Abstract
Rationale: Airborne transmission of Mycobacterium tuberculosis results from incompletely characterized host, bacterial, and environmental factors. Sputum smear microscopy is associated with considerable variability in transmission.
Objectives: To evaluate the use of cough-generated aerosols of M. tuberculosis to predict recent transmission.
Methods: Patients with pulmonary tuberculosis (TB) underwent a standard evaluation and collection of cough aerosol cultures of M. tuberculosis. We assessed household contacts for new M. tuberculosis infection. We used multivariable logistic regression analysis with cluster adjustment to analyze predictors of new infection.
Measurements and Main Results: From May 2009 to January 2011, we enrolled 96 sputum culture-positive index TB cases and their 442 contacts. Only 43 (45%) patients with TB yielded M. tuberculosis in aerosols. Contacts of patients with TB who produced high aerosols (≥10 CFU) were more likely to have a new infection compared with contacts from low-aerosol (1–9 CFU) and aerosol-negative cases (69%, 25%, and 30%, respectively; P = 0.009). A high-aerosol patient with TB was the only predictor of new M. tuberculosis infection in unadjusted (odds ratio, 5.18; 95% confidence interval, 1.52–17.61) and adjusted analyses (odds ratio, 4.81; 95% confidence interval, 1.20–19.23). Contacts of patients with TB with no aerosols versus low and high aerosols had differential tuberculin skin test and interferon-γ release assay responses.
Conclusions: Cough aerosols of M. tuberculosis are produced by a minority of patients with TB but predict transmission better than sputum smear microscopy or culture. Cough aerosols may help identify the most infectious patients with TB and thus improve the cost-effectiveness of TB control programs.
The concentration of Mycobacterium tuberculosis in sputum of patients with TB, usually defined by acid-fast bacilli (AFB)-positive versus AFB-negative smear microscopy, has been associated with transmission for decades. Yet, there is experimental and epidemiologic evidence of marked variability in transmission from patients with TB.
Direct measurement of cough-generated aerosols from patients with pulmonary TB seems to be a new quantitative predictor of infectiousness. Our results have several implications. First, by identifying the most infectious patients with pulmonary TB, our findings belie the current paradigm that considers all sputum AFB-positive patients as homogeneously infectious, providing a framework for more rational and cost-effective infection control decisions. Second, by reducing the number of TB contacts at highest risk of infection while increasing the yield of infected individuals, aerosols may improve the implementation of latent TB infection treatment programs by targeting contacts by virtue of their exposure. Finally, M. tuberculosis aerosol measurement provides a quantitative surrogate of inhaled dose in exposed contacts that reduces exposure misclassification and thus is likely to offer insights into TB immunopathogenesis and facilitate studies of vaccines, drugs, and immune responses.
Tuberculosis (TB) remains one of the most significant challenges in global health (1). Exclusively transmitted by the airborne route, the TB disease cycle begins when a human host inhales infectious aerosols (airborne particles <5 µm in aerodynamic diameter) containing viable Mycobacterium tuberculosis (2). There is experimental (3, 4) and epidemiologic (5–10) evidence of marked variability in transmission from patients with TB. Riley and coworkers (11) demonstrated this variability of infectiousness in pulmonary TB more than 50 years ago but this knowledge has never been translated into practice or policy (4, 11), and the mechanisms underlying this variability remain poorly understood.
Successful TB transmission results from a variety of incompletely characterized determinants. These factors relate to the contagiousness of the source case; environmental conditions, such as ambient air humidity, ventilation, and exposure to ultraviolet light; and the pulmonary ventilation rate in exposed human hosts (12). Risk factors for transmission include the sputum bacillary concentration and cough frequency in the source case, and the proximity or duration of the exposure (13). The concentration of M. tuberculosis in sputum of patients with TB (14), usually defined by acid-fast bacilli (AFB)–positive versus AFB-negative smear microscopy, has been associated with transmission for decades. As a result, the sputum AFB smear is widely considered to define TB infectiousness. However, it is estimated that less than 30% of sputum AFB-positive source cases transmit infection to their contacts (15). Conversely, sputum AFB-negative sources have accounted for up to 13–17% of M. tuberculosis transmission in several settings (16–18). Thus, there is a need for an improved laboratory marker of TB transmission. We conducted this study among patients with pulmonary TB with sputum AFB greater than or equal to 1+ at screening to determine if colony-forming units (CFU) of M. tuberculosis in cough-generated aerosols predict recent infection in household contacts and to compare its predictive ability with that of sputum AFB smear microscopy grade.
We enrolled consecutive patients with pulmonary TB attending the Mulago Hospital National Tuberculosis and Leprosy Program (NTLP) clinic in Kampala, Uganda, and their household contacts. Patients with TB were eligible if they (1) were 18 years or older, (2) had a new TB episode with at least one sputum specimen that was initially AFB greater than or equal to 1+ with subsequent growth of M. tuberculosis in culture, (3) were untreated or had received 5 days or less of anti-TB treatment, and (4) lived with three or more contacts. We excluded patients with medical conditions that could be worsened by cough (19) or if they were too ill to consent, or unable to understand or to comply with the protocol. A contact was defined as an individual of any age sleeping under the same roof with the index TB case for 3 or more months before enrollment and no prior history of TB disease. We obtained informed consent and assent in accordance with age-specific ethical guidelines from participating institutions.
