Detection, survival and infectious potential of Mycobacterium tuberculosis in the environment: a review of the evidence and epidemiological implications

Abstract

Much remains unknown about Mycobacterium tuberculosis transmission. Seminal experimental studies from the 1950s demonstrated that airborne expulsion of droplet nuclei from an infectious tuberculosis (TB) patient is the primary route of transmission. However, these findings did not rule out other routes of M. tuberculosis transmission. We reviewed historical scientific evidence from the late 19th/early 20th century and contemporary studies investigating the presence, persistence and infectiousness of environmental M. tuberculosis. We found both experimental and epidemiological evidence supporting the presence and viability of M. tuberculosis in multiple natural and built environments for months to years, presumably following contamination by a human source. Furthermore, several studies confirm M. tuberculosis viability and virulence in the environment using guinea pig and mouse models. Most of this evidence was historical; however, several recent studies have reported consistent findings of M. tuberculosis detection and viability in the environment using modern methods. Whether M. tuberculosis in environments represents an infectious threat to humans requires further investigation; this may represent an untapped source of data with which to further understand M. tuberculosis transmission. We discuss potential opportunities for harnessing these data to generate new insights into TB transmission in congregate settings.

Introduction

Much of our current understanding of the transmissibility of tuberculosis derives from inference and accident rather than from intentional scientific study

Kent A. Sepkowitz, 1996 [1]

Every year, more than 10 million new cases of tuberculosis (TB) occur globally [2] and ongoing Mycobacterium tuberculosis transmission is the primary driver of incident disease. Despite significant advances in TB diagnostics, immunology and genomic epidemiology [3], much remains unknown about individual- and population-level transmission dynamics. Our current tools allow for the study of M. tuberculosis transmission only after a TB case is diagnosed and seminal studies of TB infectiousness, including the landmark studies of Wells and Riley in the mid-20th century [4–6], were conducted in modified hospital wards following TB diagnosis. “Community-based” exposure studies focus predominantly on household contacts, which account for <20% of infections in high TB burden settings [7–12]. Due to the prolonged infectious period of TB [13–15] and the potential for transmission from brief, casual exposures, less than one-third of cases can be epidemiologically and genetically linked [7–12]. Fundamental questions about where TB transmission occurs in communities, relationships between exposure and infection or disease risk, and heterogeneity in transmission at the population level remain poorly understood [1, 16, 17]. Much of this is due to our limited tools for studying transmission in community settings.

Prior to the mid-19th century, several competing theories existed regarding transmission of M. tuberculosis. For example, CALMETTE and GUÉRIN [18] proposed in 1905 that TB could be transmitted through contaminated food. In the 1950s, innovative and groundbreaking studies from Riley and Wells demonstrated that droplet aerosols from infectious TB patients resulted in substantial rates of tuberculin conversion in exposed guinea pigs [4–6]. These experiments were recently re-created in guinea pig air sampling facilities in Peru and South Africa, with consistent results [19–23]. These studies found that TB transmission to guinea pigs occurred from small droplet aerosols expulsed by patients and that the infectiousness of source cases was highly heterogeneous [6]. This work clarified the fundamentals of airborne transmission of pulmonary TB and is currently widely accepted as the primary mechanistic transmission route [6, 17]. These and other findings by Loudon investigating the generation of droplet nuclei through various airway activities, such as coughing and singing, were highly influential [24–27]. These studies paved the way for the study of TB transmission in contemporary studies, as well as the development and implementation of effective infection control strategies in healthcare facilities [17, 22].

While these critical studies shifted the focus of research to airborne transmission, their results did not rule out other routes of M. tuberculosis transmission. Earlier evidence for the detection of M. tuberculosis in, and potential transmission from, environmental settings may have been set aside with the discovery of the predominance of transmission through airborne droplet nuclei. The study of M. tuberculosis in the environment could potentially address limitations in our understanding of transmission, which is currently achieved almost exclusively through studying patient-derived samples.

We review historical scientific evidence from the late 19th/early 20th century as well as contemporary studies investigating the presence, persistence and infectious potential of M. tuberculosis in environmental settings. We evaluate the strengths and limitations of this body of evidence, identify unresolved questions and areas requiring further research, and discuss how studies of M. tuberculosis in environmental samples could facilitate new approaches towards the study of TB transmission.

