P, Wilson D, Cohen T, Tanaka MM, 2016. ClassTR: Classifying Within-Host Heterogeneity Based on Tandem Repeats with Application to Mycobacterium tuberculosis Infections. PLoS Computational Biology 12, 1–16. 10.1371/journal.pcbi.1004475 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Cohen T, Chindelevitch L, Misra R, Kempner ME, Galea J, Moodley P, Wilson D, 2016. Within-host heterogeneity of mycobacterium tuberculosis infection is associated with poor early treatment response: A prospective cohort study. Journal of Infectious Diseases 213, 1796–1799. 10.1093/infdis/jiw014 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Cohen T, van Helden PD, Wilson D, Colijn C, Mclaughlin MM, Abubakar I, Warren RM, 2012. Mixed-strain Mycobacterium tuberculosis infections and the implications for tuberculosis treatment and control. Clinical Microbiology Reviews 25, 708–719. 10.1128/CMR.00021-12 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Cohen T, Wilson D, Wallengren K, Samuel EY, Murray M, 2011. Mixed-strain Mycobacterium tuberculosis infections among patients dying in a Hospital in KwaZulu-Natal, South Africa. Journal of Clinical Microbiology 49, 385–388. 10.1128/JCM.01378-10 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Corraini P, Olsen M, Pedersen L, Dekkers OM, Vandenbroucke JP, 2017. Effect modification, interaction and mediation: An overview of theoretical insights for clinical investigators. Clinical Epidemiology 9, 331–338. 10.2147/CLEP.S129728 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Coscolla M, Gagneux S, 2014. Consequences of genomic diversity in mycobacterium tuberculosis. Seminars in Immunology. 10.1016/j.smim.2014.09.012 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Coscolla M, Gagneux S, 2010. Does M. tuberculosis genomic diversity explain disease diversity? Drug Discovery Today: Disease Mechanisms. 10.1016/j.ddmec.2010.09.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Das S, Narayanan S, Hari L, Mohan NS, Somasundaram S, Selvakumar N, Narayanan PR, 2004. Simultaneous infection with multiple strains of Mycobacterium tuberculosis identified by restriction fragment length polymorphism analysis. International Journal of Tuberculosis and Lung Disease 8, 267–270. [PubMed] [Google Scholar]
- Dickman KR, Nabyonga L, Kateete DP, Katabazi FA, Asiimwe BB, Mayanja HK, Okwera A, Whalen C, Joloba ML, 2010. Detection of multiple strains of Mycobacterium tuberculosis using MIRU-VNTR in patients with pulmonary tuberculosis in Kampala, Uganda. BMC Infectious Diseases 10, 349 10.1186/1471-2334-10-349 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Fang R, Li X, Li J, Wu J, Shen X, Gui X, DeRiemer K, Liu L, Mei J, Gao Q, 2008. Mixed infections of Mycobacterium tuberculosis in tuberculosis patients in Shanghai, China. Tuberculosis. 10.1016/j.tube.2008.02.002 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Fitzpatrick LK, Okwera A, Mugerwa R, Ridzon R, Ellner J, Onorato I, 2002. An investigation of suspected exogenous reinfection in tuberculosis patients in Kampala, Uganda. International Journal of Tuberculosis and Lung Disease 6, 550–552. 10.5588/09640569513084 [DOI] [PubMed] [Google Scholar]
- Ghosh S, Moonan PK, Cowan L, Grant J, Kammerer S, Navin TR, 2012. Tuberculosis Genotyping Information Management System: Enhancing Tuberculosis Surveillance in the United States. Infection, Genetics and Evolution 12, 782–788. 10.1016/j.meegid.2011.10.013 [DOI] [PubMed] [Google Scholar]
- Guide T, 2005. Multilocus Variable Number Tandem Repeat Genotyping of Mycobacterium tuberculosis.
