Urinary lipoarabinomannan in individuals with sputum-negative pulmonary tuberculosis

Tuberculosis (TB) is a disease known to cause major health disturbances worldwide and is also a prime infectious agent of mortality. Across the globe about 10 million people were affected with TB and 1.4 million succumbed in 20191. The clinical diagnosis of pulmonary TB (PTB) is based on the symptoms and chest X-ray. Microbiologic confirmation of PTB is done by using diagnostic tools such as Ziehl–Neelsen (ZN) sputum microscopy, molecular tests for TB and culture for Mycobacterium tuberculosis (M.tb). The gold standard for PTB diagnosis is sputum culture. However, due to the lack of access to culture facilities and its long turn-around time most TB control/elimination programmes worldwide employ Z-N sputum microscopy for detecting of acid-fast bacilli in sputum smears as their main diagnostic tool. Sputum smear microscopy being feasible and inexpensive in most field conditions effectively identifies the most infectious of PTB cases2. However, this technique has lower sensitivity to detect smear-negative culture-positive PTB. The direct sputum smear microscopy by Z-N staining has sensitivity as low as 22-43 per cent for a single smear3 and up to 60 per cent under best conditions in comparison to mycobacteriological cultures45. A wide variation in the sensitivity of sputum smear microscopy can be a major obstacle for early TB case detection. Accordingly, National TB Elimination Programme (NTEP) of India defines the use of diagnostic tests like cartridge-based nucleic acid amplification test (CBNAAT) and culture for M.tb for microbiological confirmation of TB in sputum smear-negative presumptive PTB cases. However, the diagnosis of these individuals is still challenging due to the paucibacillary nature of the disease and the low sensitivity of the routinely employed first-line screening procedures. The accuracy of CBNAAT (Xpert) is only one third to detect M.tb in smear-negative P-TB cases; however, the second CBNAAT (Xpert) test increased the sensitivity by 27.8 per cent in smear-negative disease, yielding upto 61 per cent4. A screening test that is accurate for detecting active smear-negative PTB is the need of the hour.

Lipoarabinomannan (LAM), the outer surface glycan, is a TB antigen that was identified in the urine from the individuals with active TB. Measurement of LAM by ELISA and lateral flow tests reportedly provide low sensitivity6. However, lateral flow (LF) urine LAM assay is helpful in the diagnosis and screening of active TB in people living with HIV (PLHIV) in whom CD4 count was found to be low. Those PLHIV who have TB have shown to have a higher morbidity and mortality, smear-negative PTB in specific. Delay in making diagnosis can be an important cause of excess morbidity and mortality7. At present, the WHO policy is not to use LF urinary LAM for the diagnosis of TB among HIV-negative individuals due to poor sensitivity of existing LAM assays (strong recommendation and low quality of evidence). There are three main hypotheses for poor sensitivity of available LAM assays. First, LAM is secreted into the urine of individuals with PTB only in the context of HIV-mediated nephropathy which leads to HIV-induced glomerular dysfunction. Previous clinical studies have found no association between urine-LAM detection and proteinuria (indicative of glomerular damage)89. Second, TB-infected individuals can release LAM into the urine, with active extrapulmonary involvement of the kidney and its allied system. The antigen directly enters the urine from the infected tissue.10 The final hypothesis explains that the amount of LAM in active PTB cases is low to be detected by existing assays. The formation of various immune complexes, non-LAM proteins or inhibitors present in the urine makes it difficult for the identification of LAM11. Urinary LAM detection by cage-based nanotechnology, a novel class of sandwich immunoassay using copper complex dye in a hydrogel nanocage is not affected by the immunologic or physiologic consequences of HIV infection for LAM detection in human urine912. Thus, with the help of urinary LAM analysis, with the chemical bait on one side of the sandwich, discriminated HIV-negative but active TB-positive individuals can be identified12. This method is not restricted to individuals with TB colonization of the kidney1012 and also, copper dye bait captures LAM effectively from any coupled protein that affects results. We sought to study the diagnostic yield of urinary LAM among HIV-negative individuals with clinically diagnosed smear-negative PTB cases.

