Lipoarabinomannan in urine during tuberculosis treatment: association with host and pathogen factors and mycobacteriuria.

Abstract

Detection of lipoarabinomannan (LAM), a Mycobacterium tuberculosis (Mtb) cell wall antigen, is a potentially attractive diagnostic. However, the LAM-ELISA assay has demonstrated variable sensitivity in diagnosing TB in diverse clinical populations. We therefore explored pathogen and host factors potentially impacting LAM detection.
LAM-ELISA assay testing, sputum smear and culture status, HIV status, CD4 cell count, proteinuria and TB outcomes were prospectively determined in adults diagnosed with TB and commencing TB treatment at a South African township TB clinic. Sputum TB isolates were characterised by IS61110-based restriction fragment length polymorphism (RFLP) and urines were tested for mycobacteriuria by Xpert® MTB/RIF assay.
32/199 (16.1%) of patients tested LAM-ELISA positive. Median optical density and proportion testing LAM positive remained unchanged during 2 weeks of treatment and then declined over 24 weeks. LAM was associated with positive sputum smear and culture status, HIV infection and low CD4 cell counts but not proteinuria, RFLP strain or TB treatment outcome. The sensitivity of LAM for TB in HIV-infected patients with CD4 counts of ≥ 200, 100-199, 50-99, and < 50 cells/μl, was 15.2%, 32%, 42.9%, and 69.2% respectively. Mycobacteriuria was found in 15/32 (46.9%) of LAM positive patients and in none of the LAM negative controls.
Urinary LAM was related to host immune factors, was unrelated to Mtb strain and declined steadily after an initial 2 weeks of TB treatment. The strong association of urine LAM with mycobacteriuria is a new finding, indicating frequent TB involvement of the renal tract in advanced HIV infection.

Full text

Page 1

RESEARCH ARTICLEOpen Access
Lipoarabinomannan in urine during tuberculosis
treatment: association with host and pathogen
factors and mycobacteriuria
Robin Wood1,2,3, Kimberly Racow1*, Linda-Gail Bekker1,2, Keren Middelkoop1,2, Monica Vogt1, Barry N Kreiswirth4
and Stephen D Lawn1,5
Abstract
Background: Detection of lipoarabinomannan (LAM), a Mycobacterium tuberculosis (Mtb) cell wall antigen, is a
potentially attractive diagnostic. However, the LAM-ELISA assay has demonstrated variable sensitivity in diagnosing
TB in diverse clinical populations. We therefore explored pathogen and host factors potentially impacting LAM
detection.
Methods: LAM-ELISA assay testing, sputum smear and culture status, HIV status, CD4 cell count, proteinuria and TB
outcomes were prospectively determined in adults diagnosed with TB and commencing TB treatment at a South
African township TB clinic. Sputum TB isolates were characterised by IS61110-based restriction fragment length
polymorphism (RFLP) and urines were tested for mycobacteriuria by Xpert®MTB/RIF assay.
Results: 32/199 (16.1%) of patients tested LAM-ELISA positive. Median optical density and proportion testing LAM
positive remained unchanged during 2 weeks of treatment and then declined over 24 weeks. LAM was associated
with positive sputum smear and culture status, HIV infection and low CD4 cell counts but not proteinuria, RFLP
strain or TB treatment outcome. The sensitivity of LAM for TB in HIV-infected patients with CD4 counts of ≥ 200,
100-199, 50-99, and < 50 cells/μl, was 15.2%, 32%, 42.9%, and 69.2% respectively. Mycobacteriuria was found in 15/
32 (46.9%) of LAM positive patients and in none of the LAM negative controls.
Conclusions: Urinary LAM was related to host immune factors, was unrelated to Mtb strain and declined steadily
after an initial 2 weeks of TB treatment. The strong association of urine LAM with mycobacteriuria is a new finding,
indicating frequent TB involvement of the renal tract in advanced HIV infection.
Background
Lipoarabinomannan (LAM), a major lipopolysaccharide
component of the cell wall of the genus Mycobacterium
and related actinomyces, was first characterised in the
1980’s [1]. LAM is present at the cell surface where it
can readily interact with host receptors and act as an
immunomodulator [2,3]. LAM is also highly immuno-
genic and anti-LAM antibodies are produced during
mycobacterial infection [4]. The detection of anti-LAM
antibodies has been proposed for diagnosis of active
tuberculosis [5,6]. Both LAM antigen and anti-LAM
antibody may be found aggregated in circulating anti-
body-antigen immune complexes [6,7].
LAM antigen is a 19,000 (± 8,500) daltons sized lipo-
polysaccharide which can be recovered in large quanti-
ties from Mycobacterium tuberculosis (Mtb) cultures [1],
and is detectable in serum [8], sputum [9,10] and urine
in a wide variety of tuberculosis (TB) clinical settings
[10-19]. Urine LAM testing has shown markedly vari-
able diagnostic accuracy for TB in field studies with a
generally low sensitivity [20,21]. However, sensitivity of
the assay has been reported to be increased in HIV-TB
co-infected patients with advanced immune suppression
[17,19] and also in those with high TB bacillary burden
[17,18]. A simple, low-cost, point-of care version of this
assay has been shown to have considerable utility when
* Correspondence: Kimberly.Racow@hiv-research.org.za
1Desmond Tutu HIV Centre, Institute of Infectious Diseases and Molecular
Medicine, University of Cape Town Faculty of Health Sciences, Cape Town,
South Africa
Full list of author information is available at the end of the article
Wood et al. BMC Infectious Diseases 2012, 12:47
http://www.biomedcentral.com/1471-2334/12/47
© 2012 Wood et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons
Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in
any medium, provided the original work is properly cited.

