Theo Smart
The lateral-flow urine LAM (liparabinomannan) test for tuberculosis — a simple
inexpensive strip test for tuberculosis — is a feasible point-of-care test in hospitalised
South African adults living with HIV and, if people are then quickly put on
effective treatment, would be a very cost-effective diagnostic strategy in
such patients, according to a South African study presented on Wednesday afternoon
at the Nineteenth International AIDS Conference (AIDS 2012) in Washington DC.
“It is a cost-effective diagnostic strategy, having an
incremental cost-effectiveness ratio (ICER) of $1370 per disability-adjusted
life-year (DALY) averted, and this is less than the per capita GDP in South
Africa of over $7000,” said Dr Di Sun of Johns Hopkins Bloomberg School of
Public Health, who presented the findings. “This remained robust across a wide
range of sensitivity and uncertainty analysis.”
Five or six years ago, TB-HIV activists began drawing
attention to the shocking inadequacy of the existing TB diagnostics — smear microscopy and culturing — which
were relatively unchanged in over 120 years since they had been established by
Dr Robert Koch, who discovered Mycobacterium tuberculosis (M.TB).
Smear microscopy fails to detect the majority of cases in people who are
co-infected with HIV, resulting in delays in diagnosis and treatment, and all
too often in death. Even culture fails to detect about a quarter of the TB
cases that have to be diagnosed clinically.
As a field, TB diagnostics research and development was
virtually non-existent. But the
activists caught the attention of the Bill & Melinda Gates Foundation and
others, who invested heavily in the Foundation for Innovative Diagnostics and
other concerns, setting off a flurry of activity. A product pipeline evolved,
out of which GeneXpert has emerged to great fanfare, despite being too
expensive and having running requirements too complex to put into every primary
health care facility — or for that matter even district hospitals.
“Dropping the machine with a parachute and cartridge will
not do the job — you need quite a lot of logistical back-up to install this
machine in lower resource settings — you need a stable electrical power supply
which is very challenging in the settings we are working in, and you also need
air conditioning for both the machine and the cartridges; the temperature
should not rise above 30 degrees,” said Dr Steven Van Den Broucke of MSF during
a TB diagnostics session almost entirely devoted to the Xpert MTB/Rif tests at
the conference on Thursday.
The only exception was during the overview on TB diagnostics
given by Professor Gavin Churchyard of the Aurum Institute, who mentioned a
couple of other assays entering into use, including those that try to detect lipoarabinomannan (LAM), a component
of MTB’s cell wall that can be detected in urine samples, when it gets released
from metabolically active or degraded MTB.
Urine tests for TB are appealing for a couple of reasons,
according to Dr Churchyard. “Urine is easy to obtain,” he said, particularly
from patients who may have trouble producing sputum – a common problem in
people living with HIV who have extrapulmonary TB. Plus it lacks infection
control issues associated with handling blood or sputum.
An earlier version of the LAM urine antigen test was ELISA-based,
which would have to be used in a centralised laboratory. However, a new form of
the test uses the Determine testing platform, requires no sample processing and
produces results in 25 minutes.
In other words, it is a test that can be performed at the
point of care, whichever medical facility the patient is in.
The test isn’t perfectly sensitive or specific. Its highest
sensitivity is in people with high MTB burden, who have more detectable antigen
in urine, in immunosuppressed patients, and in those with disseminated TB.
The test seems to perform much better in people living with
advanced HIV, particularly those with CD4 cell counts below 100. Dr Churchyard
presented a table summarising the results using the point-of-care test thus far in people
living with HIV.
Determine
TB-LAM
Author/
Year
|
N
|
Setting
|
Sensitivity
|
Specificity
|
|
|
|
Overall
|
CD4<100
|
|
Peter,
2012
|
335
|
Inpatients
|
45%
|
|
96%
|
Lawn,
2012
|
516
|
ART
clinic
|
28%
|
52%
|
99%
|
Dorman S,
2012
(Interim unpublished data)
|
561
|
Outpatients
Inpatients
|
45%
|
80%
|
90%
|
Dr Sun noted that the lateral flow test has a few other
profound advantages: it requires minimal training, and no expensive additional equipment that has to be airlifted to
remote facilities in resource-limited settings.