Evaluation of TB cases included three sputa specimens for AFB smear microscopy (auramine O fluorescent stain) (20) and cultures performed on 7H10 Middlebrook agar and liquid media (BACTEC MGIT 960, Becton Dickinson, Franklin Lakes, NJ). The radiologic extent of disease was graded on a four-category ordinal scale. We evaluated cough using a self-reported visual analog scale (21) and three cough peak flow rates (MicroDirect, Inc., Lewiston, ME). We cultured M. tuberculosis from cough aerosols using a cough aerosol sampling system (CASS), a simple, safe, and reproducible method initially developed in the United States (22) and subsequently evaluated at this site with minor modifications (19). A detailed description of the aerosol sampling method is provided in the online supplement. Briefly, patients were instructed to voluntarily cough into the CASS as frequently and as strongly as they could for 5 minutes. They then rested for 5 minutes before a second 5-minute period of coughing. After aerosol collection, the CASS agar plates were read at 1 week to detect contaminants, and then at 3 and 6 weeks to count CFU of M. tuberculosis, which were confirmed by polymerase chain reaction; the 6-week CFU counts were taken as the primary outcome measure of the cough aerosol collection. Before aerosol collection, we recorded room temperature and humidity. The NTLP program provided standard TB treatment and offered routine HIV testing. HIV-infected patients had a CD4 performed and were referred for HIV care.
We used modified US Centers for Disease Control and Prevention recommendations for household contact investigations (23). Before enrollment, we trained staff in tuberculin skin test (TST) (Tubersol; Sanofi Pasteur, Swiftwater, PA) placement and reading and completed interreader and intrareader evaluations (kappa agreement >90%). We obtained blood for an interferon-γ release assay (IGRA) (QuantiFERON Gold-In-Tube; Cellestis, Chadstone, Australia) before TST placement. Household contacts with a negative TST or IGRA at baseline were retested after 6 weeks. To minimize misclassification of TST converters (24), we used two different definitions of TST conversion (see the online supplement). Contacts were not tested for HIV and thus, HIV infection was not taken into account in determining TST and IGRA results. We visited the participants’ dwellings to assess individual contact time, crowding, and ventilation conditions. Following NTLP recommendations (25), infected contacts at highest risk of developing active TB disease and secondary TB suspects were referred for evaluation and treatment.
Our primary exposure variable was the cough aerosol status of the index TB patient in the household. To evaluate a relationship between number of CFU in aerosol and TST conversion, we divided aerosol status into three groups after first looking at the trend in risk of TST conversion by finely divided categories of CFU. Our final groupings were (1) aerosol negative; (2) low aerosol (1–9 CFU); and (3) high aerosol (≥10 CFU). We chose a 10-CFU cut-off because this was the point at which we noted an increase in TST conversion risk. We also performed a sensitivity analysis using different CFU cut-points. Our primary outcome was TST conversion in contacts (using Criterion 1). To examine the possibility of TST or IGRA boosting in contacts, we performed a quantitative analysis of interferon-γ (international units per milliliter) values for each group of contacts. We use standard box plots to represent normalized (QuantiFERON Gold-In-Tube: TB antigen–nil) interferon-γ levels after calculating a delta value for each individual from paired blood samples (6-wk level–baseline level). We calculated descriptive statistics to identify clinical and demographic characteristics and differences between households of aerosol-positive (CFU >0) and aerosol-negative index patients with TB. To determine factors associated with TST conversion in contacts, we performed unadjusted analyses using our primary definition of conversion. Predictor variables included clinical and epidemiologic characteristics of the index case and household contacts, and environmental factors. Variables associated with TST conversion in contacts in the unadjusted analysis (P < 0.15) were retained in the final multivariable analysis. To account for clustering within a household, we fit logistic regression models of conversion using generalized estimating equations (SAS version 9.1, SAS Institute, Inc., Cary, NC) (26) using an independent working correlation matrix. Because we identified no significant interactions, we present only main effects. We present our results as unadjusted and adjusted odds ratios (OR) with 95% confidence intervals (CI).
The study was approved by the Makerere University Faculty of Medicine Research and Ethics Committee, the Uganda National Council for Science and Technology, and the Institutional Review Boards of the University of Medicine and Dentistry of New Jersey and Boston University Medical Campus.
Between May 2009 and January 2011, we screened 838 patients with TB (Figure 1). The 728 patients excluded were either ineligible (n = 710) or refused (n = 18). Of 110 enrolled families, 14 were withdrawn because they declined participation, were lost to follow-up, or had missing results. Eighteen (4%) eligible contacts were excluded. Thus, this analysis includes 96 index TB cases and their 442 household contacts.
Figure 1. Study profile. AFB = acid-fast bacilli; CASS = cough aerosol sampling system; HHC = household contacts; IC = index tuberculosis cases; IGRA = interferon-γ release assay; TB = tuberculosis; TST = tuberculin skin test.
[More] [Minimize]Index patients with TB were a median age of 28.9 years with advanced pulmonary TB disease as measured by duration of cough (median, 12 wk); sputum smear microscopy; solid and liquid culture results; and, chest radiograph (Table 1). The 21 (22%) HIV-infected patients had a median CD4+ cell count of 264 cells/µl and 12 (57%) were on antiretroviral therapy at enrollment.