Page 2

screening for TB prior to antiretroviral therapy [22]. In
addition to host factors which may affect LAM detection
such as immune status, the quantitative expression of
LAM on the mycobacterial surface has also been shown
to be strain dependent [23,24].
It has been postulated that LAM is released from
metabolically active or degrading mycobacterial organ-
isms into the serum, with subsequent filtration by the
kidneys, passing into the urine where it can be detected
by enzyme-linked immunosorbent assay (ELISA) [14].
The molecular size of LAM is similar to myoglobin
(16,700 daltons), the primary oxygen carrying hemopro-
tein in striated muscle [25], which readily passes
through the normal glomerulus into urine following
muscle injury [25]. However, in contrast to myoglobin,
LAM is a highly immunogenic molecule frequently asso-
ciated with anti-LAM antibodies readily detectable in
serum [4-6]. Systemically released LAM may therefore
circulate in large immune complexes [26], which would
not be able to pass through normal renal glomeruli to
the urine [27]. In contrast, free or antibody-complexed
LAM released from mycobacteria within the renal tract
could pass directly into urine without the need to pass
though the glomerular membrane.
We therefore hypothesised that the variable sensitivity
of the urine LAM assay for TB diagnosis may be deter-
mined by a variety of pathogen and host factors. We
first explored temporal changes in urine LAM during
TB therapy when increased mycobacterial cell killing
[28] would be expected to increase LAM release. We
then explored the relationship between urine LAM and
host factors, including HIV status, level of immune sup-
pression, renal proteinuria and response to treatment.
To explore relationships between LAM and mycobacter-
ial factors we compared sputum TB strain patterns and
the presence of mycobacteriuria in LAM positive
patients and LAM negative patients.
Methods
Study site and population
The study population consisted of sequential adult TB
patients presenting between September 2009 and April
2011 to the TB clinic of a peri-urban township which
has been described elsewhere [29]. All patients were
diagnosed with TB, notified to the South African
National TB Control Programme (NTBCP) and com-
menced on standard rifampicin based short course TB
treatment administered under directly observed supervi-
sion [30]. Routine TB diagnostic and treatment proce-
dures were performed as outlined by the NTBCP [28].
Additional TB and HIV information was collected pro-
spectively in case report forms and by review of the
local TB and HIV registers. Written informed consent
was obtained from all patients and the study was
approved by the Research Ethics Committee of the
Faculty of Health Sciences of the University of Cape
Town. No previously reported studies have examined
the kinetics of urine LAM during TB treatment. We
chose a convenience sample size of 200 which was
based primarily on the average number of TB patients
presenting at Masiphumelele Clinic per year and also to
ensure a diverse study population in terms of sputum
and HIV status.
Sample collection and storage
Patients were requested to provide a urine sample at
each routine visit to the TB clinic every day during
week 1 and on week 2, week 8, week 16, and week 24 of
TB treatment. Urine samples were provided in separate
sterile containers, aliqotted into 5-7 ml containers, and
refrigerated at -200C for later analysis for LAM-ELISA,
urinary protein:creatinine ratio (P:C ratios) and Xpert®
MTB/RIF [31]. The aforementioned analyses were per-
formed in accredited laboratories of the South African
National Health Laboratory Service (NHLS). Sputum
samples collected for direct sputum smears during the
course of routine management were additionally
requested to be cultured for Mtb at an accredited TB
laboratory of the NHLS.
Urinary LAM ELISA analysis
LAM ELISA testing was performed strictly according to
manufacturer’s instructions (Clearview®TB ELISA,
Alere Health Services, USA). Assays were performed in
duplicate together with negative controls. Samples were
considered positive if the mean optical density (OD)
value minus the control OD value was greater than or
equal to 0.1 and negative if the mean OD value minus
the control OD value was less than 0.1. Patients were
categorised as LAM positive if their initial pre-treatment
(day 1) urine sample was positive as per conventional
screening procedure [10-19]. During the study we col-
lected an additional 10 urine samples at specific time
points throughout TB treatment, which increased the
probability of acquiring false positive results. We there-
fore adjusted our case definition to include patients who
tested LAM positive at day 1 and those who tested
LAM negative at day 1 but subsequently tested LAM
positive on two or more occasions during TB treatment.
Urinary protein and creatinine analysis
The P:C ratio (units = g/mmol) on a single urine speci-
men provides an estimate of approximate daily total
protein excretion [32]. LAM positive and an equal num-
ber of LAM negative control urines were analysed for
protein (g) and creatinine (mmol) to estimate P:C ratios
on day 1 and week 24 specimens. LAM negative con-
trols were chosen from HIV positive patients with
Wood et al. BMC Infectious Diseases 2012, 12:47
http://www.biomedcentral.com/1471-2334/12/47
Page 2 of 11

