Rapid urine LAM test for HIV-associated TB – potential to reduce deaths?

January 9, 2020 0 By Jose Scott


The World Health Organization estimated that
in 2011 there were 1.1 million new or recurrent cases of TB in people living with HIV worldwide,
which is around 13% of total TB cases. This burden of HIV-associated TB is highly concentrated
in the countries of sub-Saharan Africa, which account for 79% of all cases. Disease rates
are highest in countries towards the south of the continent where HIV prevalence is highest.
Here between 50% and 80% of TB cases are HIV-coinfected. One country alone, South Africa, accounts
for almost 30% of all HIV-TB cases worldwide. Outside Africa, TB is also a common opportunistic
infection in people living with HIV in south-east Asia and South America and also among HIV-infected
injection drug users in the countries of eastern Europe and central Asia. HIV-TB causes approx 430,000 deaths per year,
although these are classified as HIV deaths in the International Classification of Diseases.
Around one third of people who die with TB are HIV-coinfected and conversely, around
a quarter of global HIV/AIDS deaths have TB as the underlying cause. So, TB is the leading
cause of death in people with HIV worldwide. TB diagnosis in resource-limited settings
relies heavily on sputum smear microscopy and chest radiology, and both of these are
impaired in people with HIV coinfection. TB diagnosis by sputum smear requires the release
of TB bacilli (Mycobacterium tuberculosis) in sufficient numbers so that there are at
least 10,000 organisms per ml of sputum, and that is is the limit of detection of the assay.
However, HIV impairs the host response to TB, such that in coinfected patients there
is reduced immunopathology in the lungs, which means that there is less inflammation and
less lung tissue damage. As a result, lower concentrations of bacilli are liberated into
sputum. As a result, sputum smears are much more likely to be negative. This issue is
further compounded by the fact that when patients have advanced disease and are very weak, such
as inpatients, it can be very difficult for them to expectorate good quality sputum samples.
In addition to smear microscopy, chest radiology is less useful in those with HIV coinfection.
Again, this is because of reduced immunopathology. The chest radiographic appearances are often
non-specific and lack the typical characteristics of pulmonary TB that are seen in HIV-negative
patients. HIV-coinfection also increases the frequency of extrapulmonary disease, which
again makes it more difficult to diagnose. The huge challenge of diagnosis of HIV TB
is graphically illustrated by a number of post-mortem studies of patients dying with
HIV/AIDS in hospitals in sub-Saharan Africa. These studies have repeatedly shown that between
a third and half of patients had evidence of TB, much of which also remained undiagnosed
at the time of death. Culture of clinical samples, especially in
liquid media, is the assay with the highest sensitivity for TB diagnosis. However, this
is slow, often yielding results in weeks rather than days. It is technologically demanding
and expensive and is only feasible in centralised laboratories. For these reasons it is not
generally available in many resource-limited settings.
However, a real landmark development is the Xpert MTB/RIF assay which was endorsed by
WHO in December 2010. This is a simplified fully automated real-time PCR assay system
that is cartridge-based and requires very limited training for operation. It takes just
two hours to generate a result and it has much higher sensitivity that sputum smear
microscopy. A single test can detect all smear-positive cases and approximately 70%-75% of smear-negative
cases, so that is a huge advance. However, there are a number of draw-backs. The hardware
is sophisticated and expensive. Each cartridge, even at heavily subsidized prices, costs $10.
Moreover, its use will largely be confined to laboratory settings as it is difficult
to implement this at the actual point-of-care within the clinic environment. When use is
laboratory-based, results cannot be used to inform immediate treatment decisions, and
that is a major drawback. So what is really needed is a low-cost point-of-care
assay that can be used to reliably diagnose TB and allow TB treatment to be started at
the same clinic visit. The Determine TB-LAM assay is one such test, and it has recently
become commercially available. Although this has not yet been endorsed and we need much
more evidence from ongoing studies, this assay has real potential to play a useful role in
diagnosis of HIV-TB. Lipoarabinomannan, or LAM for short, is a
glycolipid component of the cell wall of Mycobacterium tuberculosis. It is produced in large quantities
by the organism and I think this probably relates to the fact that it is immunomodulatory,
favouring survival of the organism in the human host. LAM can be detected in the urine
of a proportion of patients with TB, even if the primary focus of the disease is remote
from the renal tract, for example pulmonary disease. LAM is thought to gain entry to the
urine either via the bloodstream or via direct involvement of the renal tract by TB in those
with disseminated disease. Regardless of the mechanism, detection of LAM in urine can be
used as a means of TB diagnosis. TB diagnosis via detection of LAM in urine
can be done in the laboratory using a simple polyclonal antibody sandwich ELISA in a 96-well
plate format, and the initial diagnostic accuracy studies of the LAM assay were done using this
format of the assay. The ELISA has now been developed into a simple lateral-flow ‘strip-test’
format – which is an immunochromatographic assay that looks a bit like a urine pregnancy
test. So to run the test, all you need is a fresh
urine sample and you just apply 60 µL to the sample pad at the bottom of the test strip,
and it doesn’t require any prior processing. The strip is left for 25 minutes, during which
time the urine soaks up the test strip and immobilised capture and detection antibodies
labelled with colloidal gold lead to the development of a purple band in the test window. If that
band is of sufficient intensity when compared to a reference card, it is scored as a positive
test for LAM, and infers probable diagnosis of TB. There has been a series of studies published
since 2009 using urine LAM assays, and we’ve shown that the utility of LAM detection for
TB diagnosis is really restricted to patients with HIV infection and advanced immunodeficiency.