Historical evidence

Following Robert Koch’s discovery of the M. tuberculosis bacillus in 1882, many scientists and researchers sought to determine how the bacterium was transmitted. Their studies focused on detecting M. tuberculosis in various environments, culturing it to demonstrate its viability and, in some cases, injecting it into animals to confirm that it caused TB disease. Much of this evidence was generated in the late 19th/early 20th centuries. Here, we review these historical studies concerning the detection, viability and infectiousness of M. tuberculosis in natural and built environments.

Certain caveats should be considered in the interpretation of historical studies. First, many of these studies were done before the diversity of mycobacteria was understood and it is possible that nontuberculous mycobacteria (or other members of the M. tuberculosis complex) were described in some of the studies. While many of the studies reported use of clinical M. tuberculosis isolates or sputum from TB patients, others reported direct detection in various environments such as soil or water, in which nontuberculous mycobacteria are commonly found. Studies lacking proper controls are susceptible to inaccurate inference about the role of M. tuberculosis in the environment. Given this uncertainty in the historical microbiology, we have attempted to emphasise studies in which controls were tested, clinical isolates from human cases were used or TB was induced in animal models for confirmation.

Natural environments

A number of studies investigated the survival of tubercle bacilli in natural environments, including soil and water (table 1). In 1888, CHANTEMESE and WIDAL [28] performed several experiments on river water attempting to quantify the number of days that the tubercle bacilli remained active. They first collected several tubes of water from the Seine River in Paris, France, subsequently sterilising some tubes. They then inoculated all tubes (sterilised and not) with patient cultures of tubercle bacilli. The tubes were kept at different temperatures (8–12°C and 15–20°C). After 50 days, the tubercle bacilli stored at both temperatures could still be cultured. A year later, STRAUS and DUBARRY [29] inoculated 10 tubes of 10 cm³ of distilled water and water from the Ourcq River, also in France. Culture from a TB patient was inoculated into tubes containing these different water samples. Samples maintained in river water were culturable for up to 30 days and those in distilled water were culturable for up to 115 days. Three guinea pigs were then inoculated with the river water into which M. tuberculosis had been inoculated. One guinea pig developed abscesses from which M. tuberculosis was culturable, while the other two guinea pigs remained healthy. In addition to these tests of environmental persistence and viability, several studies detected M. tuberculosis in environmental water and sewage samples collected near TB sanatoria, suggesting contamination from human sources [28].

TABLE 1.

Historical studies investigating Mycobacterium tuberculosis in the environment (pre-1960)

First author, year [ref.]	Medium	Method of M. tuberculosis detection
Outdoor, natural environments
Feltz, 1887 [30]	Soil	Guinea pig model
Chantemese, 1888 [28]	River water	Culture of samples; guinea pig model
Straus, 1889 [29]	River water and distilled water	Culture of samples; guinea pig model
Mitchell, 1900 [31]	Soil exposed to sunlight	Guinea pig model
Indoor, built environments
Cornet, 1889 [32]	Dust from tuberculosis medical ward and walls of rooms of TB patients; dust from streets and surgical wards [controls] with no TB patients	Guinea pig model
Mitchell, 1900 [31]	Soil	Guinea pig model
Twitchell, 1905 [34]	Handkerchief, woollen blanket, wood	Culture of samples; guinea pig model
Rogers, 1920 [33]	Dust from floors in open TB wards, from a morgue floor, and windows and shelves of a morgue; lastly, a morgue floor was swept every morning	Guinea pig model
Cumming, 1920 [35]	Utensils of a TB patient; hands of a TB patient	Guinea pig model

TB: tuberculosis.

Other studies demonstrated prolonged viability of M. tuberculosis in soil. In 1887, FELTZ [30] mixed soil and sputum rich in M. tuberculosis bacilli. He then exposed the soil to differing degrees of sunlight, periodically collected extracts of the soil and inoculated it into guinea pigs. Inoculated soil to which M. tuberculosis had been added 137 days prior produced TB lesions. A decade later, MITCHELL and COUCH [31] attempted to investigate the virulence of M. tuberculosis in soil after exposure to sunlight. They placed sputum from TB patients onto heat-sterilised soil, exposed the soil to sunlight for varying lengths of time and injected the soil into guinea pigs. When inoculated with soil exposed to sunlight for <35 h, TB manifested in guinea pigs; longer durations of sunlight exposure resulted in no infections in guinea pigs. M. tuberculosis was also culturable from soil exposed for <35 h. This study suggested that sunlight may have some influence on the presence of M. tuberculosis in the environment from some outdoor settings.