- Kargarpour Kamakoli M, Farmanfarmaei G, Masoumi M, Khanipour S, Gharibzadeh S, Sola C, Fateh A, Siadat SD, Refregier G, Vaziri F, 2020. Prediction of the hidden genotype of mixed infection strains in Iranian tuberculosis patients. International Journal of Infectious Diseases 95, 22–27. 10.1016/j.ijid.2020.03.056 [DOI] [PubMed] [Google Scholar]
- Li C, 2013. Little’s test of missing completely at random. Stata Journal 13, 795–809. 10.1177/1536867x1301300407 [DOI] [Google Scholar]
- Mallard K, McNerney R, Crampin AC, Houben R, Ndlovu R, Munthali L, Warren RM, French N, Glynn JR, 2010. Molecular detection of mixed infections of Mycobacterium tuberculosis strains in sputum samples from patients in Karonga District, Malawi. Journal of clinical microbiology 48, 4512–4518. 10.1128/JCM.01683-10 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Martín A, Herranz M, Navarro Y, Lasarte S, Serrano MJR, Bouza E, De Viedma DG, 2011. Evaluation of the inaccurate assignment of mixed infections by mycobacterium tuberculosis as exogenous reinfection and analysis of the potential role of bacterial factors in reinfection. Journal of Clinical Microbiology 49, 1331–1338. 10.1128/JCM.02519-10 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Martinez N, Kornfeld H, 2014. Diabetes and immunity to tuberculosis. European Journal of Immunology 44, 617–626. 10.1002/eji.201344301 [DOI] [PMC free article] [PubMed] [Google Scholar]
- McIvor A, Koornhof H, Kana BD, 2017. Relapse, re-infection and mixed infections in tuberculosis disease. Pathogens and Disease 75, 1–16. 10.1093/femspd/ftx020 [DOI] [PubMed] [Google Scholar]
- Merker M, Blin C, Mona S, Duforet-Frebourg N, Lecher S, Willery E, Blum MGB, Rüsch-Gerdes S, Mokrousov I, Aleksic E, Allix-Béguec C, Antierens A, Augustynowicz-Kopeć E, Ballif M, Barletta F, Beck HP, Barry CE, Bonnet M, Borroni E, Campos-Herrero I, Cirillo D, Cox H, Crowe S, Crudu V, Diel R, Drobniewski F, Fauville-Dufaux M, Gagneux S, Ghebremichael S, Hanekom M, Hoffner S, Jiao WW, Kalon S, Kohl TA, Kontsevaya I, Lillebæk T, Maeda S, Nikolayevskyy V, Rasmussen M, Rastogi N, Samper S, Sanchez-Padilla E, Savic B, Shamputa IC, Shen A, Sng LH, Stakenas P, Toit K, Varaine F, Vukovic D, Wahl C, Warren R, Supply P, Niemann S, Wirth T, 2015. Evolutionary history and global spread of the Mycobacterium tuberculosis Beijing lineage. Nature Genetics 47, 242–249. 10.1038/ng.3195 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Muwonge A, Kankya C, Olea-popelka F, Biffa D, Ssengooba W, Berit D, Skjerve E, Johansen TB, 2013. Infection , Genetics and Evolution Molecular investigation of multiple strain infections in patients with tuberculosis in Mubende district , Uganda. Infection, Genetics and Evolution 17, 16–22. 10.1016/j.meegid.2013.03.039 [DOI] [PubMed] [Google Scholar]
- Nardell EA, 2004. Catching Droplet Nuclei. American Journal of Respiratory and Critical Care Medicine 169, 553–554. 10.1164/rccm.2401003 [DOI] [PubMed] [Google Scholar]
- O’Garra A, Redford PS, McNab FW, Bloom CI, Wilkinson RJ, Berry MPR, 2013. The Immune Response in Tuberculosis, Annual Review of Immunology. 10.1146/annurev-immunol-032712-095939 [DOI] [PubMed] [Google Scholar]
- Pang Y, Zhou Y, Wang S, Song Y, Ou X, Zhao B, Zhang Z, Zhao Y, 2015. Prevalence and risk factors of mixed Mycobacterium tuberculosis complex infections in China. Journal of Infection. 10.1016/j.jinf.2015.03.012 [DOI] [PubMed] [Google Scholar]
- Pareek M, Evans J, Innes J, Smith G, Hingley-wilson S, 2017. Europe PMC Funders Group Ethnicity and mycobacterial lineage as determinants of tuberculosis disease phenotype 68, 221–229. 10.1136/thoraxjnl-2012-201824.Ethnicity [DOI] [PMC free article] [PubMed] [Google Scholar]