Material & Methods

This case-control study was undertaken at the Department of TB and Chest, National Institute of Tuberculosis and Respiratory Diseases, New Delhi, between 2017 and 2019 after clearance from the Institutional Ethics Committee. Integrated counselling and testing centre services were offered to all the enrolled participants and a written informed consent was obtained from all before the start of the study. Three groups of study participants were included, 50 smear-negative clinically diagnosed PTB cases, 15 sputum smear-positive cases (positive control) and 15 post tubercular sequel cases (negative control) according to the NTEP guidelines. Microbiological confirmation of diagnosis of TB was made using TB diagnostic tests like CBNAAT and mycobacteria growth indicator tube (MGIT) culture. Line probe assay (LPA) was processed from culture isolates. Fifteen post TB-squeal (negative control) who had completed the anti-TB treatment of more than three months and M.tb was not detected in their sputum by TB diagnostic tests CBNAAT and MGIT culture. All the patients were more than 14 yr of age. Sputum examination for ZN stain microscopy, CBNAAT and MGIT culture was done as per the standard guidelines of NTEP of India. M. tuberculosis culture was performed by MGIT-960. Smear status of the patient population was known only to the primary investigator and the laboratory personnel processing the urine sample for LAM was blinded.

Urine lipoarabinomannan (LAM) test: Urine LAM test was done with cage nanotechnology-based sandwich ELISA using copper complex dye technique. Urine sample was collected in a sterile container and centrifuged for 20 min at 2000 rpm within 30 min of collection. Supernatant was carefully collected and stored at -20° to -80°. Determination of urinary LAM was done using Quantitative sandwich ELISA technique with SINCERE© Human Lipoarabinomannan ELISA Kit (Sincere Bio, Beijing, china) as per manufacturer’s instructions. Optical density was read at 450 nm when colour in the wells changed from blue to yellow using ELISA reader within 15 min of adding the stop solution. The value of the analyte was calculated from a standard curve prepared by using professional curve fitting software.

Results

Fifty smear-negative individuals with clinically diagnosed PTB (as per NTEP guidelines of India) were enrolled. Their mean age was 29.4 yr. Thirty five (70%) study participants were in the age group of 15-30 yr. Thirty one (62%) participants were male. All participants recruited had cough, 43 had fever, 15 had a history of night sweats and 46 had a history of weight loss. Thirty six (72%) participants were new while 14 (28%) were previously treated. M. tb was culture positive in 37 (74%) participants. M. tb was detected by CBNAAT in 47 (94%) and LPA in 35 (70%) individuals.

Among the 10 cases detected by CBNAAT, nine were new cases and one was previously treated. Three sputum samples were collected from every participant for CBNAAT, LPA and culture sample size is small to draw any affirmative conclusions.

Fourteen previously treated individuals were found to have active disease by the treating physician. Thirteen out of 14 had M. tb grown in the culture. One culture was contaminated.

Fifteen smear-positive individuals with PTB were also enrolled as control. Their mean age was 31 yr and eight (63%) were of the age group 15-30 yr. Eleven (73%) participants were male. All these 15 smear-positive individuals with PTB had cough fever and weight loss. Nine participants had history of night sweats. Thirteen participants (87%) were new cases and two (13%) were previously treated. M. tuberculosis culture was positive in 13 (87%) but M. tb was detected both by CBNAAT and LPA in all 15 participants. Fifteen individuals with post tubercular sequel were also enrolled as controls. Their mean age was 56 yr and all were >30 yr. Of these 10 (67%) were male. M. tb was not detected in any of these individuals with a post tubercular sequel by smear microscopy, CBNAAT and MGIT-liquid culture but cough was documented in all of them. None of them reported any symptom of fever, weight loss or night sweats. Baseline parameters of the three groups are described in Table I. Rifampicin resistance was not detected in any microbiologically confirmed TB cases with molecular diagnostic test for TB (CBNAAT and LPA). None of the enrolled cases was found to be reactive for HIV.