Built environments

In addition to natural environments, many studies investigated the viability of M. tuberculosis found in built environments and on fomites such as clothing and cooking utensils (table 1). In the late 1800s, several studies investigated dust collected from rooms of TB patients as a potential mode of transmission. In 1888, CORNET [32] collected dust from TB medical wards as well as from hospitals, asylums and apartments housing TB patients. After subcutaneous inoculation of the dust in 91 guinea pigs, 15 were later autopsied and found to have developed TB. This study also evaluated negative control groups with no TB exposure, using dust from surgical wards (none of eight inoculated guinea pigs developed TB) as well as from streets and public buildings (none of 41 inoculated guinea pigs developed TB). This study provided strong empiric evidence using guinea pig controls and several locations with a high TB risk that exposure to M. tuberculosis through dust may lead to disease in guinea pigs.

In 1920, ROGERS [33] also conducted a series of experiments using dust. In all experiments, dust was treated with sterile 2% sodium hydroxide solution, and then centrifuged and injected into guinea pigs. In the first three experiments, he injected guinea pigs with dust and found evidence of disease in four out of 11 guinea pigs injected with dust from floors of TB wards, four out of four guinea pigs injected with dust from a morgue floor, and two out of five guinea pigs injected with dust from the windows and shelves of a morgue. A final experiment involved placing 10 guinea pigs in a sterilised wire cage kept about 3 feet (1 m) above the morgue floor for an average of 5 h per day. The floor was swept every morning. After 18 days, all guinea pigs were autopsied and seven were subsequently diagnosed with disseminated TB. These experiments suggested that exhaled TB, which had settled on surfaces, could be re-aerosolised and generate infections. Importantly, however, ROGERS [33] did not have adequate controls.

TWITCHELL [34] placed dried sputum from a TB patient on a handkerchief, wood and a woollen blanket, and waited up to 70 days, subsequently subcutaneously inoculating the sputum into the groin of one guinea pig per object. After 4–6 weeks, each guinea pig was killed, autopsied and inspected for the presence of TB. TB lesions were detected in each guinea pig. TWITCHELL [34] also investigated the viability of M. tuberculosis on carpets. He placed dried sputum from a patient on a carpet for 39 days and then inoculated a guinea pig. The guinea pig subsequently developed disease. When the same experiment was conducted with the sputum exposed to sunlight for up to 7 h, the bacilli died in a short time span, suggesting that sunlight may affect M. tuberculosis on these objects.

In 1920, CUMMING [35] conducted a series of experiments with the goal of testing whether common household objects could represent fomites for TB. In a first experiment, three spoons from separate smear-positive TB patients were washed after each meal with a cloth in hot water. This wash water was subsequently centrifuged and injected subcutaneously into guinea pigs. Of 31 guinea pigs injected and subsequently autopsied, 11 died from TB. In a separate experiment, the hands of TB patients were placed in warm water for several minutes after which epithelia from the hands were scraped with a scalpel and injected into seven guinea pigs. Three of these guinea pigs subsequently died due to TB.

Flies and M. tuberculosis

Several early studies examined the ability of flies to carry and spread M. tuberculosis [36–40]. In 1904, LORD [39] reported on several studies on TB in flies performed in the late 19th/early 20th century. In one experiment, he confined 30 flies in a glass jar and fed them sputum from TB patients over 4 days (figure 1). Two sets of controls were performed: 1) six flies were put into a jar and fed nontuberculous sputum, and 2) six flies were confined to a jar and fed only water, sugar and meat. Among the flies fed sputum from TB patients, M. tuberculosis was recovered in the excrement from the flies as well as the intestines of all 30 flies. No M. tuberculosis was found in excrement or intestines from control flies. The investigators then let the excrement from the flies who were fed TB patient sputum sit for differing periods of time (1, 8, 15, 28 and 55 days). Excrement from each duration of time was fed to guinea pigs. The guinea pigs who were fed excrement that sat for 1, 8 and 15 days all developed TB, while the two guinea pigs fed faeces older than 15 days remained healthy. Just a few years later, BUCHANAN [37] exposed flies to a Petri dish with M. tuberculosis and then transferred them to a sterile Petri dish with growth media, which resulted in growth in the second dish; he thereby demonstrated that flies could spread M. tuberculosis to other surfaces.

FIGURE 1.

