Researchers at the Institute of Tropical Medicine (ITM) have successfully developed and tested a combination of simple clinical, radiological and laboratory tools to diagnose smear-negative tuberculosis (SN-TB). The strategy outperforms those currently directed at countries lacking the money or medical infrastructure to use methods available in industrialised countries, including the recommended World Health Organization (WHO) approach. The results of the six-year study in Peru provide much needed scientific evidence on the diagnosis of SN-TB cases in HIV-negative patients. While strategies for diagnosing SN-TB have been recently updated for HIV-positive cases, there was almost no evidence for HIV-negative patients thus far.
Smear-negative tuberculosis refers to TB cases which go undetected when sputum is examined on a microscope slide, a standard diagnostic method. They add up to around 35% of total TB cases. When doubts persist after a negative sputum test, WHO recommendations include the response to an 'empirical' treatment with antibiotics to decide whether a patient has SN-TB. This approach is a point of reference for health practitioners around the world, but the collaborative ITM study in Peru concludes that it is ineffective and needs updating.
"The WHO algorithm is based on expertise rather than on scientific evidence. Our findings show it simply doesn't work. We put forward a tested alternative that does work, based on a combination of accessible clinical, radiological and laboratory tools. This evidence-based diagnosis, which also works when money or advanced medical infrastructure are lacking, is good news for TB patients around the world," said Dr. Alonso Ricardo Soto Tarazona of the Hospital Nacional Hipólito Unanue, Lima, Peru.
Based on scientific evidence gained over the last few years, the WHO has updated its approach for the diagnosis of SN-TB in patients infected with HIV. Combined HIV-TB infections are frequent in many African countries. Out of 8,8 million people who contracted TB worldwide in 2010, 1,1 were also HIV-positive. Unfortunately, recommendations for HIV-negative cases have not changed in decades. The study in Peru, a country with a high incidence of TB and a concentrated HIV-epidemic, therefore contributes significant evidence to an area where scientific proof is lacking.
Diagnosing SN-TB in three simple steps
The new detection strategy is a combination of three diagnostic tools:
· Sputum concentration
· A score-based tool to make a clinical prediction
· Manual liquid cultures, a decades old technique revived by the researchers
First, researchers analyse a concentrated extract of sputum, obtained after centrifugation. If the bacterium is not found, the next step is a score-based clinical prediction based on easily identifiable symptoms and radiological findings.
"We identified coughing up blood and weight loss as symptoms associated with an increased probability of smear-negative tuberculosis. Ability to expectorate and age over 45 years, on the other hand, were negatively associated. Shadows in the upper region of the lungs or a diffuse presence of tiny dots, observed via X-ray, increased the probability a person was ill," said Soto Tarazona.
The resulting score (high, intermediate or low) informs further steps. A suspected patient with a high score is immediately referred to treatment. Patients with low probability scores should be screened for other diseases. An intermediate score, associated with a stagnating clinical picture, leads to step three: the use of manual liquid cultures. Automatised liquid cultures, in which the TB agent multiplies in a predetermined medium, are commonly used in industrialised countries to determine the cause of infectious diseases. However, the technology required is expensive and manual techniques are all but lost.
"We delved into decades old scientific literature to rediscover the origins of liquid culture, which started as a manual technique. You can say we saved it from oblivion and showed that this technique is still useful. Manual liquid culture is a valuable alternative when a costly high-tech device simply is not an option. This is often the case in developing countries, but also in emerging economies," said Soto Tarazona.
Study shows large language models susceptible to misinformation