A combination of administrative, environmental, and personal infection control measures could successfully reduce by half the incidence of hospital-acquired extensively drug-resistant tuberculosis (XDR-TB) in rural South Africa, according to the results of a mathematical modelling study published in the October 27th edition of the Lancet.
The spectre of multidrug-resistant TB hangs over countries with an escalating HIV problem. The public health problem of multidrug-resistant TB, characterised by resistance to isoniazid and rifampicin, has been confounded by the emergence of multidrug-resistant TB isolates which are resistant to second line anti-TB drugs. These XDR-TB isolates are resistant to isoniazid, rifampicin, any flouroquinolone, and second-line drugs.
XDR-TB has been reported in 37 countries as of May 2007. The largest cluster of XDR-TB has been reported in South Africa where it has been diagnosed in every province. KwaZulu Natal province alone has reported over 200 cases.
An important insight about the epidemiology of XDR-TB in Kwa-Zulu Natal was provided by the first 53 cases of XDR-TB in the rural town of Tugela Ferry. All infected patients were HIV-positive, had a median survival time of 16 days from sputum collection, 98% mortality, and had never been exposed to second line drugs. Significantly, 67% of the patients had been recently hospitalised and two were healthcare workers. This was the first indication that XDR-TB in Tugela Ferry was hospital-acquired.
Scientific data are urgently needed to guide policy on the control of such hospital-acquired XDR-TB in resource-poor settings. A team of investigators from South Africa and the United States addressed this problem by employing a mathematical model to assess the effectiveness of available hospital-based infection control strategies on the future burden of XDR-TB in rural South Africa.
Clinical data for the mathematical model were obtained from Tugela Ferry and the Church of Scotland Hospital in Kwa-Zulu Natal province. An outpatient program for TB treatment and an HIV antiretroviral therapy (ART) program have been in place since 1993 and 2004, respectively.
The mathematical model incorporated over two years of inpatient and community-based data from a case-control study of 57 non-multidrug-resistant, 52 multidrug-resistant, and 61 XDR-TB patients. The model simulated rates of inpatient airborne and community TB transmission, and the effect of HIV and ART. Various administrative, environmental, and personal infection control measures were simulated.
The administrative measures included reducing hospitalisation time from 21 days to 5 days, deferred admission for some patients, drug-susceptibility studies, and involuntary detention of patients who refuse or default from treatment. The environmental measures included improved natural ventilation, mechanical ventilation, air filters, and ultraviolet germicidal irradiation.
The isolation scenarios simulated were individual isolation versus isolation clusters of five or ten patients. Rapid identification and isolation of multidrug-resistant TB patients were also simulated.
Personal protective measures included the use of respirators for staff and surgical masks for patients, voluntary HIV testing for staff and patients followed by initiation of ART for those who qualified, and re-deployment of HIV-infected staff to other hospital locations to reduce the risk of TB infections.
In the absence of any new interventions the model predicted about 1,300 cases of XDR-TB cases in Tugela Ferry by the end of 2012, more than half of which are likely to be hospital-acquired.
Although mask use alone by patients would prevent less than 10% of cases in the overall epidemic, it could prevent a large proportion of cases of XDR-TB in hospital staff. The combination of mask use with reduced hospitalisation time and a shift to outpatient therapy could prevent nearly a third of XDR-TB cases.
The combination of mask use, reduced hospitalisation, and shift to outpatient treatment when complemented with improved ventilation, rapid drug resistance testing, HIV treatment, and tuberculosis isolation facilities could prevent 48% (range 34–50%) of XDR-TB cases. By contrast, involuntary detention could result in an unexpected rise in incidence due to restricted isolation capacity.
The findings of Basu et al provide the important data required for the implementation of an evidence-based policy for the control of hospital-acquired XDR-TB in resource-poor settings. However, since XDR-TB transmission is on-going in the community, there is a need to develop and implement parallel community-based programmes, say the authors.
Basu S et al. Prevention of nosocomial transmission of extensively drug-resistant tuberculosis in rural South African district hospitals: an epidemiological modelling study. Lancet 370: 1500-1507, 2007.