Flies and tuberculosis (TB): 20th century experimentation and fly extermination campaigns. Experiments in August 1904 on Mycobacterium tuberculosis and flies by Frederick Lord [39]: flies trapped in a jar for feeding of sputum from a TB patient in a hospital ward (top left). Advertisement for eradication of flies for paediatric health in the Toronto Daily Star in 1912 (right). “Fly Swatting” contest from The Montreal Star in 1912: buckets of flies brought in to collect prizes (bottom left).

Other researchers used natural experiments to further study the viability of M. tuberculosis in flies. In 1887, HOFFMAN [40] found flies in the house of a recently deceased TB patient. He collected the flies, dissected them, and identified tubercle bacilli in their intestines and faeces by microscopy. In 1904, HAYWARD [38] caught several flies feeding on TB sputum in his laboratory and subsequently put them into clean cages. After several days he found TB bacilli in the faeces on 10 out of 16 cover-slips. Examination of unexposed, control flies resulted in no tubercle bacilli.

In the early 20th century, several cities (including Toronto and Montreal in Canada, London in the UK, and Detroit in the USA) launched large-scale campaigns to eradicate flies for the sake of public health [41–43]. During this time period, TB was a leading cause of paediatric mortality in the USA, Canada and much of the world, and flies were widely believed to be a vector of infectious diseases, including TB. “Fly Swatting” contests for children offered prizes for those that captured, killed and brought in the most flies (figure 1). These campaigns were promoted by physicians and the media (figure 1) who believed improved sanitation would have positive ramifications for paediatric health, including TB [41].

Contemporary evidence

Since the demonstration of transmission by airborne droplet nuclei, few studies have investigated M. tuberculosis in the environment (table 2) [44–47]. GHODBANE et al. [44] inoculated soil with M. tuberculosis, Mycobacterium bovis and Mycobacterium canetti, and assessed the survival of the distinct mycobacteria in the soil for 12 months. The soil was cultured monthly and all three types of mycobacteria were still present in the soil after 12 months. The soil contaminated with M. tuberculosis was then intermixed with the food for five healthy mice for 60 days. All five of these mice developed granulomas, whereas all of the control mice remained healthy. Similar results were found by KOZLOV and ROTOV [45], who seeded three strains of M. tuberculosis in a natural turf–podzol sandy soil and found that they were culturable for up to 3 months. These studies suggest that M. tuberculosis may retain viability and infectiousness after prolonged periods in soil. Natural sunlight may affect the persistence of M. tuberculosis in soil; however, this was not investigated. In addition, whether humans or other animals exposed to soil contaminated with M. tuberculosis could also become infected is unclear.

TABLE 2.

Contemporary evidence for environmental contamination of Mycobacterium tuberculosis (post-1960)

First author, year [ref.]	Methodology and medium	Description of results
Kozlov, 1977 [45]	Three strains of M. tuberculosis cultured for several months in a natural turf-podzol sandy soil	Samples culturable for up to 3 months after seeding
Ghodbane, 2014 [44]	Inoculated soil with M. tuberculosis, M. bovis and M. canetti; survival of the distinct mycobacteria in the soil for 12 months	Soil samples culturable; all types of mycobacteria found in soil after 12 months; mice inoculated with contaminated soil all developed granulomas; control mice did not grow M. tuberculosis
Tshilombo, 2015 [46]	Prospective in vitro study; mixed 20 mL of 106 CFU·mL−1 M. smegmatis with 125 mg of sterile dust; air sampling was conducted pre- and post-re-aerosolisation, and the number of CFUs was measured based on plate count	M. smegmatis survived in dust for 19 days and could be successfully re-aerosolised, remaining viable
Velayati, 2015 [47]	1500 random samples of soil and water	M. tuberculosis isolated from 1% of soil samples and 10% of water samples; persisted for 9 months