Page 3

laboratory confirmed TB using a random number table.
If day 1 urine was not available, the earliest available
specimen from the first week was used for analysis.
Missing week 24 urine specimens were not replaced.
Restriction fragment length polymorphism (RFLP) analysis
Pre-treatment sputum isolates that tested positive for
Mtb by culture were inoculated in duplicate into 7H9
liquid medium supplemented with oleic acid, albumin,
dextrose, and catalase (OADC) and 15% glycerol and
then stored at -70°C. Frozen duplicate culture stock was
shipped to the Public Health Research Institute (PHRI)
Tuberculosis Center at University of Medicine and Den-
tistry of New Jersey (UMDNJ). Culture stocks were sub-
cultured on Lowenstein-Jensen slants, and DNA was
extracted from each isolate. IS6110-based RFLP analysis
was performed as described elsewhere [33]. RFLP pat-
terns were analyzed using Bio Image pattern matching
software (BioImage Systems Inc., Jackson, Michigan,
USA.). IS6110-based RFLP derived DNA fingerprints
were assigned a strain code following a nomenclature
system that has been described elsewhere [34]. TB RFLP
strain patterns from LAM positive and LAM negative
patients were compared with each other and with pre-
viously described strains from this community [35,36].
Urine Xpert®MTB/RIF analysis
Single pre-treatment aliquots of urine from patients who
tested LAM positive and an equal number of randomly
matched LAM negative controls, as described above,
were analysed with the Xpert®MTB/RIF assay. In order
to further increase the yield of the Xpert®MTB/RIF
assay for patients testing LAM positive but Xpert®
MTB/RIF negative, urine collected from LAM positive,
Xpert negative patients during the first week of TB
treatment were pooled and re-tested. Each patient pro-
vided up to 5 daily urine samples of 4 ml each, which
were thawed, pooled and centrifuged. The resulting
supernatant was decanted and resuspended in a phos-
phate buffer to produce a 0.75 ml sample for Xpert
testing.
Statistical analysis
Data were analyzed using STATA 11.0 (StataCorp, Col-
lege Station, Texas). Bivariate analyses employed Wil-
coxon sum rank, chi-square and Fisher’s exact tests, as
appropriate. Logistic regression models were developed
to examine the association between both LAM and
Xpert®positivity and CD4 strata. Mixed effect, random
intercept models were developed to examine changes in
LAM OD over time. Mean changes in proportion
of patients with positive LAM results and the signifi-
cance of trends pre and post week 2 of TB treatment
were assessed using linear regression models. An
autoregressive model was used to account for the auto-
correlation of the data, and the impact of time was
examined through an interaction term. All statistical
tests are 2-sided at a = 0.05.
Results
Study cohort
Two hundred consecutive adults were diagnosed and
started on TB therapy in the clinic, and invited to parti-
cipate in the study between September 2009 and April
2011. One subject was found to be under study entry
age (17 years old) and the 199 met entry criteria and
were recruited to the study. The TB diagnostic group-
ing, HIV status and urine LAM testing results are
shown in Figure 1. 92/199 (46.2%) of patients were spu-
tum smear positive and of the total sputum smear posi-
tive patients, 85/92 (92.4%) were culture positive and 7/
92 (7.6%) were culture negative. 27/199 (13.6%) were
sputum smear negative and sputum Mtb culture positive
and 80/199 (40.2%) were presumptive TB cases (sputum
smear and culture negative). HIV sero-status was posi-
tive in 146/199 (73.4%), negative in 52/199 (26.1%) and
1/199 (0.5%) refused HIV testing. LAM was positive in
32/199 (16.1%) of patients, and of those, 29/119 (24.4%)
had laboratory confirmed TB, 25/92 (27.2%) were spu-
tum smear positive and 24/62 (38.7%) were HIV positive
smear positive.
The overall sensitivity of the LAM-ELISA assay in
sputum smear or culture confirmed cases was 24.4%. Of
the 146 HIV positive patients, the sensitivity was 69.2%
in those with a CD4 count < 50 cells/μl, 42.9% in those
with a CD4 count 50-99 cells/μl, 32% in those with a
CD4 count 100-199 cells/μl, and 15.2% in those with a
CD4 count of ≥ 200 cells/μl.
LAM positive cohort
32/199 (16.1%) of patients were classified as LAM posi-
tive; 27 patients tested positive on day 1 and a further 5
patients, who were LAM negative on day 1, tested LAM
positive on multiple occasions during TB treatment. Of
the latter group, 2 patients had 2 positive LAM results
(median LAM-ELISA OD = 0.195 and 0.228), 1 had 3
positive LAM results (median LAM-ELISA OD = 0.114),
1 had 4 positive LAM results (median LAM-ELISA OD
= 0.169) and 1 had 5 positive LAM results (median
LAM-ELISA OD = 0.44). 15 patients with single LAM
positive urines were not classified as LAM positive.
LAM analysis during TB treatment
The number of urine samples collected from LAM posi-
tive patients, the proportion of samples testing positive
and the median LAM-ELISA OD value at each time
point during TB treatment are shown in Figure 2. OD
measures decreased over the period of observation with
Wood et al. BMC Infectious Diseases 2012, 12:47
http://www.biomedcentral.com/1471-2334/12/47
Page 3 of 11

Page 4

a 0.004 unit reduction in OD per day (p < 0.001). How-
ever, OD measures remained stable during the first 2
weeks (p = 0.93), dropped significantly between week 2
and week 8 (p = 0.005), and although further decline
continued from week 8 to week 24, this did not reach
statistical significance (p = 0.36). Similarly, autoregres-
sion models showed that the proportion of all LAM
positive patients who were LAM positive at each time
point, remained stable in the first 2 weeks (p = 0.99),
and decreased from week 8 to week 24, although this
trend did not reach statistical significance (p = 0.77).
LAM and host risk factors
Host factors including age, gender, patient category, site
of TB disease, HIV status, CD4 cell count and antiretro-
viral status at baseline are presented for LAM positive
and LAM negative patients in Table 1. Being LAM posi-
tive was not statistically related to age, gender, patient
category, or routine designation of pulmonary and
extra-pulmonary TB. However, LAM positivity was asso-
ciated with laboratory confirmation of TB (p < 0.001)
and being HIV positive (p = 0.002). Amongst HIV
positive patients, LAM positivity was associated with
lower median CD4 cell count (p = 0.006) but not with
antiretroviral use (p = 0.078).
The median values with interquartile ranges (IQR) of
P:C ratios of 32 LAM positive and 32 LAM negative
patients at day 1 and 24 LAM positive and 26 LAM
negative patients at week 24 are shown in Figure 3. P:C
ratios were higher on day 1 for LAM positive patients
compared with LAM negative patients (p < 0.001) but
similar for both groups at week 24 (p = 0.8). Elevated
proteinuria (> 0.150 g/24 hours) was noted in 47
patients (28 LAM positive, 19 LAM negative) on day 1
and 15 patients (8 LAM positive, 7 LAM negative) at
week 24. Moderate proteinuria (> 1 g/24 hours) was
present in 4 patients (3 LAM positive, 1 LAM negative)
at day 1 and 1 patient (LAM negative) at week 24.
Furthermore, 33/167 (19.8%) of LAM negative patients
and 12/32 (37.5%) of LAM positive patients had a smear
positive grading of +++, 9/167 (5.4%) and 5/32 (15.6%)
were smear positive ++, 15/167 (9%) and 6/32 (18.8%)
were smear positive +, 7/167 (4.2%) and 1/32 (3.1%)
were smear positive scanty, and 103/167 (61.68%) and
27 smear-ve/culture+ve
23 HIV+, 4 HIV-
92 smear+ve
62 HIV+, 30 HIV-
80 smear-ve/culture-ve*
61 HIV+, 18 HIV-, 1 unknown
25 LAM+ve
24 HIV+,1 HIV-
200 TB notifications
3 LAM+ve
1 HIV+, 1 HIV-,
1 unknown
23 LAM-ve
19 HIV+,4 HIV-
4 LAM+ve
4HIV+
67 LAM-ve
38HIV+,29HIV-
77 LAM-ve*
60HIV+,17HIV-
1 exclusion 17 years old
*1 subject did not provide a sputum sample
199 enrolled
146 HIV+, 52 HIV-, 1unknown
Figure 1 LAM Study Consort Diagram. Disposition by HIV and TB sputum bacteriological status of 200 consecutive patients presenting to a
South African township TB clinic and undergoing LAM-ELISA urine testing
Wood et al. BMC Infectious Diseases 2012, 12:47
http://www.biomedcentral.com/1471-2334/12/47
Page 4 of 11