It is not useful if you’re HIV-negative ; and it’s not useful if you’re HIV-positive patients
and have a high CD4 count. So you can only really use this test applied to selected patients.
In those who we know are HIV-infected and have CD4 counts less than say 200 cells per
µL, the test is useful. The lower the CD4 cell counts and the sicker the patients (as
defined by a range of characteristics), the greater the sensitivity for TB diagnosis.
The likely reason for this is that the sicker the patient, the much more likely they are
to have disseminated disease and therefore have LAM in the urine.
The sensitivity of the assay entirely depends on the particular characteristics of the patients
tested. So in ambulatory out-patients with CD4 cell counts of less than 50 cells/uL,
the test sensitivity is maximal in around two-thirds of cases. Among in-patients who
are of course sicker, the proportion will be higher. Conversely, among ambulatory patients
with higher CD4 counts, the sensitivity will be lower. So it all depends on the patient
population you’re applying the test to. The major advantages of this test are obvious.
The test is low-cost (currently marketed at $3.50 per test), it’s rapid, results are available
within 25 minutes at a single clinic consultation, and it doesn’t require any equipment or sophisticated
hardware. Urine samples are easy to obtain even among sick patients in hospital who may
find it very difficult to cough up sputum samples. Urine is safe and easy to handle
and testing could readily be done by nursing staff following limited training. Most importantly,
it can be used at the point-of-care at the clinical consultation, allowing TB diagnosis
and the start of treatment at a single clinic visit, and this is the way by which it may
really reduce treatment delays. The disadvantages of the test are that it
can only be applied to specific patient subgroups; those who have got confirmed HIV and those
who have advanced immunodeficiency, whether that is defined by CD4 count criteria or WHO
clinical staging. The sensitivity of the assay is limited and so it can only be used as a
test to diagnose TB but we can’t use it to rule out TB. While several well-conducted
studies of the LAM ELISA and the point-of-care assays have found very high assay specificity,
some studies have found it to be more limited . Whether this is true or whether it’s an
artefact of study design remains to be fully clarified. Finally, we’re not sure what the
optimal cut-off for scoring the point-of-care strip tests is, scoring them positive or negative.
We need to take into account both the ease of use by health-care workers and also the
diagnostic accuracy. The very faint bands may be true positives, but if they’re very
difficult to read by healthcare workers it may lead to a false positive rate as well. So we need a range of further studies to clarify
the use of the TB-LAM point-of-care assay. We need a greater evidence base on the diagnostic
accuracy, we need operational research to clarify its use in the clinical environment
rather than the laboratory environment and finally, we also need large-scale studies
to assess the impact of implementation on clinical outcomes. Can this assay reduce deaths? There are several characteristics of the Determine
TB-LAM assay that make it potentially very useful in resource-limited settings. One of
the key observations is that although the sensitivity of the assay is somewhat limited,
the sensitivity is highest in the very sub-set of patients who have got the worst prognostic
characteristics and are at highest risk of death. Secondly, the assay allows TB to be
diagnosed and treated at a single clinic visit, and that’s a huge step forwards. So there
is the real potential to accelerate both diagnosis and treatment and thereby reduce mortality
risk. The assay, we must understand, is not a stand-alone
assay, and needs to be used in combination with other assays. In fact, there is incremental
sensitivity when used in combination with sputum smear microscopy or sputum Xpert MTB/RIF
testing. So, the LAM assay shortens the time to diagnosis in the sickest patients, allowing
immediate treatment, and it increases overall sensitivity when used in combination with
other assays whose results may be available in the following days. So this assay may come
to play an important role in the diagnostic algorithm. There are two key clinical populations in
which this assay may come to play an important role. First, the assay may be used to screen
for TB in HIV-infected out-patients who are accessing antiretroviral treatment clinics,
and also in in-patients it can be used to diagnose TB in HIV-infected in-patients who
are sick. By shortening the time to diagnosis in these key patient groups, we would hope
that their mortality risk would be reduced. So the real goal of this assay and the potential
of the assay to make a difference is to reduce deaths from HIV-TB.