| Cough Aerosol Positive | |||
|---|---|---|---|
| Characteristic | Cough Aerosol Negative | Low Aerosol (1–9 CFU) | High Aerosol (≥10 CFU) |
| Index TB cases | |||
| N | 53 | 18 | 25 |
| Age, yr | 29.3 (23.1–36.5) | 28.4 (20.8–40.3) | 28.5 (24–40) |
| Sex | |||
| Male | 26 (49) | 8 (44) | 15 (60) |
| Female | 27 (51) | 10 (56) | 10 (40) |
| Body mass index, kg/m2 | 19.0 (17.8–20.7) | 18.9 (18.1–20.7) | 18.7 (17.2–20.7) |
| Karnofsky performance score | |||
| 70 | 2 (4) | 0 (0) | 0 (0) |
| 80 | 2 (4) | 0 (0) | 2 (8) |
| 90 | 49 (92) | 18 (100) | 23 (92) |
| HIV status | |||
| Uninfected | 44 (83) | 10 (56) | 19 (76) |
| Infected | 8 (15) | 8 (44) | 5 (20) |
| CD4 cell count, cells/ml | 315 (226–351) | 172 (116–264) | 288 (217–360) |
| Weeks sick before enrollment | 12 (8–20) | 12 (8–12) | 12 (8–15) |
| Chest radiograph findings | |||
| Extent of lung disease | |||
| Normal | 0 (0) | 1 (6) | 0 (0) |
| Minimal | 4 (8) | 2 (13) | 1 (4) |
| Moderate | 15 (29) | 6 (37) | 9 (38) |
| Far advanced | 32 (63) | 7 (44) | 14 (58) |
| Cavitations | |||
| Absent | 14 (27) | 8 (50) | 8 (33) |
| Present | 37 (73) | 8 (50) | 16 (67) |
| Sputum characteristics | |||
| Volume, ml | 5 (4–10) | 5 (3–7.5) | 5 (5–15) |
| Acid-fast bacilli smear | |||
| Negative | 1 (2) | 0 (0) | 0 (0) |
| Scanty | 1 (2) | 0 (0) | 0 (0) |
| 1+ | 8 (15) | 3 (17) | 3 (12) |
| 2+ | 11 (21) | 2 (11) | 1 (4) |
| 3+ | 32 (60) | 13 (72) | 21 (84) |
| MGIT 960 culture (DTP) | 6 (4–8) | 5 (4–8) | 6 (4–10) |
| CASS characteristics | |||
| Cough during aerosol collection | |||
| Weak | 21 (40) | 8 (44) | 7 (28) |
| Strong | 32 (60) | 10 (56) | 18 (72) |
| Days on TB treatment before aerosol collection | 1 (1–2) | 1 (1–2) | 1 (0–1) |
| Aerosol CFU | 0 (0) | 2 (1–5) | 26 (17–55) |
| Household contacts | |||
| N | 254 | 90 | 98 |
| Age | 12 (7–24) | 15 (8–24) | 13 (6–26) |
| BCG scar | |||
| Present | 188 (74) | 68 (76) | 74 (76) |
| Absent or uncertain | 66 (26) | 22 (24) | 24 (24) |
| HIV status | |||
| Infected | 4 (2) | 3 (4) | 3 (3) |
| Uninfected | 232 (91) | 67 (74) | 78 (80) |
| Unknown | 18 (7) | 20 (22) | 17 (17) |
| TST induration diameter, mm | |||
| Baseline | 18 (0–24) | 20 (0–25) | 20 (4–25) |
| Six weeks (in contacts with first TST <5 mm only) | 0 (0–14) | 0 (0–18) | 20 (0–24) |
| IGRA positive, n/N (%) | |||
| Baseline | 163/239 (68) | 56/83 (68) | 67/96 (70) |
| Six weeks (in contacts with first IGRA negative only) | 28/77 (36) | 16/34 (47) | 21/27 (78) |
Fifty-five percent of patients with TB produced no aerosols with culturable M. tuberculosis; 19% produced low aerosols (1–9 CFU); and 26% high aerosols (≥10 CFU). Cough aerosol positivity varied across sputum AFB smear microscopy grades (Figure 2A); the proportion of patients with high aerosols in each of the sputum-positive AFB smear grades (1+, 2+, and 3+) was 21%, 7%, and 32%, respectively. There was considerable variation in CFU (median, 14 CFU; range, 1–378 CFU) among aerosol-positive patients (Figure 2B). Patients with high aerosols had median cough peak flow rates that were 60–80 L/min higher than patients with negative or low aerosols (see Table E1 in the online supplement).
Figure 2. Distribution of CFU of Mycobacterium tuberculosis in aerosols from 96 patients with pulmonary tuberculosis. (A) Distribution of cough aerosol results according to sputum acid-fast bacilli (AFB) smear microscopy grade. Patients with 0 CFU M. tuberculosis in aerosols (white circles). Patients with 1–9 CFU of M. tuberculosis in (low) aerosol (gray circles). Patients with greater than or equal to 10 CFU of M. tuberculosis in (high) aerosol (black circles). Dashed horizontal lines indicate the median for subjects who produced aerosol (CFU ≥1). Below the graph, the median CFU and range are given for subjects who produced aerosol (CFU ≥1). Because aerosol CFU data are presented on a log scale, 1 was added to all CFU for plotting purposes. (B) Distribution of CFU counts in cough aerosols. TB = tuberculosis.
[More] [Minimize]The number of contacts exposed to aerosol-negative, low- and high-aerosol index TB cases was 254, 90, and 98, respectively (Table 1). Enrolled contacts were 52% female, had a median age of 13 years, and 74% had a bacille Calmette-Guérin vaccination scar. Nearly all were biologically related to the TB index case. Participant dwellings were small and crowded; 52% of contacts slept in the same bed or room as the index TB case. At baseline, the prevalence of TB infection in contacts was comparable in all three cough-aerosol groups by either TST greater than or equal to 10 mm or IGRA criteria (see Figure E1A). The use of different TST cut-off values (5, 10, and 15 mm of induration) had a limited impact on TST sensitivity (variation between 5% and 8%).