Page 5

8/32 (25%) were smear negative. Overall, LAM positive
patients had a higher smear grading compared to LAM
negative patients (p = 0.001).
LAM and TB treatment outcomes
Treatment outcomes were similar in LAM positive and
LAM negative patients. Deaths 2/32 (6.3%) and 8/167
(4.8%), treatment completion or cure 24/32 (75%) and
130/167 (77.8%), default/failure/transfer out 4/32
(12.5%) and 21/167 (12.6%), and miscellaneous 2/32
(6.3%) and 8/167 (4.8%) occurred in the LAM positive
and LAM negative patients respectively.
RFLP analysis
65 Mtb isolates were successfully cultured and under-
went RFLP analysis, identifying 42/65 (64.6%) different
unique strain codes. There was considerable genetic
diversity between strains with 6/9 (66.6%) discrete
synonymous single-nucleotide polymorphism-based
(sSNP) phylogenetic lineages represented [37]. Identified
strains were classified as belonging to CC 14/42 (33.3%),
W-Beijing 6/42 (14.3%) and 22/42 (52.4%) a variety of
other genotype families. 6/42 (14.3%) of strains were
common to both LAM positive and LAM negative
patients, 13/42 (31%) of unique strains were from LAM
positive patients and 23/42 (54.8%) of unique strains
were from LAM negative patients.
21/42 (50%) of strains had been previously recognised
as circulating in this community between 2001 and 2006
[35] or identified in a more recent survey [36], while 12/
42 (28.6%) of LAM positive and 9/42 (21.4%) of LAM
negative strains were identified for the first time. We
therefore found no evidence of LAM association with
specific TB strains, TB strain families or sSNP phyloge-
netic lineages.
Urine xpert®MTB/RIF assay
10/32 (31.3%) day-1 urine samples and a further 5/22
(22.7%) of pooled urine samples tested Xpert®MTB/RIF
assay positive. All 15/15 (100%) Xpert®positive urines
0
0.1
0.2
0.3
0.4
0.5
0
0.2
0.4
0.6
0.8
1
D1D2D3 D4 D5
Time on TB treament
D6 D7Wk2 Wk8Wk16Wk24
% LAM positive
% LAM+
Median OD
LAM Pos Tests 2717 1215151723231161
Total Tests3222171817 2127 2929 27 24
Median OD0.390.19 0.250.340.220.32 0.290.35 0.040.010.00
IQR+0.740.480.630.550.570.650.990.840.240.060.01
IQR-
Figure 2 Proportion of LAM positive patients testing LAM positive at days 1-7, week 2, week 8, week 16 and week 24 and median
LAM-ELISA optical density at each time point. Proportions are represented as bars (values shown on left-hand vertical axis) and median
optical density is represented as a solid line (values shown on right-hand vertical axis). The number of patients who tested at each time point,
median OD and IQR are shown in corresponding table below the figure
0.140.11 0.070.160.160.120.120.160.01-0.01-0.01
Median Optical Density
Wood et al. BMC Infectious Diseases 2012, 12:47
http://www.biomedcentral.com/1471-2334/12/47
Page 5 of 11

