TB screening assessed

“TB screening misses 70% of latent cases,” reported The Guardian . The newspaper says that experts have called for a change to tuberculosis (TB) screening policy. They suggest that a relatively new blood test should now be used to screen arrivals to the UK from the Indian subcontinent for hidden TB as well as arrivals from other high-risk areas of the world. This, they say, would mean treatment could be given to prevent most cases of the latent or hidden form of the disease from developing into full infectious TB.

This is a well designed study, and goes towards answering a clear and important question for policy decision makers. It is accompanied by an analysis of the overall cost of a change to policy and, importantly, the cost of averting an additional case of TB in people from different countries, allowing the researchers to suggest the best approach to screening those who come to the UK from these areas of the world.

Several papers have focused on the fact that the previous technique for screening for active TB, using X-ray alone, missed 70% of latent TB. This new strategy identified 92% of latent cases, therefore “missing” only 8%. UK guidance currently specifies the groups of people who are offered screening for active TB and includes those arriving in the UK from countries known to have high rates of TB. By specifying countries for screening for latent TB as well as active cases, and using this new test, it is likely that more people can be treated and cured of this increasingly common illness.

Recent NICE guidance, updated earlier this year, has a section on new entrant screening and advises a co-ordinated programme linked to local services that is designed to detect latent TB and start treatment where needed. A positive IGRA test is one of the suggested tests, along with a positive tuberculin skin test in people under 35 years. This study was unpublished at the time and NICE asked for this sort of cost-effectiveness study to target the treatment of latent TB better. There is a small difference between the guidance and the conclusions of this study, relating to the countries that are recommended for this sort of screening for latent TB. Details of this are given below.

Where did the story come from?

The study was carried out by researchers from Imperial College London and other TB services around the UK. The research was funded by the Medical Research Council. The study was published in the peer-reviewed medical journal The Lancet Infectious Diseases .

The news coverage is generally accurate. The newspapers do all emphasise the poor accuracy of chest X-rays when used as a screening test for TB, though the study did not look at this. They all then go on to describe the new study and its main findings, along with the researchers’ call for a change to screening policy. Quotes from a range of commentators are also included, including a comment from the Department of Health that the research backs up the latest guidance from the National Institute for Health and Clinical Excellence (NICE) on TB, issued in March 2011.

What kind of research was this?

This was a cohort study accompanied by a cost effectiveness analysis. Between 2008 and 2010 the researchers analysed data collected from 1,229 immigrants to the UK, from immigration centres in Westminster, Leeds and Blackburn. All three centres were using a relatively new blood test called the interferon-gamma release-assay (IGRA) specifically to test for TB. Only people aged 35 or younger who had been screened for latent or hidden TB infection using this test were included in the analysis. In a decision analysis model, the results for people from different countries of origin were modelled separately so that the researchers could test the strategy for different levels of underlying TB.

The research was carefully conducted and has provided a clear answer to the question of how accurate the test is when used in a population similar to those in these centres. It has also provided an estimate of cost effectiveness and the cost per case of avoided TB, the results of which look favourable and will help to inform immigrant screening policy. There are some practical limitations to how the study was conducted in terms of selection of patients and the assumptions the researchers had to make in the decision model. There are also different IGRA tests available, so this one may not necessarily be the best one. Despite these points, the testing strategy looks promising.

What did the research involve?

The researchers explain that cases of diagnosed TB have risen in the UK from 6,167 to 9,040 in the 10 years to 2009, and this is mainly due to rising numbers of cases in foreign-born immigrants. They say that national guidance for immigrant screening is hampered by a lack of data. They wanted to address this shortfall by finding out the number of cases of latent infection in immigrants to the UK and by examining the prevalence (rate of latent cases found per 100,000 population) so that they could define the groups that should be screened. They also wanted to model the cost-effectiveness of different strategies so that they could estimate the number of additional cases of full TB that could potentially be avoided with each strategy and at what cost to the tax payer.

TB is a bacterial infection caught by breathing in the bacteria that cause it. These bacteria are spread through sneezing or coughing by someone who has TB. There are two main types of TB, active and latent. In active TB, some people become ill a few weeks or months after breathing in the bacteria and can spread the disease. However, in most people, the body’s immune system kills the bacteria and the person does not get ill. In other people, the bacteria are not killed but stay in the body at a low level, and the person does not get ill and is not infectious. This is called latent TB. The bacteria can start to multiply again months or years later (for example, if the person’s immune system is weakened by another disease such as HIV) and active TB can develop.