After excluding contacts that were infected at baseline, the number of contacts “at-risk” of TST conversion in each of the three groups was 66 (aerosol negative), 28 (low aerosol), and 26 (high aerosol). Compared with contacts of aerosol-negative patients with TB (30%), the odds of TST conversion in contacts of low-aerosol cases (25%) was similar (OR, 0.77; 95% CI, 0.27–2.17; P = 0.62) but over fivefold greater (OR, 5.18; 95% CI, 1.52–17.61; P = 0.009) in contacts of high-aerosol patients (69%) (Figure 3). Although not reaching statistical significance, a similar pattern was seen when we only included contacts (data not shown) of patients with TB with sputum AFB greater than or equal to 3+ (OR, 3.14; 95% CI, 0.75–13.12; P = 0.12). When we used IGRA alone (OR, 9.53; 95% CI, 2.51–36.24; P = 0.0009) or TST and IGRA conversion together (OR, 20.9; 95% CI, 3.94–110.8; P = 0.0004) to define new TB infection, we observed a stronger positive association with cough aerosols (Figure 3A; see Figure E2A). By comparison, the same analysis using sputum AFB smear grade to classify exposure groups did not show clear or consistent risk stratification (Figure 3A).
Figure 3. Tuberculin skin test (TST) and interferon-γ release assay (IGRA) results in household contacts uninfected at baseline, according to aerosol or sputum results in index tuberculosis (TB) cases. (A) Proportion of contacts with TST, IGRA, or TST and IGRA (both together) conversion at 6 weeks. We present data using two different criteria to define TST conversion. Criterion 1: first TST less than 10 mm, second TST greater than or equal to 10 mm, and difference between first and second TST greater than or equal to 10 mm. Criterion 2: first TST less than 5 mm, second TST greater than or equal to 10 mm, and difference between first and second TST greater than or equal to 6 mm. Comparisons (odds ratios adjusted for clustering and 95% confidence intervals) are based on Criterion 1. The n under each column denotes the number of contacts fulfilling each outcome criterion. The N under each column denotes the number of household contacts “at risk” of conversion in each exposure category. We highlight the N to emphasize a change in exposure categories when using aerosol versus sputum. (B) Quantitative IGRA analysis. Standard box plots of interferon-γ values for each group of contacts. Values are normalized interferon-γ (TB antigen − Nil) results after calculating a delta for each individual from paired blood samples (6-wk level − baseline level).
[More] [Minimize]To examine the possibility of boosting, we analyzed TST and IGRA results as a continuous variable. The TST induration size in contacts that developed a TST greater than or equal to 5 mm after the baseline visit was larger in contacts from high-aerosol households (Table 1); similarly, we observed a strong dose–response correlation between interferon-γ levels and aerosol CFU (Figure 3B). Furthermore, a sensitivity analysis varying CFU cut-points showed similar results but additionally suggests that higher cut-points to define high aerosol production improves the predictive value of cough aerosols and reveals a consistently increasing trend with higher aerosol CFU (see Figure E2B).
By 6 weeks (see Figure E1B), the proportion of contacts with TST greater than or equal to 10 mm was similar in contacts from aerosol-negative (78%) and low-aerosol (76%) households (OR, 0.85; 95% CI, 0.41–1.79; P = 0.67) but significantly higher (92%) in contacts from high-aerosol households (OR, 3.10; 95% CI, 1.13–8.48; P = 0.03). This result did not change when we used a greater than or equal to 5 mm (P = 0.03) or greater than or equal to 15 mm (P = 0.04) TST cut-off level or IGRA positivity (P = 0.03) to define TB infection (see Figure E1B).
The median age (range) of contacts with TST conversion in each of the three aerosol groups (negative, low, and high) was 9 (0.6–76), 11 (2–26), and 8 (1–43) years, respectively. In analyzing index TB patients, household contact, and environmental factors associated with TST conversion in contacts (Table 2), the only significant factor in an unadjusted analysis was the TB case producing high aerosols (OR, 5.17; 95% CI, 1.52–17.61). In a multivariable model adjusted for clustering and other covariates, high aerosol status remained the only variable associated with TST conversion (OR, 4.81; 95% CI, 1.20–19.23; P = 0.03). This result did not significantly change using other multivariable models.