Page 6

were from LAM positive patients and none from LAM
negative patients. Therefore all Xpert®positive results
occurred only in LAM positive patients. Similarly, there
was a strong association between HIV and TB status;
14/15 (93.3%) Xpert®positive tests occurred in those
that were HIV positive with laboratory proven TB. One
Xpert®positive patient was HIV negative with presump-
tive TB. Of the 85 HIV positive patients with laboratory
proven TB patients, the proportion that were LAM posi-
tive and the proportion that were Xpert®positive are
shown stratified by CD4 cell count in Figure 4. The pro-
portions LAM positive were 9/13 (69.2%), 6/14 (42.9%),
8/25 (32%), 5/33 (15.2%) and Xpert®positive were 7/13
(53.8%), 3/14 (21.4%), 2/25 (8%), 2/33 (6.1%), in CD4
cell count strata < 50 cells/μl, 50-99 cells/μl, 100- 199
cells/μl, and ≥ 200 cells/μL respectively. The bivariate
logistic regression modelling showed that CD4 cell
count strata are a significant predictor of LAM status
with an Odds Ratio (OR) of 0.45 (95% CI: 0.28-0.72; p =
0.001). CD4 cell count was then stratified into 4 strata
and a multivariate logistic regression analysis was done.
With each increase in CD4 cell count strata, the OR of
being LAM positive decreased: 0.33 (95% CI:0.07-1.62; p
= 0.174), 0.21 (95% CI: 0.05-0.89; p = 0.034) and 0.08
(95% CI:0.02-0.36; p = 0.001) in CD4 cell count strata,
50-99 cells/μl, 100-199 cells/μl, and ≥ 200 cells/μL com-
pared to < 50 cells/μl stratum as reference stratum,
respectively. A similar trend between CD4 cell count
strata and Xpert®exists, however it did not reach sta-
tistical significance with an OR of 0.50 (95% CI: 0.23-
1.06; p = 0.07).
positive urine samples were 53.8%, 21.4%, 8%, 6.1% in
CD4 cell count strata < 50 cells/μl, 50-99 cells/μl, 100-
199 cells/μl, and ≥ 200 cells/μL respectively. The logistic
Table 1 Host factors and urine LAM Status of 199 study participants
Host FactorsTotal n = 199LAM Negative
n = 167 (83.9%)
LAM Positive*
n = 32 (16.1%)
p-value**
Median age years [IQR]
34 [28-43]35 [30-40]34 [28-44]0.929†
Gender
0.904††
Male 11092 (55.1) 18 (56.3)
Female 8975 (44.9)14 (43.8)
Patient TB category
0.968††
New TB case 125105 (62.9)20 (62.5)
Re-treatment TB case7462 37.1)12 (37.5)
Site of disease
0.613‡
Pulmonary168141 (84.4)27 (84.4)
Extra-Pulmonary1816 (9.6) 2 (6.3)
Pulmonary and Extra-Pulmonary 1310 (6)3 (9.4)
Smear or culture positive
< 0.001‡
Positive 11990 (53.9)29 (90.6)
Negative 7976 (45.5) 3 (9.4)
Not done11 (0.6)0 (0)
HIV status
0.002‡
Positive 146117 (70.1) 29 (90.6)
Negative5250 (29.9)2 (6.3)
Unknown10 (0)1 (3.1)
Median CD4 cells/μL. [IQR]a
132.5 [70-283]156 [80-309]95 [42-134]0.006†
< 502617 (14.5) 9 (31)
50-992822 (18.8)6 (20.7)
100-1993931 (26.5)8 (27.6)
≥ 2005347 (40.2)6 (20.7)
ART at start of TB treatmenta
0.078††
Yes3330 (25.6)3 (10.3)
No 11387 (74.4)26 (89.7)
*LAM positive patients are defined as those who were LAM positive on day 1 or on two or more occasions during TB treatment. A LAM result was considered
positive when the average sample LAM-ELISA optical density (OD) minus the negative control was greater than or equal to 0.1 and negative when the average
sample OD minus the negative control was less than 0.1. **p-values indicate differences between LAM positive and LAM negative participants. p-values were
calculated using Wilcoxon rank sum†, chi-square††, and Fisher’s exact test‡.aof the 146 HIV positive subjects
Wood et al. BMC Infectious Diseases 2012, 12:47
http://www.biomedcentral.com/1471-2334/12/47
Page 6 of 11

Page 7

regression models showed that CD4 cell count strata is
a significant predictor of LAM status (OR = 0.45; p =
0.001). With each increase in CD4 cell count strata, the
odds ratio (OR) of being LAM positive decreased. A
similar pattern was seen with Xpert®but did not reach
statistical significance (OR = 0.50; p = 0.07)
Discussion
In an adult South African TB clinic population approxi-
mately one quarter of patients with laboratory proven
TB were found to have detectable LAM in urine. We
found increased sensitivity of the LAM test in HIV
infected individuals with advanced immune suppression,
a finding which is consistent with previous studies
[10,17,18,21,22]. Mild to moderate proteinuria was
increased in both LAM positive and LAM negative HIV
positive patients early in TB therapy which diminished
with TB treatment. Heavy proteinuria sufficient to be
associated with leakage of large immunoglobulins and
immune-complex LAM was not recorded among LAM
0
.1
.2
.3
.4
.5
Day 1 Week 24
LAM NegativeLAM Positive
p=0.011** p=0.289**
P<0.001*
Day 1 Week 24
LAM negative LAM positive LAM negative LAM positive
No. of tests 32 32 26 24
Median P:C ratio 0.02 0.05 0.01 0.01
IQR+ 0.04 0.06 0.02 0.02
IQR-
Figure 3 Protein: Creatinine Ratios of 32 LAM positive patients and 32 randomly chosen LAM negative patients at Day 1 and Week 24
of TB treatment. There was an overall statistically significant difference in P:C ratios between day 1 and week 24 (p < 0.001, Wilcoxon signed
rank test*). There was a statistically significant difference in P:C ratios between LAM positive and LAM negative patients at day 1 (p = 0.011,
Wilcoxon rank sum test **) but not at week 24 (p = 0.289, Wilcoxon rank sum test **)
0.01 0.03 0.01 0.01
Wood et al. BMC Infectious Diseases 2012, 12:47
http://www.biomedcentral.com/1471-2334/12/47
Page 7 of 11

Page 8

positive patients [27]. The major new finding of our
study was the identification of Mtb organisms in the
urine of nearly half of patients who were urine LAM
positive and in none of the LAM negative controls. The
presence of urinary LAM was not limited to a single
infecting Mtb strain but was associated with an equally
wide variety of individual strain patterns and strain
families as found in LAM negative cases. Therefore, our
findings indicate that urine LAM may be more related
to the site of Mtb infection rather than the specific
infecting strain.
Our initial hypothesis that the quantity of urine LAM
as reflected by OD would be increased during early TB
treatment was not substantiated. However, sensitivity of
LAM testing was increased by inclusion of patients
whose urine test became positive during TB treatment.
It seems probable that the increased yield of LAM posi-
tive tests was due to increased sampling frequency
rather than to increased release of LAM from organisms
during therapy. However, the almost total clearance of
LAM from the urine after 24 weeks of TB treatment
likely reflects a measure of therapeutic response.
The Xpert®MTB/RIF assay isolates mycobacterial
organisms from clinical samples within the test cartridge
by filtration prior to ultrasonic lysis which releases Mtb
DNA for subsequent amplification [38]. We therefore
used the Xpert®assay as a measure of the presence of
whole Mtb organisms within urine but recognized that
0
0.2
0.4
0.6
0.8
1
<50 50-99 100-199 ≥200
Proportion test positive
CD4+ cell count (cells/uL)
LAM-ELISA
Xpert MTB/RIF
Logistic Regression model for LAM results by CD4 Categories (n=85)
CD4 Categories OR 95% CI P-values
<50 (reference)
50-99
100-199
≥200
1 NA NA
0.174
0.034
0.001
0.33
0.21
0.08
0.07-1.62
0.05-0.89
0.02-0.36
Figure 4 Proportions of urine samples from 85 HIV positive patients with laboratory confirmed TB who were LAM positive and Xpert®
MTB/RIF positive stratified into 4 CD4 cell strata: < 50 cells/μl, 50-99 cells/μl, 100-199 cells/μl, and ≥ 200 cells/μL. The proportions of
LAM positive urine samples were 69.2%, 42.9%, 32%, 15.2% and Xpert®
Wood et al. BMC Infectious Diseases 2012, 12:47
http://www.biomedcentral.com/1471-2334/12/47
Page 8 of 11

