Contact details

Dr Adrian R Escombe

Honorary Research Fellow
Division of Investigative Science

Email:

Dr Rod Escombe

RESEARCH OVERVIEW

Tuberculosis kills over 2 million people every year, and the majority of these deaths occur in the developing world. TB is an airborne disease, transmitted by tiny droplets produced when an infected person coughs. These droplet nuclei, between 1-5 µm in diameter, are so small they remain suspended in the air for long periods, waiting to be breathed in to the lungs of a new susceptible host. Airborne tuberculosis transmission occurs particularly in overcrowded settings where TB patients may be found: hospitals, clinics, homeless shelters, and prisons are particularly important foci of TB transmission. The TB epidemic is exacerbated by HIV co-infection and the emergence of drug-resistant TB strains. Whilst just one in ten HIV-negative patients go on to develop active TB after inhaling an infectious dose of airborne TB, this rises to 40% in persons co-infected with HIV. Once a person develops active TB, it can be passed on via the infectious droplet nuclei produced by coughing and talking, or by cough inducing medical procedures such as intubation, bronchoscopy or sputum induction. Effective treatment for TB with antibiotics quickly reduces the infectiousness of TB patients. However the emergence of drug-resistant TB means that many TB patients are treated empirically with first line drugs, which are largely ineffective. Many of these patients attend health care facilities repeatedly, and all the time may be coughing up infectious droplets containing drug-resistant TB bacteria.

TUBERCULOSIS INFECTION CONTROL

Preventing TB transmission in the first place through TBinfection control measures is a public health priority. TB transmission and occupational TB are common in resource-limited settings, especially where TB and HIV are prevalent. In health care facilities it is not just nurses, doctors, and other staff that are at risk of catching TB, but also patients' relatives and other visitors. Furthermore, patients with drug-sensitive TB may catch drug-resistant TB. The expansion of HIV care programmes may inadvertently increase TB transmission by congregating highly susceptible individuals with those likely to have TB disease in overcrowded facilities such as anti-retroviral clinics.

Guidelines for preventing TB transmission in health care settings advocate three strategies: administrative control measures; environmental control measures; and personal protective measures. Administrative control measures aim to ensure the prompt diagnosis, isolation and treatment of TB patients. Environmental control measures aim to reduce the concentration of airborne infectious droplets by diluting them with fresh air ventilation, or by inactivating them with an air cleaning strategy such as ultraviolet (UV) light or air filtration. Personal protective measures involve the wearing of personal respirators, sometimes called masks.

Our research in Peru firstly investigated the infectiousness of HIV-positive TB patients, and then evaluated TB infection control measures, specifically natural ventilation by opening windows, and the use of upper room ultraviolet lights in hospital isolation rooms. Subsequent work has focussed on implementing research into practice, using natural ventilation in Peruvian hospitals and prisons.

INFECTIOUSNESS OF TB PATIENTS CO-INFECTED WITH HIV

Classic studies performed in Baltimore in the 1950s housed guinea pigs on the roof of a TB ward, breathing ward air. These elegant studies demonstrated the airborne nature of TB transmission, the variability of TB patient infectiousness, and the sterilising effect of UV lights placed inside air ducts. We recreated these studies at Hospital Nacional Dos de Mayo (below), a busy government hospital in Lima, Peru. A schematic of the guinea pig air sampling facility is shown below. All air from a mechanically ventilated, negative-pressure TB-HIV ward was passed over an average of 92 guinea pigs on the roof.  Guinea pigs underwent tuberculin skin testing, autopsy, and organ culture for M. tuberculosis. Sputum samples were collected from patients. DNA fingerprinting was performed on all M. tuberculosis strains isolated from both patients and guinea pigs. Over 505 days, 97 HIV-positive pulmonary TB patients were treated in the ward, and TB disease was detected in 144 of 292 exposed guinea pigs.  Marked variability in patient infectiousness was demonstrated, with TB transmission occurring from only 8.5% of patients. 90% of TB transmission occurred from inadequately treated multi-drug resistant (MDR) TB patients. Three highly infectious MDR-TB patients produced 226, 52 and 40 airborne infectious units (quanta) per hour. These studies highlight the importance of rapid TB drug-susceptibility testing to allow prompt initiation of effective treatment in drug-resistant cases. Furthermore, these studies highlight the importance of environmental control measures to reduce ongoing TB transmission in congregate settings and the need for TB infection control to be prioritized in order to prevent health-care facilities from disseminating the drug-resistant TB that they are attempting to treat.

