Researchers Create Scoring System to Predict TB Risk in Adults

Researchers from Imperial College in London have created a scoring system designed to predict 10-year tuberculosis (TB) risk in adult contacts of index cases.

When it comes to screening for tuberculosis (TB), current World Health Organization (WHO) guidelines are clear: for index cases of pulmonary TB, all household contacts should be investigated via screening and surveillance.

Unfortunately, in practice, these recommendations are often easier said than done, particularly in the resource-poor areas in which TB is prevalent. However, the authors of a study published on August 18, 2017 in the journal Lancet Infectious Diseases, believe they have developed a tool to make this process easier—and affordable—for clinicians in these high-risk communities: a scoring system designed to predict 10-year TB risk in adult contacts of index cases. The research project received funding from a number of important sources in the TB world, including the Wellcome Trust, the Bill & Melinda Gates Foundation, the World Health Organization (WHO), and TB REACH.

“In the majority of resource-constrained settings where much of the world's TB occurs, management of contacts is given low priority because TB programs are focused on diagnosing and managing patients with TB. [These countries] have few resources allocated to TB prevention, particularly among adult contacts who are often ineligible for preventive treatment,” study co-authors Matthew J. Saunders, MBChB, and Carlton A. Evans, FRCP, both of Imperial College, London, said in an email to Contagion®. “Furthermore, contacts often have to follow a long road from being identified to complete screening, and ultimately, to be prescribed preventive therapy… completing a variety of tests to exclude active TB, frequently including a tuberculin skin test, and attending multiple appointments at clinics that are difficult to access and associated with direct and indirect costs.”

However, Saunders and Evans believe their scoring tool has the potential to simplify this process. From 2002 to 2006, they and their colleagues enrolled 2,017 household contacts (15 years of age and older) of 715 index cases with pulmonary TB who lived in shanty towns in Ventanilla, Peru, and they followed the contacts for more than a decade (median follow-up 10.7 years), until February 2016. Of the 2,017 contacts, 178 (9%) developed TB during 19,147 person-years of follow-up.

For each year of follow-up, the authors determined the incidence of TB per 100 person-years. Using a Cox proportional hazards model to identify index cases, household contacts, and 9 household risk factors for the disease—including sustained exposure to the index case, lower community household socioeconomic position, indoor air pollution, and “living in a household with a low number of windows per room”—they created a “continuous, integer-point risk score… [using] whole numbers rather than exact regression coefficients to create an easily calculable score for field use.” The score is calculated based on the contact’s body mass index; from that number, a certain number of points is subtracted for each risk factor present (eg, 4 points for a contact with a history of TB, 2 points for households that cook foods with crop wastes, wood, or kerosene). The authors classified those with scores >19 as low risk, those with scores between 18 and 12 as medium risk, and those with scores of ≤11 as high risk.

“Because our risk score informs actual 10-year risk of developing TB, healthcare providers are able to determine their own cut offs for a ‘high-risk’ contact, based on prescriber preferences and the availability of resources in each setting,” Drs. Saunders and Evans explained. “In Peru, we have chosen to opt for a more inclusive definition of ‘high-risk,’ where any contact judged to have a higher TB risk than the risk in the community is advised to start preventive therapy.”

In their study population, they found that the 10-year risks of TB in the low-, medium-, and high-risk groups were 2.8%, 6.2%, and 20.6%, respectively. In all, the 535 contacts classified as high risk accounted for 60% of the TB cases identified. The authors also noted that their scoring system predicted TB “independently of tuberculin skin test and index-case drug sensitivity results.” They validated the score using an urban cohort recruited in Callao, Peru.

“Our score provides a paradigm shift from the current approach where all contacts are treated the same and enables health workers to focus their resources and target TB prevention interventions to contacts who are particularly likely to benefit,” according to Drs. Saunders and Evans. “Because our score does not need any laboratory or clinical tests, in the short-term we believe our score could be immediately implemented… in other resource-constrained settings where currently very few adult contacts who are at high-risk of TB are considered for preventive therapy. However, challenging established beliefs about the use of preventive therapy and encouraging scale-up among all contact of patients with TB, not just children aged less than 5 years [of age], is more complicated and will require concerted effort and engagement from healthcare providers at all levels.”

Brian P. Dunleavy is a medical writer and editor based in New York. His work has appeared in numerous healthcare-related publications. He is the former editor of Infectious Disease Special Edition.