| Household Contacts | ||||||
|---|---|---|---|---|---|---|
| Characteristic | Number At Risk | Number (%) with TST Conversion | Unadjusted* Odds Ratio (95% Confidence Interval) | P Value‡ | Adjusted Odds Ratio (95% Confidence Interval) | P Value |
| Index cases | ||||||
| Age (per 10-yr increment) | 0.93 (0.56–1.55) | 0.79 | ||||
| Sex | ||||||
| Male | 53 | 18 (34) | 1 | |||
| Female | 67 | 27 (40) | 1.31 (0.68–2.52) | 0.41 | ||
| Body mass index, kg/m2 | ||||||
| <15.0 | 2 | 1 (50) | 1 | |||
| 15.0–18.5 | 33 | 13 (39) | 0.65 (0.29–1.46) | |||
| >18.5 | 78 | 30 (38) | 0.63 (0.32–1.22) | 0.68 | ||
| HIV status | ||||||
| Infected | 19 | 6 (32) | 1 | |||
| Uninfected | 100 | 39 (39) | 1.39 (0.49–3.91) | 0.55 | ||
| Family size | ||||||
| 2–3 contacts | 27 | 7 (26) | 1 | |||
| 4–5 contacts | 42 | 17 (40) | 1.94 (0.56–6.72) | |||
| ≥6 contacts | 51 | 21 (41) | 1.99 (0.51–7.83) | 0.49 | ||
| Cough measurements at baseline | ||||||
| Visual analog scale (per unit increase) | 1.08 (0.85–1.37) | 0.51 | 1.00 (0.93–1.06) | 0.12 | ||
| Cough peak flow§ (per 10 L/min increase) | 1.00 (0.99–1.01) | 0.91 | ||||
| Cough during aerosol collection | ||||||
| Weak | 46 | 15 (33) | 1 | |||
| Strong | 74 | 30 (41) | 1.41 (0.56–3.56) | 0.48 | ||
| Chest radiograph | ||||||
| Normal/minimal | 14 | 7 (50) | 1 | |||
| Moderate | 49 | 18 (37) | 0.58 (0.16–2.17) | |||
| Far advanced | 49 | 19 (39) | 0.63 (0.19–2.11) | 0.73 | ||
| Cavitations | ||||||
| Absent | 43 | 15 (35) | 1 | |||
| Present | 69 | 29 (42) | 1.35 (0.50–3.66) | 0.55 | ||
| Sputum appearance | ||||||
| Mucoid | 35 | 14 (40) | 1 | |||
| Mucopurulent | 27 | 6 (22) | 0.43 (0.12–1.52) | |||
| Mucosalivary | 32 | 12 (38) | 0.90 (0.24–3.37) | |||
| Salivary/watery | 24 | 12 (50) | 1.50 (0.29–7.68) | 0.45 | ||
| Sputum acid-fast bacilli smear | ||||||
| Negative/scanty/ 1+ | 24 | 12 (50) | 1 | |||
| 2+ | 26 | 6 (23) | 0.30 (0.05–1.68) | |||
| 3+ | 70 | 27 (39) | 0.63 (0.16–2.48) | 0.42 | ||
| 7H11 culture, CFU | ||||||
| <200 | 16 | 4 (25) | 1 | |||
| ≥200 | 104 | 41 (39) | 1.95 (0.64–6.00) | 0.30 | ||
| MGIT 960 culture (DTP) | ||||||
| ≥7 | 44 | 14 (32) | 1 | 1 | ||
| <7 | 71 | 26 (37) | 1.24 (0.42–3.69) | 0.70 | 1.14 (0.38–3.41) | 0.81 |
| Cough aerosol, CFU | ||||||
| Negative | 66 | 20 (30) | 1 | 1 | ||
| Low (1–9 CFU) | 28 | 7 (25) | 0.77 (0.27–2.17) | 0.83 (0.29–2.42) | 0.74 | |
| High (≥10 CFU) | 26 | 18 (69) | 5.17 (1.52–17.61) | 0.08 | 4.81 (1.20–19.23) | 0.03 |
| Household contacts | ||||||
| Age, yrs | ||||||
| ≥18 | 32 | 10 (31) | 1 | 1 | ||
| 5–17 | 63 | 24 (38) | 1.35 (0.60–3.07) | 1.14 (0.44–2.97) | 0.97 | |
| <5 | 23 | 11 (48) | 2.02 (0.74–5.52) | 0.40 | 1.29 (0.40–4.19) | 0.66 |
| Sex | ||||||
| Male | 59 | 20 (34) | 1 | |||
| Female | 61 | 25 (41) | 1.35 (0.49–3.77) | 0.56 | ||
| BCG scar | ||||||
| Absent/uncertain | 31 | 11 (35) | 1 | |||
| Present | 89 | 34 (38) | 1.12 (0.48–2.61) | 0.78 | ||
| Dwellings | ||||||
| Number of habitable rooms | 36 (40) | |||||
| ≥2 | 90 | 8 (36) | 1 | |||
| <2 | 22 | 0.86 (0.23–3.15) | 0.81 | |||
| Sleeping arrangement with index | ||||||
| Different room | 71 | 23 (32) | 1 | 1 | ||
| Same room, different bed | 39 | 19 (49) | 1.98 (0.79–4.98) | 1.06 (0.38–2.94) | 0.91 | |
| Same room, same bed | 10 | 3 (30) | 0.89 (0.18–4.37) | 0.33 | 0.64 (0.10–4.08) | 0.64 |
This is the first study to demonstrate an association between cough-generated aerosols of M. tuberculosis and transmission among humans. Our study design and use of TST and IGRA to measure incident TB infection provide robust evidence that exposure to high aerosols increases the risk of infection in contacts, despite high background infection rates in this setting. The results from this study suggest that a minority of patients with pulmonary TB transmit more than others (27, 28), and that in future studies, cough aerosols may serve as a more refined predictor of infectiousness and a more precise surrogate of exposure on inhaled dose in recently infected individuals. The findings from this study have generated several important questions that need to be investigated in future work.
We have now performed aerosol studies in more than 200 patients with TB and have found that only a minority of patients with pulmonary TB generates aerosols during coughing. Although unexplained at this time, we doubt that this observation is spurious or is related to an insensitive aerosol collection method. Whereas in our first two studies less than 30% of TB cases produced aerosols (19, 22), the greater frequency (45%) of aerosol-positive patients in this study is explained by our recruitment strategy. We collected aerosols earlier and enrolled mostly ambulatory patients who had stronger coughs compared with the hospitalized patients recruited in our previous study at this site (19). Our aerosol production results have been consistent across two different settings (United States and Uganda) and our reproducibility studies have shown reliable results (19). The Andersen cascade impactor was chosen for our sampling method because it has been the reference standard for viable bioaerosol measurement for decades, supported by a considerable body of literature. It provides validated particle size distribution data (29) and it has been used previously to collect mycobacterial aerosols (30, 31). Furthermore, the same aerosol collection method yielded positive cough aerosol cultures of Pseudomonas aeruginosa from 96% of patients with cystic fibrosis who were chronically infected (32).