This research has been published in Clinical Infectious Diseases CID 2007 and PLoS Medicine.PLoS Med 2008

NATURAL VENTILATION TO PREVENT TB TRANSMISSION

Fresh air ventilation is the mainstay of environmental control measures to reduce the airborne concentration of infectious droplet nuclei. Ventilation may be delivered by mechanical systems, with air supply and/or extractor fans, or more simply by just opening the windows. Mechanical ventilation can provide fresh air exchange measured as air changes per hour (ACH), and also negative pressure, which aims to keep contaminated air inside an isolation room. The problem with mechanical ventilation is its cost: not just for installation, but for ongoing careful maintenance such as changing filters and cleaning motors, and in addition electricity running costs. Because of this, high air exchange mechanical ventilation is usually restricted to high risk areas such as respiratory isolation rooms. However, TB infection control needs to be prioritised in other areas of health care facilities: emergency departments, out-patient departments, and waiting rooms. Natural ventilation by simply opening windows and doors offers an attractive alternative to mechanical ventilation because it is almost cost free, and requires little maintenance.
Little was known about the rates of fresh air exchange achievable with natural ventilation, so we used a carbon dioxide tracer gas technique to measure ventilation in 70 naturally ventilated clinical rooms where infectious patients are likely to be encountered, including TB wards, respiratory wards, HIV wards, out-patient clinics and emergency departments. Architectural and environmental variables were measured to elucidate the determinants of natural ventilation.
Opening windows and doors provided median ventilation of 28 ACH, more than double that of mechanically ventilated negative-pressure rooms ventilated at the 12 ACH recommended for high-risk areas, and eighteen times that with windows/doors closed (p<0.001). Facilities built >50 years ago, characterised by large windows and high ceilings, had greater ventilation than modern naturally ventilated rooms (40 vs. 17 air-changes/hour; p<0.001).
Even within the lowest quartile of wind speeds, natural ventilation exceeded mechanical (p<0.001; see graph above showing ACH for different quartiles of wind speed, for closed, partially, or fully opened windows and doors). Examples of old fashioned wards built along the design principles of Victorian TB sanatoria are shown: the large windows and high ceilings promoted high rates of natural ventilation, and reduced the risk of airborne TB transmission.

This research was published in PLoS Medicine in 2007. PLoS Med 2007

UPPER ROOM ULTRAVIOLET LIGHT TO PREVENT TB TRANSMISSION

M. tuberculosis is highly susceptible to short wavelength UV light known as UV-C. Carefully designed UV-C lamps may be positioned in the upper part of hospital rooms, to provide high UV intensity in the upper part of the room, without harming occupants in the lower part of the room. Mixing of room air, which may be augmented by simple fans, takes air contaminated by coughing TB patients through the upper part of the room where it is sterilised. In this way, nosocomial TB transmission may be reduced. Upper-room UV light is already recommended as an adjunct for reducing TB transmission in health care facilities. However, its use is not widespread due mainly to the absence of efficacy studies in clinical settings.

We therefore adapted our guinea pig air sampling model to measure the efficacy of upper room UV light for preventing TB transmission in the TB-HIV wards at Hospital Nacional Dos de Mayo in Lima. A new animal facility was built on the roof, this time with 3 guinea pig exposure chambers (shown above). In the control chamber, animals breathed untreated ward air. In the second chamber, animals breathed similar ward air, but only on alternate days, days when upper room UV lights were switched on in the ward. A third group of guinea pigs breathed untreated ward air, but negative air ionizers, which precipitate particles out of the air by putting a negative charge on them, were placed above their cages.
TB infection in guinea pigs was defined by monthly tuberculin skin-tests. All guinea pigs underwent autopsy to test for TB disease, defined by characteristic autopsy changes or by the culture of M. tuberculosis from organs. 