In this study, the participants were all foreign-born new entrants who had come to the UK within the past five years and were aged 35 years or younger. They were screened between January 2008 and July 2010 in Westminster, Leeds and Blackburn following referral by “port-of-entry” screening systems, health protection units or after registration with primary-care services. These centres serve a total of 1.6 million people, of which 6.5% are born abroad.

All participants were screened first with a symptom questionnaire followed by four blood tests, including the one-step IGRA test of interest. Immigrants who were symptomatic or who had a positive IGRA result were referred for chest radiography and further clinical assessment to see if they had active tuberculosis. Those with latent infection were offered treatment with either three months of two drugs or six months of one drug, in accordance with their wishes and standard UK guidance.
The researchers also asked about age and sex, BCG vaccination status (ascertained through documentary evidence, reliable history of vaccination or a characteristic scar) and country of origin.

The researchers used standard techniques for their cost effectiveness analysis. They costed the benefit of using this test from a UK National Health Service perspective, modelling the use of IGRA testing for 20 years. Two main questions were asked:

• What are the costs of screening at different incidence thresholds?
• Is screening at specific thresholds a cost effective use of resources and, if so, at what threshold?

What were the basic results?

Of 1,229 immigrants, 245 (20%) tested positive in the IGRA tests, 982 (80%) tested negative and two people (0.2%) had indeterminate results.

They say that positive results were independently linked to the TB incidence in immigrants’ countries of origin. This means that test results were more likely to be positive in countries with higher rates of TB after taking into account other factors that were also linked to an increase in rates (male sex and age).

They say that the current national policy for detecting active TB used a chest X-ray in people from countries where more than 40 per 100,000 population per year develop TB. If this was used to screen for latent TB, it would fail to detect 71% of individuals with latent infection.

From the modelling analysis, they found that the most cost-effective strategy would be to screen people for latent TB from countries with a TB incidence of more than 250 cases per 100,000 per year. Using the IGRA test would result in an additional cost of £17,956 for each case of tuberculosis prevented compared with the next most effective strategy.

The next most cost effective strategy would be also to screen immigrants from the Indian subcontinent, where there are more than 150 TB cases per 100,000 people per year. It was estimated that this would identify 92% of infected immigrants and prevent an additional 29 cases of TB over 20 years compared with no screening.

How did the researchers interpret the results?

The researchers say that implementation of screening for latent infection would be cost effective. They recommend the level of incidence (150 cases per 100,000 per year) that identifies most immigrants with latent tuberculosis, and which is likely to prevent substantial numbers of future cases of active TB.

Conclusion

Until recently, it has been unclear who best to screen for latent TB. This research supports recent decisions made by NICE regarding how to screen and adds to the evidence on who, from a cost-effectiveness perspective, it might be best to target. This was clearly an area that needed research, as the screening for active TB using a chest X-ray was not effective at identifying latent TB. There are several points the researchers make about their research:

• UK national policy from NICE since 2006 specifies that immigrants who intend to stay in the UK for more than six months need to be identified at their port of entry, and those from certain countries with normal chest X-rays need screening for latent TB. This includes children aged less than 16 years from countries with a tuberculosis incidence or more than 40 per 100 000 per year, and 16–35-year-olds from either sub-Saharan countries or from those with a disease incidence of more than 500 per 100,000 per year. Individuals older than 35 years are not screened because the risks of treatment outweigh the potential benefits.
• The suggestion here is to widen the groups referred for testing to include screening for latent TB for people aged 16 to 35 years from countries that have rates above 150 cases per 100,000 per year and those from the Indian sub-continent. This is, in effect, a different screening approach and will result in more people being screened from the countries with lower rates of TB.
• In this cohort, the prevalence of latent infection was moderately high at 20%, and it is not clear what the accuracy and cost effectiveness of the test would be in populations with a lower prevalence of latent TB. The researchers say that their study may be biased towards showing an increased prevalence of latent infection, as people who were concerned that they may have the illness may be more likely to attend for screening.
• The researchers were unable to test the accuracy of the test against tuberculin skin testing, which is an alternative way of testing for latent infection, as this skin test is not routinely performed on new entrants.
• The economic models rely on some assumptions, including that all patients identified with latent infection are treated and cured. Different estimates could result from assuming some drug-resistance, for example.

Overall, this is a useful study which is likely to be discussed by those deciding on immigration screening policies. The one-step IGRA blood test may prove to be the preferred option, however it is too soon to say this is the best approach. It is not the only test and further work is needed to compare different screening protocols (such as tuberculin skin test with IGRA compared with skin test alone or IGRA alone). There are also different types of IGRA test, some of which may have different costs.