Urinary lipoarabinomannan (LAM): Mean urinary LAM value in 47 PTB sputum smear-negative cases, microbiologically confirmed by CBNAAT was 0.76±0.23 ng/ml. Mean urinary LAM value in 37 sputum smear-negative culture-positive PTB cases was 0.84±0.18 ng/ml. In three of these sputum smear-negative culture-positive PTB cases, M. tb was not detected by sputum CBNAAT and their mean urinary LAM value was 0.7±0.026 ng/ml. In 13 sputum smear and culture negative cases but microbiologically confirmed by CBNAAT PTB cases, mean urinary LAM value was 0.49±0.07 ng/ml.

Mean urinary LAM value in all 15 smear-positive PTB cases was 1.068 ng/ml. In 13 sputum smear-positive and M. tb culture-positive PTB cases, the mean urinary LAM value was 1.115±0.31 ng /ml. In two sputum smear-positive and M. tb culture-negative PTB cases, the mean urinary LAM value was 0.765±0.004 ng/ml.

The mean urinary LAM value in the post tubercular sequel cases was 0.478 ng /ml. All these 15 post tubercular sequel cases were sputum smears and culture-negative. M. tb was not detected by sputum CBNAAT in any of these individuals (Table II).

Sensitivity, specificity, positive predictive (PP) value and negative predictive value (NPV) of the urinary LAM in smear-negative P-TB is shown in (Table III). Considering sputum culture as the gold standard, the area under the curve was 0.98 [95% confidence interval (CI): 0.95-1] in receiver operating characteristic (ROC) curve of urinary LAM for detection of TB and the cut-off value of urinary LAM was 0.58 ng/ml with 100 per cent sensitivity and 92.3 per cent specificity. Using CBNAAT as gold standard area under the curve was 0.61 (95% CI: 0.45-0.76) in ROC curve of urinary LAM for detection of TB. The cut-off value of urinary LAM for detection of TB using sputum CBNAAT as gold standard was 0.745 ng/ml with 88.9 per cent sensitivity and 100 per cent specificity (Figs. 1 and 2).

Close

Fig. 1

ROC curve of urinary LAM for detection of TB with taking sputum culture as gold standard. Area under the curve was 0.98 (95% CI: 0.95-1). Sensitivity: 100 per cent; specificity: 92.3 per cent. ROC, Receiver operating characteristic; LAM, lipoarabinomannan; CI, confidence interval; TB, tuberculosis.

Export to PPT

Close

Fig. 2

ROC curve of urinary LAM for the detection of TB with taking CBNAAT as gold standard. Area under the curve was 0.61 (95% CI: 0.45-0.76 Sensitivity: 88.9 per cent; specificity: 100 per cent specificity. CBNAAT, cartridge-based nucleic acid amplification test.

Export to PPT

Statistical methods: Significance among the three groups was found using the one-way ANOVA test. The P-value is <0.00001. The result was significant at P <0.01. The f-ratio value is 24.36696. ROC was prepared using urinary LAM for detection of TB with taking sputum culture and CBNAAT as gold standard. The urinary LAM values are separated into two groups as M.tb detected by CBNAAT and M.tb detected by culture in Table II. The cut-off values were calculated by Youden index, based on which specificity sensitivity, NPV and PPV were calculated using CBNAAT and culture as gold standard tests with a CI of 95 per cent

Discussion

LAM is a major non-protein lipid cell wall component present in Mycobacteria. The structure of LAM varies among different mycobacterial species and has potent virulence properties that can alter the host immune response. Circulating LAM or other antigens can be secreted from dead and/or active Mycobacteria, and are detected in serum, sputum and urine13. In active TB diseases pulmonary as well as extra pulmonary, the antigen enters the blood circulation secreted from the inflammatory TB site. These easily travel through the basement membrane in the kidney glomeruli and are secreted in the urine where these can be detected101415161718. However, urinary LAM value in individuals with TB varied anywhere between 1 ng/ml and 100 ng/ml depending on the clinical manifestation of TB. High values of urinary LAM were was correlated positively to other markers of high mycobacterial burden192021. It was also found that urinary LAM strongly correlates with density of mycobacteria in sputum samples and is often positive in individuals with smear-positive TB than smear-negative counterpart22.