M. bovis: Mycobacterium bovis; M. canetti: Mycobacterium canetti; M. smegmatis: Mycobacterium smegmatis.

A recent study investigated re-aerosolisation of dust containing Mycobacterium smegmatis, a common surrogate marker for M. tuberculosis. TSHILOMBO et al. [46] conducted a prospective in vitro study whereby they mixed 20 mL of 10⁶ CFU·mL⁻¹ M. smegmatis with 125 mg of sterile dust. Air sampling was conducted pre- and post-re-aerosolisation and the number of CFUs was measured. The authors found that M. smegmatis survived in the dust for 19 days and could be successfully re-aerosolised, remaining viable. Although a limitation is the use of M. smegmatis, this study suggests the possibility that re-aerosolisation of M. tuberculosis in dust and other environments may be possible. Further research is needed to validate and extend the findings of this study [46]. Studies in real-world settings are also needed to assess the generalisability of these findings outside the laboratory. VELAYATI et al. [47] collected 1500 randomly selected soil and water samples in three counties in Tehran, Iran. M. tuberculosis was isolated by Löwenstein–Jensen culture media from 1% of soil samples and 10% of water samples, with confirmation by phenotypic and molecular tests. Soil and water samples were additionally stored and retested for culturability over time, which demonstrated that M. tuberculosis could be recovered from stored soil and water samples for 9 months after sample collection. MIRU-VNTR (mycobacterial interspersed repetitive unit-variable number tandem repeat) typing of environmental and clinical isolates revealed partial overlap in M. tuberculosis families present in the area.

Implications and future directions

Potential for TB transmission through environmental samples

It is not surprising that M. tuberculosis, an organism that originally evolved from a soil saprophyte and, by virtue of its capsule and mycolic acid and lipid-rich cell wall, is comparatively resilient against environmental stresses, would be capable of surviving for prolonged periods outside of humans [48, 49]. Historical and contemporary evidence for M. tuberculosis viability and infectiousness in environmental samples does not establish that environmental reservoirs are important routes of transmission. The evidence does, however, raise the possibility that transmission could occur through aerosolisation of environmental bacilli. Currently, little attention is given to environmental surfaces or fomites, including clothing and other objects, in the prevention of transmission. If environmental exposures do confer substantial risk, we will need to develop methods to prevent these transmission routes as global control efforts turn towards elimination. Areas with densely populated housing and hospital settings that have little sunlight may be particularly susceptible environments for the spread of M. tuberculosis.

To investigate this potential route of transmission, there is a need to test whether, and how readily, M. tuberculosis can be re-aerosolised from surfaces in a form that can generate infections in animals. There is evidence that re-aerosolisation causing human infection can occur through nosocomial transmission during surgical procedures from incision and irrigation of tuberculous abscesses [50, 51], and through autopsies [52].

Current animal models of TB typically involve bronchoscopic installation or aerosol exposure via nebulisation of freshly cultured M. tuberculosis. These models could be used to investigate the risk of infection from aerosolised M. tuberculosis on environmental surfaces (e.g. floors), as a function of the duration over which the bacteria are present prior to the exposure.

Methods for environmental detection of M. tuberculosis

There is a need for standardised, highly accurate methods for environmental detection of M. tuberculosis. M. tuberculosis can be cultured from soil and other materials, but sensitivity may be limited due to bacterial overgrowth and the presence of “differentially culturable” (or “viable but nonculturable”) organisms [53, 54]. A better understanding of M. tuberculosis viability in various environmental matrices will require methods for optimal promotion of their growth following recovery from the environment.

Molecular detection of M. tuberculosis has been demonstrated in filtered air samples [55–57], but to the best of our knowledge there are no studies investigating its detection on environmental surfaces. This is despite an increasingly robust literature on detection of various pathogens in natural and built environments [58–60]. While conventional molecular methods (e.g. PCR) do not distinguish viable from nonviable organisms, a number of molecular methods have been developed to do so, including detection of mRNA or selective detection of intracellular DNA [61–63].

Viability stains have recently been developed for M. tuberculosis, which could facilitate the investigation of its viability in natural and built environments [64]. Validating these methods for M. tuberculosis detection could enable higher throughput studies on M. tuberculosis in the environment.

In addition to the development of rigorous laboratory protocols for the detection and growth of M. tuberculosis from environmental sampling, there is a need for a better understanding of how to perform sampling in various environments. This includes questions pertaining to the frequency, location, sample collection methods and site selection, as well as understanding how disruptions (cleaning, airflow, water, etc.) might affect M. tuberculosis persistence in those environments.