Page 9

the assay does not necessarily indicate the presence of
viable organisms [38]. We found that almost half of the
LAM positive patients tested Xpert®positive using rela-
tively small volumes of urine and therefore the propor-
tion of Xpert positive results may have been
underestimated. Larger urine sample volumes might
have increased yield and allowed for Mtb culture which
could have elucidated Mtb viability.
Three proposed models of the possible fate of LAM
molecules released from Mtb organisms in different
body compartments are illustrated in Figure 5. Model
5A shows LAM released from organisms in the systemic
(non-renal) compartment into the circulation where
they bind to specific anti-LAM antibodies to form large
immune complexes with limited capacity to pass
through the normal kidney to the urine. In patients with
normal renal filtration this model would be expected to
give rise to a negative urine LAM test in the absence of
mycobacteriuria. Model 5B reflects the currently
accepted model of urinary LAM [14] in which free LAM
molecules released from Mtb organisms in the systemic
compartment into the circulation are not antibody
bound and are therefore able to pass freely through the
normal kidney into the urine. This model gives rise to a
positive urine LAM test in the absence of Mtb organ-
isms. Model 5 C is compatible with the findings of this
study in which urine LAM is released directly from Mtb
organisms within the renal tract compartment, which
gives rise to a positive LAM test in the presence of Mtb
organisms in urine. Mycobacteriuria presents evidence
of extra-pulmonary TB with renal tract involvement. To
determine the relative contribution of model 5 C to
???
??
??
??
??
??????Mtb ?????????? ??????Mtb????????? ???????????????????????????????? ????????????????????????????????? ??????????????????????????????????????????????????????????????????
??????????????????????????????????????????????????????????? ??????????????????????????????????????????????????????????????????????????????????????????????????????????????????
????????
???????
??????
????????? ???? ?? ?????????
???????????????????????????????????? ???????
????????
???????
??????
???????
???????
???????? ?
????????????????????????????????????????????????????
Mtb
Mtb
Mtb
Mtb
Mtb
Mtb
????????????????
????????????????
????
??????????
?
??
??????????????????????
???????????????????

Figure 5 Three proposed models of the possible fate of LAM molecules released from systemic or urinary Mtb organisms. A.
Systemically released LAM is bound to an antibody to form an immune complex within the circulation that impedes renal filtration of LAM
across the intact glomerular membrane. This model gives rise to a negative LAM test in the absence of Mtb. B. Circulating LAM unattached to a
specific anti-LAM antibody is freely filtered through the kidney into the urine, which gives rise to a positive LAM test in the absence of Mtb. C.
Mtb within in the renal tract releases LAM directly into urine, which gives rise to a positive LAM test in the presence of Mtb.
Wood et al. BMC Infectious Diseases 2012, 12:47
http://www.biomedcentral.com/1471-2334/12/47
Page 9 of 11