- Restrepo BI, Schlesinger LS, 2014. Impact of diabetes on the natural history of tuberculosis. Diabetes Research and Clinical Practice 106, 191–199. 10.1016/j.diabres.2014.06.011 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Van Rie A, Victor TC, Richardson M, Johnson R, Van Der Spuy GD, Murray EJ, Beyers N, Van Pittius NCG, Van Helden PD, Warren RM, 1998. Reinfection and Mixed Infection Cause Changing Mycobacterium tuberculosis Drug-Resistance Patterns. 10.1164/rccm.200503-449OC [DOI] [PMC free article] [PubMed] [Google Scholar]
- Séraphin MN, Lauzardo M, Doggett RT, Zabala J, Morris JG, Blackburn JK, 2016. Spatiotemporal clustering of mycobacterium tuberculosis complex genotypes in Florida: Genetic diversity segregated by country of birth. PLoS ONE 11 10.1371/journal.pone.0153575 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Shamputa IC, Jugheli L, Sadradze N, Willery E, Portaels F, Supply P, Rigouts L, 2006. Mixed infection and clonal representativeness of a single sputum sample in tuberculosis patients from a penitentiary hospital in Georgia. Respiratory Research 7, 1–10. 10.1186/1465-9921-7-99 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Shamputa IC, Rigouts L, Eyongeta LA, Abdullah N, Aila E, Van Deun A, Salim AH, Willery E, Locht C, Supply P, 2004. Genotypic and Phenotypic Heterogeneity among Mycobacterium tuberculosis Isolates from Pulmonary Tuberculosis Patients 42, 5528–5536. 10.1128/JCM.42.12.5528 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Sharma K, Verma R, Advani J, Chatterjee O, Solanki HS, Sharma A, Varma S, Modi M, Ray P, Mukherjee KK, Sharma M, Dhillion MS, Suar M, Chatterjee A, Pandey A, Prasad TSK, Gowda H, 2017. Whole Genome Sequencing of Mycobacterium tuberculosis Isolates from Extrapulmonary Sites. OMICS A Journal of Integrative Biology 21, 413–425. 10.1089/omi.2017.0070 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Shin SS, Modongo C, Baik Y, Allender C, Lemmer D, Colman RE, Engelthaler DM, Warren RM, Zetola NM, 2018. Mixed Mycobacterium tuberculosis-Strain Infections Are Associated With Poor Treatment Outcomes Among Patients With Newly Diagnosed Tuberculosis, Independent of Pretreatment Heteroresistance. The Journal of infectious diseases 218, 1974–1982. 10.1093/infdis/jiy480 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Tarashi S, Fateh A, Mirsaeidi M, Siadat SD, Vaziri F, 2017. Mixed infections in tuberculosis: The missing part in a puzzle. Tuberculosis. 10.1016/j.tube.2017.09.004 [DOI] [PubMed] [Google Scholar]
- Thwaites G, Caws M, Thi Hong Chau T, Thi Ngoc Lan N, Nguyet Thu Huyen M, Gagneux S, Thi Hoang Anh P, Quang Tho D, Torok E, Thi Quynh Nhu N, Thi Hong Duyen N, Minh Duy P, Richenberg J, Simmons C, Tinh Hien T, Farrar J, 2008. Relationship between Mycobacterium tuberculosis Genotype and the Clinical Phenotype of Pulmonary and Meningeal Tuberculosis. JOURNAL OF CLINICAL MICROBIOLOGY 46, 1363–1368. 10.1128/JCM.02180-07 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Viedma D, Rodriguez NA, Andre S, Jesu M, Bouza E, 2005. Characterization of Clonal Complexity in Tuberculosis by Mycobacterial Interspersed Repetitive Unit – Variable-Number Tandem Repeat Typing 43, 5660–5664. 10.1128/JCM.43.11.5660 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Warren RM, Victor TC, Streicher EM, Richardson M, Beyers N, Van Pittius NCG, Van Helden PD, 2004. Patients with Active Tuberculosis often Have Different Strains in the Same Sputum Specimen. Am J Respir Crit Care Med.,p. 2819. doi: 10.1164/rccm.200305-714OC. Epub 2003 Dec 30. PMID: 14701710. [DOI] [PubMed] [Google Scholar]
- World Health Organisation, 2019. Global Tuberculosis Report WHO.s.l: WHO; 2019. [Google Scholar]