Potential for enhancing our understanding of TB transmission through environmental detection of M. tuberculosis

A deeper understanding of M. tuberculosis transmission could improve TB prevention and control efforts [16, 17]. However, there are several fundamental challenges in studying TB transmission at present. First, individuals with TB are believed to be infectious for many months to years prior to their diagnosis [13–15]. During this time period, they may have numerous indoor contacts during which transmission could occur [11, 65, 66]. Because TB can be spread through airborne droplet nuclei between strangers sharing the same space and can even transmit to individuals without being in a room contemporaneously, contact-tracing investigations may have significant social network gaps. Efforts to link transmission based on contact investigations or social network analysis have found that only ~10–30% of genetically linked cases in endemic settings can be epidemiologically linked [12, 14, 16, 67]. An alternative goal has been to identify high-risk environments for transmission, which has implicated public transport, bars, churches, schools and workplaces based on where individuals spend time [11, 68]. However, given that individuals with active TB may have distinct social mixing patterns, or may alter these due to illness, this also represents an imperfect approach. Finally, most studies evaluating infectiousness of TB patients involve characterising their clinical features and examining their sputum (or, more recently, their cough or respiratory aerosol content) at the time of their diagnosis [69–72]. However, this approach ignores most of the TB disease spectrum prior to diagnosis and may not correlate well with the total TB exposure received by their household members or other contacts.

Environmental samples of M. tuberculosis may hold valuable epidemiological information, including where transmission occurs and where interventions should be targeted. In contrast to sampling in one location at the time of diagnosis, environmental sampling may provide a longer time and spatial window for TB. Given substantial interindividual variability in infectiousness [5, 6, 20], there may be substantial spatial heterogeneity in M. tuberculosis in the environment, influencing the chances of detection. If sensitive diagnostics become available for environmental sampling, they could theoretically be used to monitor longitudinal risk in community settings such as healthcare facilities, prisons and schools. If environmental M. tuberculosis can be genotyped, it could be possible to link environmental samples with clinical samples. Molecular epidemiology studies could then potentially reconstruct not only transmission chains, but spatially explicit transmission chains: not only “who infected who”, but where they were infected.

By utilising genotyping methods on environmental samples, it would be possible to determine when certain strains were detected in the community compared with when they were detected clinically, to understand the duration of undetected TB. Environmental sampling could also be used to quantify the sum of M. tuberculosis exposures over a time period. For example, in household contact investigations, a high degree of indoor contamination in the home may be a better indicator of exposure to household members than smear or cough aerosol status, collected at the time of diagnosis. Finally, environmental detection could direct case-finding efforts: mass screening could be targeted to sites where environmental samples are positive.

Critical knowledge gaps in TB epidemiology and the potential use of environmental sampling for M. tuberculosis are listed in table 3.

TABLE 3.

Critical knowledge gaps in tuberculosis (TB) epidemiology and the potential use of environmental sampling for Mycobacterium tuberculosis

Epidemiological areas related to TB	Potential use of environmental sampling and ideas for further investigation
Where does TB transmission occur?	Characterising M. tuberculosis abundance in various congregate settings, including public transit, schools, churches, restaurants, bars and workplaces
How much interindividual heterogeneity is there in community transmission?	Compare abundance of M. tuberculosis genotypes across settings
How does dose of exposure influence risk of TB progression?	Compare abundance of M. tuberculosis in household with risk of progression among household contacts
How long are individuals infectious prior to detection?	Whole genome sequencing of environmental samples and comparing dates of detection in the environment with date of diagnosis
How does an individual’s infectiousness change over time?	Assess rates of environmental contamination with specific isolates, identified by whole genome sequencing, over time
Was an intervention successful at reducing transmission?	Assess changes in M. tuberculosis abundance in congregate settings (e.g. hospitals, prisons, mines, schools] over time
Was an individual infectious when a potential exposure occurred?	Conduct environmental sampling at sites of potential exposure (e.g. hospital rooms]

Conclusions

Research on TB transmission has focused almost entirely on studying clinical isolates from sputum or detection of airborne M. tuberculosis through biological sensors (e.g. guinea pig models) or, more recently, air samplers in controlled environments. However, there is a considerable historical evidence base suggesting the presence of M. tuberculosis in both natural and indoor environments. Animal models suggest that M. tuberculosis may remain infectious in some of these environments, raising the possibility that re-aerosolisation could represent an unappreciated route of transmission. Moreover, even if environmental sources do not represent important sources of transmission, sampling for M. tuberculosis in indoor environments could nevertheless provide temporal and spatial dimensions for the study of TB transmission in ways not achievable through conventional patient-focused investigations. A re-assessment of the abundance and viability of M. tuberculosis in the indoor environment is warranted and, if confirmed, may yield novel insights into the transmission and community distribution of this important pathogen.