Page 10

urine LAM, further studies will be required in which
larger quantities of urine are processed to increase sam-
pling probabilities.
We were not able to show whether detected urinary
LAM existed as free LAM molecules, was complexed
with anti-LAM antibodies, or was physically associated
with mycobacterial organisms. However, a strong asso-
ciation between urine LAM detection and mycobactere-
mia has been previously reported [18] which may
indicate a likely source for mycobacteriuria.
Although LAM ELISA had moderate sensitivity in our
study overall, it is important to highlight the increased
sensitivity and utility of this diagnostic amongst those
with low CD4 count [39]. Screening for TB among
patients with advanced immunodeficiency prior to antire-
troviral therapy is challenging and rapidity is critical in
view of high mortality risk [39]. However, a simple, low-
cost, point-of-care version of the assay for detection of
LAM (Determine TB-LAM) has now been shown to have
considerable utility as a screening tool in this setting [22].
Conclusions
In summary, this study determined factors potentially
impacting the sensitivity of urine LAM for TB diagnosis
in a South African population routinely attending a TB
clinic. Sensitivity of LAM testing was independent of
Mtb strain but was associated with advanced HIV infec-
tion and laboratory confirmed TB disease. However, the
finding of urine LAM did not appear to indicate worse
TB treatment outcomes. Qualitative and quantitative
estimations of urine LAM may have future utility as bio-
markers reflecting response to TB treatment. The isola-
tion of Mtb from a large proportion of LAM positive
cases indicates that renal TB may occur more com-
monly in advanced HIV than previously recognised. The
LAM assay is a promising diagnostic for use in HIV
positive patients with low CD4 counts and a new point-
of-care version of this assay now enables rapid TB diag-
nosis in this patient population.
Acknowledgements
RW was funded by National Institutes of Health (NIH) [RO1 A1058736-01A1
and 5UO1A1069519-02] and United States Agency for International
Development (USAID) [3UO1A1069924-O2S]. LGB was funded by grant
UO1A1069519. SDL was funded by Wellcome Trust, London, UK. We are
grateful to the National Institutes of Health (NIH) and United States Agency
for International Development (USAID) and the Wellcome Trust for funding
this work, the Foundation for Innovative New Diagnostics (FIND) for
providing access to the Xpert®MTB/RIF assay cartridges with preferential
pricing, and Alere Inc., USA, for providing Clearview LAM-ELISA tests for this
study. We are grateful to Elaine Sebastian and the staff at Masiphumelele
clinic for their help with conducting and overseeing the study.
Author details
1Desmond Tutu HIV Centre, Institute of Infectious Diseases and Molecular
Medicine, University of Cape Town Faculty of Health Sciences, Cape Town,
South Africa.2Department of Medicine, University of Cape Town Faculty of
Health Sciences, Cape Town, South Africa.3Department of Science and
Technology/National Research, Foundation Centre of Excellence in
Epidemiological Modelling and Analysis, University of Stellenbosch,
Stellenbosch, South Africa.4Public Health Research Institute TB Center,
University of Medicine and Dentistry of New Jersey, Newark, NJ, USA.
5Department of Clinical Research, Faculty of Infectious and Tropical Diseases,
London School of Hygiene and Tropical Medicine, London, UK.
Authors’ contributions
All authors made substantive intellectual contributions to this study. RW, KM,
LGB, SDL contributed significantly to the study conception and design. RW,
KR, KM, SDL contributed significantly to the data analysis and interpretation.
KR supervised data collection and management. MV supervised specimen
collected, analysis, and management. BNK supervised RFLP testing. All
authors contributed significantly to the manuscript draft.
Competing interests
No interpretations of data or presentation of information of this study was
influenced by any of the author’s personal or financial relationship with
other people or organizations. None of the authors have any financial or
non-financial competing interests.
Received: 19 September 2011 Accepted: 27 February 2012
Published: 27 February 2012
References
1.Hunter SW, Gaylord H, Brennan PJ: Structure and antigenicity of the
phosphorylated lipopolysaccharide antigens from the leprosy and
tubercle bacilli. J Biol Chem 1986, 261(26):12345-12351.
2. Pitarque S, Larrouy-Maumus G, Payré B, Jackson M, Puzo G, Nigou J: The
immunomodulatory lipoglycans, lipoarabinomannan and lipomannan,
are exposed at the mycobacterial cell surface. Tuberculosis (Edinb) 2008,
88(6):560-565.
3.Mishra AK, Driessen NN, Appelmelk BJ, Besra GS: Lipoarabinomannan and
related glycol-conjugates: structure, biogenesis and role in
Mycobacterium tuberculosis physiology and host-pathogen interaction.
FEMS Microbiol Rev 2011, doi:10.1111/j.15746976.2011.00276.x.
4.Murase T, Zheng RB, Joe M, Bai Y, Marcus SL, Lowary TL, Ng KKS: Structural
insights into antibody recognition of mycobacterial polysaccharides. J
Mol Biol 2009, 392:381-392.
5.Eleftheriadis T, Tsiaga P, Antoniadi G, Liakopoulos V, Kortsaris A, Giannatos E,
Barbutis K, Stefanidis I, Vargemezis V: The value of serum
antilipoarabinomannan antibody detection in the diagnosis of latent
tuberculosis in hemodialysis patients. Am J Kidney Dis 2005, 46(4):706-712.
6.Tessema TA, Bjune G, Hamasur B, Svenson S, Syre H, Bjorvatn B: Circulating
antibodies to lipoarabinomannan in relation to sputum microscopy,
clinical features and urinary anti- lipoarabinomannan detection in
pulmonary tuberculosis. Scand J Infect Dis 2002, 34(2):97-103.
7.Ludwiczak P, Gilleron M, Bordat Y, Martin C, Gicquel B, Puzo G:
Mycobacterium tuberculosis pho mutant lipoarabinomannan molecular
structure. Microbiology 2002, 148(10):3029-3037.
8.Sada E, Aguilar D, Torres M, Herrera T: Detection of lipoarabinomannan as
a diagnostic test for tuberculosis. J Clin Micro 1992, 30(9):2415-2418.
9.Pereira Arias-Bouda LM, Nguyen LN, Ho LM, Kuijper S, Jansen HM, Kolk AH:
Development of antigen detection assay for diagnosis of tuberculosis
using sputum samples. J Clin Micro 2000, 38:2278-2283.
10.Dheda K, Davids V, Lenders L, et al: Clinical utility of a commercial LAM-
ELISA assay for TB diagnosis in HIV-infected patients using urine and
sputum samples. PLoS One 2010, 5(3):e9848.
11.Hamasur B, Bruchfeld J, Haile M, Pawlowski A, Bjorvatn B, Kallenius G, et al:
Rapid diagnosis of tuberculosis by detection of mycobacterial
lipoarabinomannan in urine. J Microbiol Methods 2001, 45:41-52.
12.Tessema TA, Hamasur B, Bjun G, Svenson S, Bjorvatn B: Diagnostic
evaluation of urinary lipoarabinomannan at an Ethiopian tuberculosis
centre. Scand J Infect Dis 2001, 33:279-284.
13.Tessema TA, Bjune G, Assefa G, Svenson S, Hamasur B, Bjorvatn B: Clinical
and radiological features in relation to urinary excretion of
lipoarabinomannan in Ethiopian tuberculosis patients. Scand J Infect Dis
2002, 34:167-171.
14.Boehme C, Molokova E, Minja F, Geis S, Loscher T, Maboko L, et al:
Detection of mycobacterial lipoarabinomannan with an antigen-capture
Wood et al. BMC Infectious Diseases 2012, 12:47
http://www.biomedcentral.com/1471-2334/12/47
Page 10 of 11

