TB or not TB?

A new test for the disease of pharaohs

The first improved test for latent tuberculosis in a century promises to revolutionise methods of diagnosis and control.

Tuberculosis has been around since the time of the pharaohs, and the microorganism that causes it, Mycobacterium tuberculosis, is still responsible for more deaths than any other single-celled infectious organism. Worldwide, there are about 8 million new cases of active disease every year, and in 2002 there were almost 7000 cases in the UK, a figure that has been steadily rising over the last decade. Yet the tools we have to stem this epidemic are remarkably old and inefficient: the BCG vaccine is 80 years old and is not always effective; anti-TB drugs are 20-30 years old, they have to be taken for six months, and resistance to them is increasing; and the method for diagnosis of latent infection is a century old and not ideal.

Latent infection - where people are infected with M. tuberculosis but show no signs of disease - is important because carriers without symptoms can remain undetected and go on to develop and spread the disease. Identifying and treating people with latent infection is thus key to TB control, but this strategy has been hampered by the limitations of the 100-year-old tuberculin skin test for TB. In both versions of this test - called Mantoux and Heaf - people are inoculated in the skin with a crude mixture of M. tuberculosis proteins. A positive diagnosis of infection is made 3-7 days later by assessing a bump - evidence of an immune response.

The main problem with the skin test is its lack of specificity - it often shows up positive in uninfected people who have been vaccinated with BCG, so it cannot distinguish between those who are vaccinated and those who are infected. The test is not especially sensitive either - some infected people test negative. Its use also presents practical difficulties and many people never return to have the test read.

Wellcome Trust Senior Clinical Research Fellow Dr Ajit Lalvani and his team in Oxford set out to see if they could better understand the immune response to TB, and use this information to develop a novel test. He and his colleagues looked at the activity of the T-cell component of the immune system, in people with active TB disease, those with latent infection, and in healthy subjects. The group discovered that two proteins raised an extremely strong immune response in both people with active TB and those with latent infection. When it turned out that these proteins were coded for within the nine-gene stretch that was missing in the BCG vaccine (see box below), they realised that the genetic differences could be exploited in a test that could distinguish between real infection and vaccination.

Unfortunately, unlike an antibody response, a T-cell response is technically difficult to detect. Dr Lalvani's team therefore attempted to develop a simplified test to measure T cells reactive to M. tuberculosis proteins. The product of their labours was a rapid version of an existing test; the new ex vivo ELISPOT can be performed overnight with T cells taken directly from blood. If a patient's T cells have recently encountered one of the two M. tuberculosis proteins in vivo, they rapidly release a factor known as interferon-gamma (IFN-g). The ex vivo ELISPOT detects the IFN-g released from individual 'effector-memory' T cells.

Having developed what they thought was a better test for TB infection, Dr Lalvani's group faced the problem of not having a 'gold standard' or benchmark test to compare it with. Was there a way around this? Dr Lalvani thought so. A key factor that determines TB transmission is the amount of time spent sharing room air with an infectious patient, so a more accurate test should correlate better with the level of exposure to an infectious case than the skin test.

An early pilot study of 50 TB contacts in London showed that ELISPOT results correlated with the extent of recent exposure to TB cases, but there were too few contacts to be certain that ELISPOT correlated significantly better with TB exposure than the skin test. So, in 2001, Dr Lalvani met with John Watson, Head of TB Control at the Health Protection Agency's Communicable Disease Surveillance Centre in Colindale, and told him about his team's results. He explained that although the ELISPOT test was promising and could potentially help to improve TB control, its effectiveness needed to be validated in the setting of a larger outbreak.

One week later the national newspapers reported on a major outbreak at a Leicester school. Dr Lalvani needed to work fast, and Dr Watson introduced him to the TB outbreak team at Leicestershire Health Authority. They shared his enthusiasm after seeing the results, and helped him set up a collaboration between his team, the health authority and the school. The aims of the study were to offer the best possible management to children and staff at risk of infection and simultaneously compare the accuracy of the old and new methods.

For Dr Lalvani, it was a major project. "I found myself speaking at school assemblies every morning for a week to explain to the children what the project was about, and speaking to the staff, and the headmaster, who gave his blessing. The school nurse played an absolutely key role, as did the Health Authority - and my team in Oxford did a great job too."

The timing was crucial. Dr Lalvani's team had from April until July, when term ended, to get as much as they could done. Written informed consent was obtained from the children and their parents. Many of the information letters and consent forms were translated into Hindi and Gujarati, because some of parents were recent immigrants who did not speak English.

According to the skin test results, there were 69 cases of active TB and 254 cases of latent TB in the school. Everything pointed to a single infectious source. In a remarkable stroke of luck, electronic school timetables allowed them to calculate how many minutes each child spent in the room air with the source case. The contacts were categorised into different groups in decreasing order of exposure to the source of infection.

But how did the ELISPOT results compare with those from the skin test? Although the results were similar, children who were more exposed to the student with full-blown tuberculosis (in terms of physical proximity and duration of exposure, such as children from the same class or school year) were significantly more likely to test positive with ELISPOT than with the skin test. Moreover, unlike the skin test, ELISPOT was not confounded by prior BCG vaccination, indicating that it was more specific than the skin test.

The team also wanted to answer another question: how much time sharing room air with an infectious person do you need before you can be certain you're going to get infected? The study answered this question for the first time ever: 130 hours. "As far as I know, this is probably the most detailed and informative study of TB transmission in an enclosed environment for 50 years," says Dr Lalvani. The findings, which were published earlier this year in the Lancet, stimulated wide media interest in print, radio, television and web outlets.

What next?

A spin-off company established by the University of Oxford is currently obtaining regulatory approval and developing ELISPOT into a product that can be used routinely in clinical practice. Back in the lab, Dr Lalvani has his eyes set on a better vaccine for TB. This will mean identifying exactly which immune responses provide protection against disease.

Dr Lalvani's team is tracking different T-cell populations in latently infected people over time and looking for differences between those who go on to develop disease, and those who maintain long-term immune control. "This will then tell us which T cells are mediating protection from TB," says Dr Lalvani. "In other words, which T-cell groups should be induced by vaccination. Because TB has such a long incubation period, it will take 10-20 years before we know how effective new experimental vaccines are. We hope that the T-cell populations we identify that correlate with protective immunity will also help those testing experimental vaccines to know if they are likely to work."

The BCG vaccine

The BCG vaccine was derived from Mycobacterium bovis, which causes TB in cows. Scientists at the Institut Pasteur regularly checked the bacterium’s virulence over several years by infecting guinea pigs, and in 1921 they discovered that it no longer killed them. To determine whether it was safe enough to use as a vaccine they tested it in a baby whose grandmother had infectious TB, a death sentence in those days. The baby didn’t contract TB, and the new strain was sent out to labs across the world.

Recent genetic studies showed that a chunk of nine genes was deleted from the BCG form of M. bovis before 1921. This gene segment is present in M. bovis in the wild, and in its human counterpart M. tuberculosis.

External links

Communicable Disease Surveillance Centre of the Health Protection Agency

TB or not TB?

A new test for the disease of pharaohs

What next?

See also

External links

Further reading