Increased sensitivity of detection of urinary LAM by lateral flow assay was found in individuals with HIV-TB co-infection. The identification of active TB disease by lateral flow LAM particularly in individuals with low levels of circulating CD4 T-cells and advanced form of HIV-induced immunosuppression, can help in early anti-TB treatment initiation15. WHO policy recommends that the lateral flow urine LAM (LF-LAM) assay “should not be used for the diagnosis of TB (strong recommendation, low quality of evidence) except as specifically described for persons with HIV infection with low CD4 counts or who are seriously ill”23. The diagnosis of active TB in HIV-infected individuals by sputum CBNAAT and urinary LAM assays were complementary. The combined sensitivity of both these tests was higher than either test alone24. In childhood TB diagnosis urinary LAM had 83 per cent sensitivity and 85 per cent specificity with a cut-off value of 0.98 mg/l using microbiological as well as clinical confirmation as the standard reference and 33 per cent sensitivity and 60 per cent specificity with a cut-off value of 1.69 mg/l using microbiological confirmation only25.

HIV related immunosuppression in individuals increases their risk for opportunistic infections by various bacteria, viruses and protozoa26. The high burden of coinfecting pathogens releases their antigens into the circulation thereby in the urine. It has been found previously that the TB pathogen burden in HIV seronegative individuals are 10 times lower than that in HIV-positive individuals assessed by PCR5. Urinary LAM is reportedly 10- to 20- fold higher in HIV-positive individuals with active TB disease than HIV seronegative TB cases567. It is suggested that the LAM binding proteins in urine may be responsible for lower sensitivity of LAM concentration in HIV-negative TB cases found by prior immunoassays8. Copper complex dye with hydrogel nanocage, a unique technology in ELISA, (nanocage-based ELISA) finds a solution by effectively sequestrating the LAM protein with high affinity and averts the interfering urinary protein. This helps in concentrating the LAM protein in the urine by a large factor and thereby allows the detection of TB among those who are not infected with HIV91011. This high-affinity copper dye bait for urinary LAM is deployed in this study for identification of TB. The advantage of this study is that urine sample is more conveniently obtained including from those who do not expectorate sputum.

Despite adequate advise for good sputum collection, the results of microscopy, CBNAAT, LPA and culture may vary depending on the quality of the sputum provided. However, any test from urine may not have much variation in the results. Thus, in this study we have better results from urine sample compared to sputum.

Despite the strengths there were some limitations in this study. The present study is designed to gather the evidence of detection of urinary LAM in HIV-negative sputum smear-negative PTB cases using copper complex dye in a hydrogel nanocage. This technology was used because it captures LAM effectively, by displacing the non-LAM urine proteins. It can detect urinary LAM during active PTB irrespective of physiologic status and/or immunologic complications of HIV infection. The detection is however, not limited to individuals with TB colonization of the kidney91011. In addition, being a urine-based test has advantage over sputum-based testing because urine is easy to collect, store and lacks the risks associated with sputum collection such as infection control. However, the sample size is small to derive an affirmative conclusion and more such rigorous studies are suggested to improve confidence in urinary LAM for the diagnosis of TB.

The urine LAM testing was carried out on thawed and not fresh samples. Furthermore, these were randomly chosen and it is unclear if this could have impacted the results. We also could not evaluate the urine LAM after treatment due to the limitation of time frame to conduct the study.