An important roadblock to developing a reliable predictor of natural TB transmission has been the long reliance on studies of sputum, despite established evidence that TB is transmitted by aerosols. Although aerosol production was correlated with sputum AFB smear grade in this and our previous studies, most subjects with high-grade smears (i.e., AFB ≥3+) were aerosol negative, suggesting great variability in infectiousness independent of sputum AFB smear microscopy. This observation runs in opposition to the current paradigm of TB transmission, which considers all smear-positive cases as equal transmitters. Similarly, aerosol generation (and therefore rate of conversion in contacts) was not associated with semiquantitative sputum culture results or MGIT DTP, two markers that are commonly assumed to be indicative of bacillary burden and presumably infectiousness. Without the aerosol measurement, however, the identification of the most infectious patients with TB is not straightforward because all patients had similar characteristics at baseline. If cough aerosol collection can be modified to incorporate a real-time read out in a point-of-care test (33, 34), the identification of the most infectious patients could allow for more cost-effective use of resources for infection control in hospitals (e.g., isolation rooms) and for targeting preventive treatment in individuals at highest risk of TB infection by virtue of their exposure.
In addition to providing a more precise marker of source infectiousness, cough aerosols may help determine the individual risk of M. tuberculosis infection after exposure, which can be variable and is poorly understood (35). Previous efforts to explain this variability have focused on identifying environmental and host immunogenetic factors in close contacts (36). More recently, there has been great interest in identifying strain virulence factors (37–40). The results from this study suggest that differential ability of source cases to transmit M. tuberculosis may be an underappreciated source of variability. A related issue has been the lack of appreciation of the importance of the inhaled dose in humans. First demonstrated by Robert Koch himself (41), the immunopathology of TB in animal models is clearly dependent on the size of the infecting inoculum (41–43). In contrast, most of the epidemiologic evidence supporting the importance of inoculum size on TB outcomes in humans is based on inference, mostly from household contact studies. Interestingly, whereas the incorporation of IGRA substantially increased the discriminatory capacity of aerosols, we observed differential TST-IGRA responses between contacts exposed to aerosol-negative and aerosol-positive index TB cases. In particular, contacts of aerosol-negative patients with TB who underwent TST conversion had very low interferon-γ levels, suggesting the possibility their TST results may be explained by boosting. Further work is needed to determine if differences in inhaled inoculum may lead to a qualitatively different infection in humans (44). If our results are reproduced in future studies, cough aerosols may serve as an exposure biomarker to study natural M. tuberculosis transmission in humans.
The predictive value of cough aerosols over time is unknown. Like the sputum smear and culture, cough aerosols measured shortly after TB diagnosis is a surrogate of past exposure. As a consequence, contacts that were already infected at enrollment may have been infected by the TB index case at an earlier time point when infectiousness (and aerosol positivity) may have differed, or in the community. Also, it is possible that the high background rate of M. tuberculosis infection in contacts, although consistent with our previous finding at this same site (36), may have obscured weaker predictors of TST conversion in this study. The generalizability of our findings is unknown for several reasons. First, we restricted study participation to sputum AFB-positive patients at screening, which in this setting biased the sample of patients toward having more advanced disease. Second, we only included two index cases that were AFB-negative/scanty in this sample and are therefore unable to evaluate aerosol production in that group. Finally, because we only enrolled households with three or more contacts, our results may not apply to less crowded dwellings. We measured cough, an important predictor of TB transmission, in several ways but none of these provided an objective, continuous measurement of cough strength.
Our results have several implications. First, by identifying the most infectious patients with pulmonary TB, our findings belie the current paradigm that considers all sputum AFB-positive patients as homogeneously infectious, providing a new framework for more rational and cost-effective infection control decisions. Second, by reducing the number of TB contacts at highest risk of infection while increasing the yield of infected individuals, aerosols may improve the implementation of latent TB infection treatment programs by targeting contacts by virtue of their exposure. Last, and perhaps most importantly, M. tuberculosis aerosol measurement provides a quantitative surrogate of inhaled dose in exposed contacts that reduces exposure misclassification and thus, is likely to offer insights into natural TB transmission and immunopathogenesis, and facilitate studies of vaccines, drugs, and immune responses.
The authors acknowledge the invaluable contribution made by Mr. Haruna Butunzi (driver); the clinical and nursing staff at the Mulago Hospital TB Clinic; the rest of the staff at the TB laboratory of the Department of Microbiology, Makerere University College of Health Sciences; and Dr. Yukari Manabe at the Makerere University College of Health Sciences Infectious Disease Institute (supervised interferon-γ release assay testing). They also thank Jennifer Drucker Hosford, MPH, Dr. Jerrold Ellner, and Dr. Robert Horsburgh for critically reviewing the manuscript.