Page 11

ELISA in unprocessed urine of Tanzanian patients with suspected
tuberculosis. Trans R Soc Trop Med Hyg 2005, 99:893-900.
Daley P, Michael JS, Hmar P, et al: Blinded evaluation of commercial
urinary lipoarabinomannan for active tuberculosis: a pilot study. Int J
Tuberc Lung Dis 2009, 13:989-995.
Mutetwa R, Boehme C, Dimairo M, et al: Diagnostic accuracy of
commercial urinary lipoarabinomannan detection in African tuberculosis
suspects and patients. Int J Tuberc Lung Dis 2009, 13:1253-1259.
Lawn SD, Edwards DJ, Kranzer K, Vogt M, Bekker L-G, Wood R: Urine
lipoarabinomannan assay for tuberculosis screening before antiretroviral
therapy: diagnostic yield and association with immune reconstitution
disease. AIDS 2009, 23:1875-1880.
Shah M, Martinson NA, Chaisson RE, Martin DJ, Variava ME, Dorman S:
Quantitative analysis of a urine-based assay for detection of
lipoarabinomannan in patients with tuberculosis. J Clin Microbiol 2010,
48:2972-2974.
Shah M, Variava E, Holmes CB, Coppin A, Golub JE, McCallum J, Wong M,
Luke B, Martin DJ, Chaisson RE, Dorman SE, Martinson NA: Diagnostic
accuracy of urine lipoaribinomannan test for tuberculosis in hospitalised
patients in a high prevalence setting. J Acquir Immune Defic Synd 2009,
52(2):145-51.
Pai M, Minion J, Sohn H, Zwerling A, Perkins MD: Novel and improved
technologies for tuberculosis diagnosis: progress and challenges. Clin
Chest Med 2009, 30:701-716.
Minion J, Leung E, Talbot E, Dheda K, Pai M, Menzies D: Diagnosing
tuberculosis with urine lipoarabinomannan: systematic review and meta-
analysis. Eur Respir J 2011, doi:10.1183/09031936.00025711.
Lawn SD, Kerkhoff AD, Vogt M, Wood R: Screening for HIV-associated
tuberculosis prior to antiretroviral therapy: diagnostic accuracy of a low-
cost, urine antigen, point-of-care assay (Determine TB-LAM). Lancet Infect
Dis 2011.
Schwebach JR, Casadevall A, Schneerson R, Dai Z, Wang X, Robbins JB,
Glatman-Freedman A: Expression of a Mycobacterium tuberculosis
arabinomannan antigen in vitro and in vivo. Infect Immun 2001,
69:5671-5678.
Neyrolles O, Guilhot C: Recent advances in deciphering the contribution
of Mycobacterium tuberculosis lipids to pathogenesis. Tuberculosis (Edinb)
2011, 91(3):187-95.
Ordway GA, Garyry DJ: Myoglobin: an essential hemoprotein in striated
muscle. J Exp Med 2004, 207(20):3441-6.
Roberts-Thomson PJ, Shepherd K: Molecular size and heterogeneity of
immunoglubulins in health and disease. Clin exp Immunol 1990,
79:328-334.
Harraldson B, Nystrom J, Deen WM: Properties of the glomerular barrier
and mechanisms of proteinuria. Physiol Rev 2008, 88:451-487.
Donald PR, Diacon AH: The early bactericidal activity of anti-tuberculosis
drugs: a literature review. Tuberculosis (Edinb) 2008, 88(Suppl 1):S75-83.
Middelkoop K, Bekker L-G, Myer L, Johnson LF, Kloos M, Morrow C, Wood R:
Antiretroviral therapy and TB notification rates in a high HIV prevalence
South African community. J Acquir Immune Defic Synd 2011, 53(3):263-269.
The South African Tuberculosis Control Programme Practical Guidelines:
South African Department of Health; 2000., [http://www.capegateway.gov.
za/Text/2003/tb_guidelines2000.pdf] Accessed 23 July 2011.
Hilleman D, Ruesch-Gerdes S, Boehme C, Richter E: Rapid molecular
detection of extrapulmonary tuberculosis by automated GeneXpert®
MTB/RIF system. J Clin Micro 2011, 49(4):1202-1205.
Gaspari F, Perico N, Remuzzi G: Timed urine collections are not needed to
measure urine protein excretion in clinical practice. Am J Kidney Dis 2006,
47(1):1-7.
van Embden JD, Cave MD, Crawford JT, et al: Strain identification of
Mycobacterium tuberculosis by DNA fingerprinting: recommendations
for a standardized methodology. J Clin Microbiol 1993, 31:406-409.
Bifani PJ, Mathema B, Liu Z, et al: Identification of a W variant out- break
of Mycobacterium tuberculosis via population-based molecular
epidemiology. JAMA 1999, 282:2321-2327.
Middelkoop K, Bekker L-G, Mathema B, Shashkina E, Kurepina N,
Whitelaw A, Morrow C, Kreiswirth B, Kaplan G, Wood R: Molecular
epidemiology of Mycobacterium tuberculosis in a South African
community with high HIV prevalence. J Infect Dis 2009, 200(8):1207-1211.
Middelkoop K, Mathema B, Myer L, Shashkina E, Kreisworth B, Kurepina N,
Wood R, Bekker L-G: Transmission of TB in a High HIV-Prevalent South African
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
Community [TUPE163] Rome: Abstracts of the 6th IAS Conference on HIV
Pathogenesis, Treatment and Prevention; 2011.
Mathema B, Kurepina NE, Bifani PJ, Kreiswirth BN: Molecular epidemiology
of tuberculosis: current insights. Clin Microbiol Rev 2006, 19:658-685.
Lawn SD, Nicol MP: Xpert MTB/RIF assay: development, evaluation and
implementation of a new rapid molecular diagnostic for tuberculosis
and rifampicin resistance. Future Micro 2011.
Lawn SD, Wood R: Tuberculosis in antiretroviral treatment services in
resource-limited settings: addressing the challenges of screening and
diagnosis. J Infect Dis 2011.
37.
38.
39.
Pre-publication history
The pre-publication history for this paper can be accessed here:
http://www.biomedcentral.com/1471-2334/12/47/prepub
doi:10.1186/1471-2334-12-47
Cite this article as: Wood et al.: Lipoarabinomannan in urine during
tuberculosis treatment: association with host and pathogen factors and
mycobacteriuria. BMC Infectious Diseases 2012 12:47.
Submit your next manuscript to BioMed Central
and take full advantage of:
• Convenient online submission
• Thorough peer review
• No space constraints or color figure charges
• Immediate publication on acceptance
• Inclusion in PubMed, CAS, Scopus and Google Scholar
• Research which is freely available for redistribution
Submit your manuscript at
www.biomedcentral.com/submit
Wood et al. BMC Infectious Diseases 2012, 12:47
http://www.biomedcentral.com/1471-2334/12/47
Page 11 of 11