Predicting tuberculosis: how a simple score card could hold the key to more effective TB prevention

18 September 2017

A new low-cost, low-tech method for predicting an adult’s risk of developing tuberculosis (TB) could represent a major step forward for TB prevention.

Women sitting outside doorway in Peru

Despite overwhelming evidence that preventive therapy effectively reduces people’s risk of developing TB after they have been exposed to it, current methods for predicting who is likely to develop TB remain inaccurate. This makes it hard to target prevention services effectively. No more than 2·7% of people who test positive for latent (undeveloped) TB infection, for example, will develop TB within two years, and the test needed to gain this result, the tuberculin skin test (TST), is expensive and clinically complex.

To address this issue, researchers developed a TB risk-score and piloted it on two communities in Peru over a 10-year period.

The score is based on nine established and readily collectable clinical and demographic TB risk factors, such as body mass index and housing ventilation. Unlike other risk-based approaches, the score does not include TST results or factors that rely on expensive equipment to measure.

Study participants were aged 15 and over and had been in the same house as someone with a confirmed case of pulmonary TB (known as the ‘index case’) for more than six hours a week in the two weeks before TB was diagnosed.

Participants were divided into groups according to their likelihood of developing TB, as predicted by the risk-score. Around a third (30%) were low risk (with a 2.8% risk of developing TB), 44% were medium risk (6.2% TB-risk), and 27% were high risk (20.6% TB-risk).

The risk-score was able to accurately predict someone’s risk of developing TB over a 10-year period. People who were well nourished, living in well ventilated houses among others with no history of TB, and low exposure to the index case, were predicted to have a 10-year risk of one in 100.

People who were underweight, living in poorly ventilated houses among people with a history of TB, and substantial exposure to the index case, were predicted to have a 10-year risk of TB approaching one in three. Associations were also found between age (either being 15–19 years or 50 years or older).

Of the 178 contacts who developed TB, 10% were in the low-risk group, 30% in the medium-risk group, and 60% in the high-risk group. The incidence of TB was highest in the first four years after exposure, and was more than double the incidence in the local population for the duration of the study.

Assuming that preventive therapy is 75% effective, these results suggest that the number of people who would need prevention therapy to prevent one additional case is 48 in the low-risk group, 22 in the medium-risk group, and six in the high-risk group.

In comparison, at least 37 people would need treatment to prevent one case when using TST results as a risk indicator. For this reason, using the risk-score could bring about a paradigm shift from the current ‘one size fits all’ approach to TB prevention to more targeted screening, surveillance, and treatment.

It may also help inform TB prevention education and result in the factors that put people at risk of TB being better addressed, for example through the provision of clean cooking stoves or ventilation maintenance.

The risk -score developed for use within this study is setting-specific, meaning it will need to be adapted to other contexts. In its current form, the risk-score does not assess factors such as drug resistance, HIV, and diabetes, which may need to be considered in high prevalence areas. It may also need to be adapted to assess exposure outside the household, which can substantially contribute to transmission in certain communities.

To reach the study in full visit The Lancet’s website.

Photo credit:

Written by Hester Phillips