Do new diagnostic tests render a study of empiric TB treatment unethical?

One of the ‘discussions in the corridors’ at the Union World Conference on Lung Health in Lille in October last year, was about the ethics of a study known as ‘ACTG 5274’ or the ‘REMEMBER’ trial. REMEMBER stands for ‘Reducing Early Mortality and Early Morbidity by Empiric Tuberculosis (TB) Treatment Regimens’. Giving someone ‘empiric’ treatment means treating them for something they might have, before it has been confirmed. The study was described in this blog a couple of months ago – it is recruiting people with HIV who present late for care (e.g. are diagnosed with a CD4 cell count below 50). In the study, people are randomly assigned to one of two groups:

1. Empiric TB treatment followed by antiretroviral therapy (ART), on the presumption they may have potentially life-threatening but hard-to-diagnose TB.
2. Immediate ART, and potentially TB treatment if they get a TB diagnosis (standard of care).

Differing opinions

In the impromptu debate at the meeting, Mark Harrington of the Treatment Action Group (TAG) argued in favour of the study, stressing how great the burden of undiagnosed TB is in people who present with such advanced HIV disease, and how high the early mortality rate, most often of TB, is among people who start ART at this point.

Professor Annelies Van Rie of the University of North Carolina said that she believed the risk of putting many weak people who didn’t have TB onto TB treatment was too great.

Mark replied, “well, that’s why we need to run a clinical trial.”

But Prof. Van Rie argued the new TB diagnostic tests made the study unnecessary or, at the very least, would make the study irrelevant by the time its results come out (the study is slated to run until 2014).

There was a fair amount of information presented at the conference on the costing and implementation of Xpert MTB/RIF test, which was discussed in HATIP #191. We thought using the new diagnostics would be too expensive, “Xpert MTB/RIF costs around $32 per test, so it may not be affordable in all settings.”

“I’m not talking about GeneXpert,” said Prof. Van Rie, “I’m talking about the LAM urinary antigen test which can be used to ‘rule in’ TB in this population, and which is now available in a strip form that can be used at the point of care, and it only costs about $3.50.”

Though not discussed much at the conference, earlier that week, Professor Steve Lawn of the Desmond Tutu Clinic and the London School of Hygiene & Tropical Medicine, suggested HATIP investigate some of the new data coming out on the TB-Determine LAM lateral flow tests. In a paper published on the test, Prof. Lawn and colleagues wrote:

“In the absence of suitable diagnostic assays, empirical treatment has been suggested as a strategy to reduce the high mortality of patients with very low CD4 cell counts in settings with the highest disease burden. However, with the development of this simple point-of-care assay, such a strategy might prove unnecessary.”

Just two years ago, Dr Lawn had advanced the concept of the empiric TB treatment study at the Union World Conference on Lung Health in Cancun. The debate was described in the blog entry referred to earlier.

At the conclusion of that debate, Professor Helen Ayles, of ZAMBART and the London School of Hygiene & Tropical Medicine, said that such a study would be unnecessary if it were possible to rapidly diagnose TB in people with low CD4 cell counts. “If we get a point-of-care diagnostic, this whole thing is out the window.”

The potential of new diagnostics

At the time, Professor Lawn pointed out that the best available test wasn’t quite accurate enough and would prove cost-prohibitive.

The Xpert MTB/RIF test could diagnose most smear-negative TB in people living with HIV, though it would take three cartridges (repeating the test three times) to do so. This will add to the expense of the test, which is already significant.

Some modelling data on the cost effectiveness of running multiple Xpert tests to diagnose smear-negative TB in people living with HIV – versus symptom screening with smear microscopy – were presented at the International AIDS Society conference in Rome last summer. As HATIP reported at the time, the modelling used rather optimistic estimates of the cost of the Xpert tests, and didn’t really consider the strategy being proposed in South Africa’s new first-line TB diagnostic algorithm.

Nevertheless, knowing that the test can be used this way poses a challenge for patient advocates who ought to demand that anyone who presents to care with low CD4 cell counts should get the best diagnostic services that can be made available, in order to avoid a life-threatening case of TB. But the cost of implementing Xpert could divert limited resources away from other low-cost measures and undermine TB control and other health system priorities. In fact, for some programmes, the cost of diagnosing some smear-negative TB cases by conducting multiple Xpert tests would be similar to the cost of putting someone on ART for a year.

Lawn et al. note in the paper on LAM that even the cost of “one [GeneXpert] cartridge (US$18 at the time of the study) [editor’s note: this does not include the other costs of running the tests, which health economists in South Africa say is closer to $32 per test] would represent the annual total health spending per head in many poor countries."

On the other hand, at $3.50 a test, the LAM urinary antigen strip test might be a far more cost-effective alternative in this population.

Data on the Determine-TB LAM lateral flow test are discussed in HATIP #193. On its own, the LAM strip test could identify two-thirds of the TB cases in HIV-positive people with CD4 cell counts below 50, and almost three-quarters of the TB cases when combined with microscopy – at least in the hands of Prof. Lawn et al. It wasn’t quite as sensitive in the two other major studies – but these used somewhat stricter criteria for a positive result than suggested by the manufacturer, in order to improve the tests specificity to an acceptable level.

Implications

Given the potential to identify a substantial proportion of the people with very low CD4 counts who have TB, is it really ethical to continue the REMEMBER trial, in which half the participants will take empiric TB treatment?

One concern is that, thus far, the LAM test has only been tested in Uganda and South Africa. Furthermore, use of the test to guide clinical management remains untested.

“Studies of the use of this assay and its effects on clinical outcomes in such patient groups are now needed,” Lawn et al. wrote.

But how and when should this happen? Should REMEMBER be put on hold, or perhaps its design changed to incorporate the use of the LAM strip tests at some sites?

We would like to open this up to discussion among our panel and readership, to include your responses in the HATIP blog. Please contact us at info@nam.org.uk.

Further information and references

The study by Prof. Lawn et al. was published in The Lancet and is available online (requires registration).

Lawn SD et al. Diagnostic accuracy of a low-cost, urine antigen, point-of-care screening assay for HIV-associated pulmonary tuberculosis before antiretroviral therapy: a descriptive study. The Lancet online: DOI:10.1016/S1473-3099(11)70251-1, October, 2011.

For further information on the REMEMBER study, visit the HATIP blogpost on the subject or the US National Institutes for Health Clinical Trials website.

A full archive of the HIV & AIDS Treatment in Practice bulletin is available at www.aidsmap.com/hatip and you can sign up free for future emails at www.aidsmap.com/bulletins.

Visit the HATIP Archive.