35% (106/304) of guinea pigs in the control group developed TB infection and this was reduced to 14% (43/303) by ionizers, and to 9.4% (29/307) by ultraviolet lights (both p<0.0001 compared with the control group – see graph). Time-to-event analysis demonstrated that TB infection was prevented by ionizers (log-rank 27; p<0.0001) and by ultraviolet lights (log-rank 46; p<0.0001).
This study has shown that upper-room UV has great potential for TB infection control. This is particularly so in settings where climate does not permit the use natural ventilation, or for use in cooler winter months, or at night when windows are likely to be closed. There is an urgent need for the design and mass production of simple UV fixtures for low-resource settings to help combat epidemic TB transmission. Upper room UV light is well suited to large congregate settings such as waiting rooms or emergency departments, where TB is often un-diagnosed, overcrowding is common, and respiratory isolation facilities are frequently unavailable.

This research was published in PLoS Medicine in March 2009.PLoS Med 2009

RESEARCH INTO PRACTICE

IMPLEMENTING NATURAL VENTILATION IN HOSPITALS IN PERU

Waiting rooms in health care facilities are often overcrowded, especially in low resource settings. The same is true for out-patient clinics. These areas are important for TB infection control because undiagnosed and untreated patients are often found in these areas, and these untreated patients are usually the most infectious. Simple architectural modifications were made to 4 waiting rooms and 2 out-patient consulting rooms in 2 hospitals in Lima, Peru. Room ventilation was measured before and after the intervention, and an airborne infection model was used to estimate the effect on TB transmission risk to health care workers in these rooms. Modifications included additional windows for cross-ventilation (see diagram below); removal of glass from un-openable windows; sealed skylight raised on 1m supports to permit air entry (see top photograph); re-building of waiting-room in the open air (see photographs below). Following the infrastructure modifications, room ventilation in the four waiting rooms increased from mean 5.5 to 15; 11 to 16; 10 to 17; and 9 to 66 ACH.

In the two consulting rooms, ventilation increased from mean 3.6 to 17; and 2.7 to 12 ACH. Median TB transmission risk calculated for all rooms was reduced from 79% to 30%. The modifications were cost free in two rooms, cost less than US$100 in two other rooms, and cost US$1000 and US$7000 each in the remaining two rooms.  These simple interventions were therefore likely to represent a highly cost effective way of preventing the transmission of TB and drug-resistant TB in crowded health care facilities.  

Old waiting room  (all specialties mixed together)	New waiting room (just for respiratory patients)

This work was presented at the International Union Against TB and Lung Disease world conference in Paris in 2008 , and the abstract here was published in the IJTLD Abstract IJTLD 2008 .

IMPLEMENTING NATURAL VENTILATION IN PERUVIAN PRISONS

Prisons commonly have high TB prevalence rates, mainly due to overcrowding.  Reducing TB and HIV in prisons was one of the objectives in a Global Fund for TB AIDS and Malaria funded programme in Peru. Whilst most TB transmission in prison probably occurs in the cells and other residential areas, crowded prison health care facilities represent another focus of transmission. We evaluated room ventilation in the health care facilities of 8 Peruvian prisons before and after the renovation or new construction of health care infrastructure guided by the principles of optimising natural ventilation. We used an airborne infection model to predict the likely reduction in TB transmission. Median baseline ventilation was 15 ACH, and this increased to 28 ACH in the newly constructed or re-modelled facilities. Theoretical TB transmission risk was almost halved in these health care facilities.
A new DOTS TB treatment dispensary area with a raised open roof to facilitate natural ventilation is shown in the top photograph. A refurbished cell for prisoners with MDR-TB is shown in the photograph above: note the new skylight and additional windows to promote cross ventilation and high fresh air exchange. A corridor that is usually crowded with waiting patients is shown on the right: new skylights were installed to promote natural ventilation.
 
A slide presentation of this work was made at the International Union Against TB and Lung Disease world conference in Cape Town in 2007, and the abstract here was published in the IJTLD. Abstract IJTLD 2007