| 1. | World Health Organization. Global tuberculosis control: WHO report 2011. Geneva, Switzerland: World Health Organization; 2011. Contract No.: WHO/HTM/TB/2011.16. |
| 2. | Roy CJ, Milton DK. Airborne transmission of communicable infection–the elusive pathway. N Engl J Med 2004;350:1710–1712. |
| 3. | Escombe AR, Moore DA, Gilman RH, Pan W, Navincopa M, Ticona E, Martinez C, Caviedes L, Sheen P, Gonzales A, et al. The infectiousness of tuberculosis patients coinfected with HIV. PLoS Med 2008;5:e188. |
| 4. | Sultan L, Nyka W, Mills C, O'Grady F, Wells W, Riley RL. Tuberculosis disseminators: a study of the variability of aerial infectivity of tuberculous patients. Am Rev Respir Dis 1960;82:358–369. |
| 5. | Borgdorff M, Nagelkerke NJ, de Haas PE, van Soolingen D. Transmission of Mycobacterium tuberculosis depending on the age and sex of source cases. Am J Epidemiol 2001;154:934–943. |
| 6. | Brooks SM, Lassiter NL, Young EC. A pilot study concerning the infection risk of sputum positive tuberculosis patients on chemotherapy. Am Rev Respir Dis 1973;108:799–804. |
| 7. | Hamburg MA, Frieden TR. Tuberculosis transmission in the 1990s. N Engl J Med 1994;330:1750–1751. |
| 8. | Morrison J, Pai M, Hopewell PC. Tuberculosis and latent tuberculosis infection in close contacts of people with pulmonary tuberculosis in low-income and middle-income countries: a systematic review and meta-analysis. Lancet Infect Dis 2008;8:359–368. |
| 9. | van Geuns HA, Meijer J, Styblo K. Results of contact examination in Rotterdam, 1967–1969. Bull Int Union Tuberc 1975;50:107–121. |
| 10. | Verhagen LM, van den Hof S, van Deutekom H, Hermans PW, Kremer K, Borgdorff MW, Soolingen DV. Mycobacterial factors relevant for transmission of tuberculosis. J Infect Dis 2011;203:1249–1255. |
| 11. | Riley RL, Mills CC, O'Grady F, Sultan LU, Wittestadt F, Shivipuri DN. Infectiousness of air from a tuberculosis ward-ultraviolet irradiation of infected air: comparative infectiousness of different patients. Am Rev Respir Dis 1962;85:511–525. |
| 12. | Nardell EA. Environmental control of tuberculosis. Med Clin North Am 1993;77:1315–1334. |
| 13. | Horsburgh CR, Jr., Rubin EJ. Clinical practice: latent tuberculosis infection in the United States. N Engl J Med 2011;364:1441–1448. |
| 14. | Sepkowitz KA. How contagious is tuberculosis? Clin Infect Dis 1996;23:954–962. |
| 15. | Fennelly KP. Variability of airborne transmission of Mycobacterium tuberculosis: implications for control of tuberculosis in the HIV era. Clin Infect Dis 2007;44:1358–1360. |
| 16. | Behr MA, Warren SA, Salamon H, Hopewell PC, Ponce de Leon A, Daley CL, Small PM. Transmission of Mycobacterium tuberculosis from patients smear-negative for acid-fast bacilli. Lancet 1999;353:444–449. |
| 17. | Hernandez-Garduno E, Cook V, Kunimoto D, Elwood RK, Black WA, FitzGerald JM. Transmission of tuberculosis from smear negative patients: a molecular epidemiology study. Thorax 2004;59:286–290. |
| 18. | Tostmann A, Kik SV, Kalisvaart NA, Sebek MM, Verver S, Boeree MJ, Soolingen VD. Tuberculosis transmission by patients with smear-negative pulmonary tuberculosis in a large cohort in the Netherlands. Clin Infect Dis 2008;47:1135–1142. |
| 19. | Fennelly KP, Jones-Lopez EC, Ayakaka I, Kim S, Menyha H, Kirenga B, Muchwa C, Joloba M, Dryden-Peterson S, Reilly N, et al. Variability of infectious aerosols produced during coughing by patients with pulmonary tuberculosis. Am J Respir Crit Care Med 2012;186:450–457. |
| 20. | World Health Organization. Laboratory services in tuberculosis control. Part II: microscopy. Geneva, Switzerland: World Health Organization; 1998. Report No.: WHO/TB/98.258. |
| 21. | Raj AA, Birring SS. Clinical assessment of chronic cough severity. Pulm Pharmacol Ther 2007;20:334–337. |
| 22. | Fennelly KP, Martyny JW, Fulton KE, Orme IM, Cave DM, Heifets LB. Cough-generated aerosols of Mycobacterium tuberculosis: a new method to study infectiousness. Am J Respir Crit Care Med 2004;169:604–609. |
| 23. | Centers for Disease Control and Prevention. Guidelines for the investigation of contacts of persons with infectious tuberculosis: recommendations from the National Tuberculosis Controllers Association and CDC, and guidelines for using the QuantiFERON-TB Gold test for detecting Mycobacterium tuberculosis infection, United States. MMWR Recomm Rep 2005(RR-15);54:1–47. |
| 24. | Menzies D. Interpretation of repeated tuberculin tests: boosting, conversion, and reversion. Am J Respir Crit Care Med 1999;159:15–21. |
| 25. | Ministry of Health, Republic of Uganda. Manual of the national tuberculosis and leprosy programme, 2nd ed. Kampala, Uganda: Ministry of Health, Republic of Uganda; 2010. |
| 26. | Zeger SL, Liang KY, Albert PS. Models for longitudinal data: a generalized estimating equation approach. Biometrics 1988;44:1049–1060. |
| 27. | Galvani AP, May RM. Epidemiology: dimensions of superspreading. Nature 2005;438:293–295. |
