© 2004 American Thoracic Society
Catching Droplet NucleiToward a Better Understanding of Tuberculosis TransmissionHarvard Medical School Cambridge, Massachusetts The droplet nuclei (dried residua of larger respiratory droplets) that transmit tuberculosis infection remain almost as mysterious as the "miasmata" that scholars once believed transmitted diseases through the air. We know droplet nuclei are there. Experimentally, if we artificially generate clouds of them we can capture and grow some on culture media. But we cannot catch and culture those naturally generated in the room by a person with pulmonary tuberculosis, no matter how infectious that person is. Many have tried. The reasons are that Mycobacterium tuberculosis organisms exist in extremely low concentrations in room air as compared with the sea of ordinary ambient bacteria and fungi around us, and they grow much slower. Culture plates with selective media intended to grow tubercle bacilli will invariably be overgrown with fungi and environmental bacteria, including environmental mycobacteria. Until now, there were only two ways to quantify the infectiousness of tuberculosis cases. The most common way has been through epidemiologic investigations of contacts, recently aided by molecular fingerprinting, but still potentially confounded (1). Have all contacts been identified? Do identical strains necessarily mean recent transmission? If a tuberculosis strain was successfully transmitted, was it because of exceptional source strength, favorable environmental conditions, increased host vulnerability, strain characteristics, or a combination of these factors? The other more direct approach was pioneered by Richard Riley more than 40 years ago, based on the experimental design of his mentor, William Firth Wells (2). Hundreds of guinea pigs breathed the air from an experimental tuberculosis ward in Baltimore. Those who became infected, as judged by skin test conversion, were autopsied and found to have a solitary peripheral lung focus of infection, the result of inhaling a single infectious droplet nucleus. By culture of the lung, and timing of the infections, the number of guinea pigs infected by individual patients on the ward was established. Not long before his death, Riley recounted some previously unknown details of those remarkable studies in these pages (3). In this issue (pp. 604–609), Fennelly and colleagues present a novel approach to directly assessing the potential infectiousness of patients with pulmonary tuberculosis (4). By having patients with pulmonary tuberculosis cough into a small chamber, Fennelly and colleagues concentrate the infectious droplet nuclei generated while excluding some of the competing ambient organisms. Using mechanical air sampling devices in the small chamber, airborne infectious droplet nuclei in the respirable size range were recovered quantitatively in a quarter of the patients tested. For the first time, the size distribution of infectious droplet nuclei generated by a cough could be estimated, depending on the "stage" on which organisms landed and grew in the air sampler. Large differences between patients in potential infectiousness were noted, as was a relatively brisk response to therapy. Reassuringly, great variability between patients and a rapid loss of infectiousness with effective treatment was also noted by Riley using guinea pig air sampling (5). Why is this work important and what is its potential? Tuberculosis is among humankind's most successful pathogens, almost every case the result of airborne transmission, and yet we know relatively little of its aerobiology. Apart from established transmission factors like cough frequency, lung cavitation, and sputum smear status, we know little about why some patients are "disseminators" while others infect few if any contacts. How important are the physical properties of the lung lining fluid that must be aerosolized to form droplet nuclei? Could these be manipulated to reduce contagion? Do tuberculosis strains vary in their ability to resist damage caused by aerosolization, by dehydration during airborne transport, and by rehydration upon inhalation by a new host? How long do tubercle bacilli survive in air under various conditions? Does ambient high humidity in tropical regions favor transmission? In addition to germicidal irradiation, are there other ways to make high-risk environments less favorable for transmission? The answers to these and other basic transmission questions will likely lead to new ways to reduce transmission, but they will not likely come from epidemiological investigations alone. We need to return to basic studies like those begun by Wells, Ratcliff, Riley, and Loudon decades ago, but not continued in recent years. The research published here by Fennelly and colleagues (4) represents a new step in that direction, focusing primarily on factors that determine the infectiousness of the source. The results, however, are admittedly preliminary, and their relevance to person to person transmission remains to be established. Despite their limitations, epidemiological investigators have as their endpoints human infection or disease, the endpoints of greatest public health importance. The closest surrogate for human infection is infection in animal models, such as mice and guinea pigs. Because they are exquisitely susceptible to human tuberculosis and because of their large size, guinea pigs were chosen by Wells and Riley as optimal living air samplers for their experimental tuberculosis ward (6). Early investigators established "parity" between the numbers of colonies on culture plates and the number of foci in the lungs of experimental animals sampling the same artificial aerosols of tubercle bacilli (7). In Fennelly's studies, however, sampling occurs at extremely close proximity to the infectious source. There is little time for droplet nuclei to dry and it is unclear whether some of these cultured organisms would remain infectious in a room very long. Fortunately, direct comparisons with the guinea pig model should be possible because experimental wards, similar to that of Riley, are being used in both Peru and South Africa. Finally, one would hope that molecular methods would advance our understanding of transmission beyond the contribution of molecular fingerprinting. Detection of tuberculosis in air from a human source using nucleic acid amplification has been reported, but must be interpreted with great caution, as the method used was not quantitative and could not differentiate living from dead organisms (8, 9). As an endpoint, nucleic acid detection is even further removed from infection than culture. Still, these methods can only improve, stimulated by bioterrorism research and the need to detect airborne organisms quickly. Multidrug-resistant tuberculosis is classified as a potential bioterrorism agent, and a welcome by-product of biodefense research may be greater understanding of tuberculosis aerobiology. FOOTNOTES Conflict of Interest Statement: E.A.N. has no declared conflict of interest. REFERENCES
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