The cost of each test is also profoundly less expensive than
Xpert MTB/RIF, which was about $17 per cartridge test — though as cost-analysis
studies in South Africa have shown, once shipping costs and other expenses were
added in, the cost of the test was closer to $32 (See HIV
& AIDS Treatment in Practice's review of LAM and other diagnostic assays
for further information). More recently, UNITAID and other funding partners
agreed to collaborate on a ‘buy-down’, essentially paying a percentage of the
cost of the cartridge so that the cost to national TB programmes in resource-limited
settings would be $10 each — though again, shipping and other costs involved in getting the
cartridges into the country may not change that much.
The LAM lateral flow assays aren’t bulky, don’t weigh much
and don’t require air conditioning. They cost roughly $3.50 per test.
But that still doesn’t mean it would necessarily be cost-effective to
roll-out the test widely, or that it would improve upon the
already available lab tests (smear microscopy, chest X-ray).
The purpose of the study presented this week was to evaluate
the cost-effectiveness of a lateral-flow urine LAM assay in HIV-infected South
African adults and the economic conditions under which it is most likely to be
preferred.
This, other studies have suggested, would be for the most ill
people living with HIV, who are waiting in the hospital for a
diagnosis.
The cost-effectiveness analysis considered certain aspects
of the diagnostic decision as being constant, such as:
ALL patients would
receive the same existing
diagnostic tests regardless of whether the LAM assay was added or not (and the
costs of those tests would be a constant.
A positive result on smear
microscopy or LAM would get treatment.
Undiagnosed, untreated TB in these
patients would lead to death.
A proportion of undiagnosed cases would be
treated anyway based upon the clinician’s judgment.
Parameter
values and unit costs were drawn from studies performed in the South African
setting. The estimates (for instance, of life gained on treatment) used were
conservative.
The primary outcome, was the incremental cost-effectiveness
ratio that adding LAM into the diagnostic decision would yield, in terms of
cost per DALY averted. Sensitivity and uncertainty analyses were performed on
all parameters.
Results
By adding
the LAM test, they would be able to diagnose 80 more true cases of TB, at a
cost of 25 false-positive TB cases. These false positives occur because the
test is only 95% specific, However, it should be noted that specificity is
determined in reference to culture as the gold standard for diagnosis — and yet
culture misses a substantial proportion of cases, particularly in this
population.
All of this
comes at an additional incremental cost of $79,000 (mostly the cost of
treatment for these cases).
Cost-Effectiveness of Adding Lateral-flow LAM
to Standard TB Diagnostics
|
Existing
diagnostics
|
Existing
diagnostics +
Urinary TB-LAM
|
Cohort size
|
1000
|
1000
|
TB cases
|
380
|
380
|
TB cases treated
|
262
|
342
|
False positives
treated
|
130
|
155
|
DALYs
|
495
|
437
|
DALYs averted
|
|
58
|
Cost
|
$299,000
|
$378,000
|
Incremental cost
|
|
$ 79,000
|
ICER $/ DALY
|
|
$1370
|
The
addition of urine lateral-flow LAM averted 58 DALYs at a cost of $1370 per DALY
averted (95% uncertainty range: $710-3396). Even if Dr Sun and colleagues had
used high values in the sensitivity analysis, the cost per DALY averted is much less than the GDP per capita of
South Africa of $7275.It should be noted however, that as of yet, there are
no empirical evidence from other trials that the addition of urinary LAM
improves survival. In addition, the results may not be generalisable to other
populations, such as the outpatient setting or in other high-burden settings settings.
However a three-way sensitivity analysis Dr Sun presented
suggested that for a test with 95% specificity to be cost effective, if the
resulting life expectancy gained from treatment is only 1.5 years, the TB
prevalence would need to be at least 5%. If however, the life expectancy is 5
years, TB prevalence must be at least 1%. But if TB treatment prevents TB-related
death, and the patient is put on effective antiretroviral therapy, life
expectancy would be much greater, making the test begin to look cost effective
in advanced people living with HIV even in settings with a substantially lower TB
prevalence.
“This illustrates the importance of extending the life
expectancy of our population of interest, which can be done by putting them on
antiretroviral treatment, which will cause LAM testing to be much more
cost-effective,” Dr Sun said.
“Cost-effectiveness
depends most strongly on LAM specificity, life expectancy, and TB prevalence,
and it is highly cost-effective with longer life expectancies,” she concluded.
Reference
Sun Di et al. Cost-effectiveness of a lateral-flow urine lipoarabinomannan test for TB diagnosis in HIV-infected South African adults. 19th International AIDS Conference, Washington DC, abstract TUAE0101, 2012. (View the abstract of the session on the conference website).
For more information on TB diagnostics, visit the archive of HIV & AIDS treatment in practice at www.aidsmap.com/hatip.