| 28. | Lloyd-Smith JO, Schreiber SJ, Kopp PE, Getz WM. Superspreading and the effect of individual variation on disease emergence. Nature 2005;438:355–359. |
| 29. | Andersen AA. New sampler for the collection, sizing, and enumeration of viable airborne particles. J Bacteriol 1958;76:471–484. |
| 30. | Riley RL, Knight M, Middlebrook G. Ultraviolet susceptibility of BCG and virulent tubercle bacilli. Am Rev Respir Dis 1976;113:413–418. |
| 31. | Wendt SL, George KL, Parker BC, Gruft H, Falkinham JO3rd. Epidemiology of infection by nontuberculous Mycobacteria. III: isolation of potentially pathogenic mycobacteria from aerosols. Am Rev Respir Dis 1980;122:259–263. |
| 32. | Wainwright CE, France MW, O'Rourke P, Anuj S, Kidd TJ, Nissen MD, Sloots TP, Coulter C, Ristovski Z, Hargreaves M, et al. Cough-generated aerosols of Pseudomonas aeruginosa and other gram-negative bacteria from patients with cystic fibrosis. Thorax 2009;64:926–931. |
| 33. | Vadrot C, Bex V, Mouilleseaux A, Squinazi F, Darbord JC. Detection of Mycobacterium tuberculosis complex by PCR in hospital air samples. J Hosp Infect 2004;58:262–267. |
| 34. | Schafer MP, Martinez KF, Mathews ES. Rapid detection and determination of the aerodynamic size range of airborne mycobacteria associated with whirlpools. Appl Occup Environ Hyg 2003;18:41–50. |
| 35. | Whalen CC, Zalwango S, Chiunda A, Malone L, Eisenach K, Joloba M, Boom WH, Mugerwa R. Secondary attack rate of tuberculosis in urban households in Kampala, Uganda. PLoS One 2011;6:e16137. |
| 36. | Guwatudde D, Nakakeeto M, Jones-Lopez EC, Maganda A, Chiunda A, Mugerwa RD, Ellner JJ, Bukenya G, Whalen CC. Tuberculosis in household contacts of infectious cases in Kampala, Uganda. Am J Epidemiol 2003;158:887–898. |
| 37. | Alland D, Whittam TS, Murray MB, Cave MD, Hazbon MH, Dix K, Kokoris M, Duesterhoeft A, Eisen JA, Fraser CK, et al. Modeling bacterial evolution with comparative-genome-based marker systems: application to Mycobacterium tuberculosis evolution and pathogenesis. J Bacteriol 2003;185:3392–3399. |
| 38. | de Jong BC, Hill PC, Aiken A, Awine T, Antonio M, Adetifa IM, Jackson-Sillah DJ, Fox A, Deriemer K, Gagneux S, et al. Progression to active tuberculosis, but not transmission, varies by Mycobacterium tuberculosis lineage in The Gambia. J Infect Dis 2008;198:1037–1043. |
| 39. | Toungoussova OS, Sandven P, Mariandyshev AO, Nizovtseva NI, Bjune G, Caugant DA. Spread of drug-resistant Mycobacterium tuberculosis strains of the Beijing genotype in the Archangel Oblast, Russia. J Clin Microbiol 2002;40:1930–1937. |
| 40. | Zhang M, Gong J, Yang Z, Samten B, Cave MD, Barnes PF. Enhanced capacity of a widespread strain of Mycobacterium tuberculosis to grow in human macrophages. J Infect Dis 1999;179:1213–1217. |
| 41. | Koch R. The etiology of tuberculosis (Die Aetiologie der Tuberculose in Berliner klinische Wochenschrift). Essays of Robert Koch. Westport, CT: Greenwood Press, Inc.; 1987. |
| 42. | Glover RE. Infection of mice with Mycobact. tuberculosis (bovis) by the respiratory route. Br J Exp Pathol 1944;25:141–149. |
| 43. | Lurie M. Resistance to tuberculosis: experimental studies in native and acquired defensive mechanisms. Cambridge, MA: Harvard University Press; 1964. |
| 44. | Hill PC, Jackson-Sillah DJ, Fox A, Brookes RH, de Jong BC, Lugos MD, Adetifa IM, Donkor SA, Aiken AM, Howie SR, et al. Incidence of tuberculosis and the predictive value of ELISPOT and Mantoux tests in Gambian case contacts. PLoS One 2008;3:e1379. |
Supported by a University of Medicine and Dentistry of New Jersey Foundation award with matching funds from the Division of Infectious Diseases at New Jersey Medical School, and funds from the Section of Infectious Diseases at Boston Medical Center. M.P.F. was supported by the National Institute of Allergy and Infectious Diseases (K01AI083097). The funding sources had no role in the design, data collection, analysis, or interpretation of this study.
Author Contributions: Conception and design, E.C.J.-L., I.A., N.R., S.K., and K.P.F. Acquisition of data, E.C.J.-L., O.N., F.M., N.R., M.S., O.M., G.M., I.A., A.O., and M.J. Analysis and interpretation, E.C.J.-L., F.M., S.M., M.P.F., S.K., A.O., M.J., and K.P.F. All authors contributed to either drafting or revising this manuscript and gave final approval. None of the authors declared a conflict of interest. E.C.J.-L. had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
The results shown here were presented in part in an oral presentation (E.C.J.-L. invited speaker) at the 5th Annual New England TB Symposium, June 23, 2011, Broad Institute, Cambridge, Massachusetts.
Originally Published in Press as DOI: 10.1164/rccm.201208-1422OC on January 10, 2013
This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org
Author disclosures are available with the text of this article at www.atsjournals.org.
