All-cause 8-week mortality is reduced by rapid point-of-care urine-based testing for tuberculosis (TB), using a test for detection of mycobacterial lipoarabinomannan (LAM) to guide rapid treatment initiation in hospitalised HIV-positive people with signs and symptoms of TB, according to a late-breaker study presented by Dr Jonny Peter of the University of Cape Town, and colleagues, at the 46th Union World Conference on Lung Health in Cape Town from 2 to 6 December 2015.
Those who received LAM testing in addition to standard of care diagnostics were 18% less likely to die at 8 weeks (aHR = 0.82; 95% CI: 0.70-0.96, p = 0.015).
The LAM strip test measures for the presence of lipoarabinomannan in urine. The test has shown variable sensitivity and specificity in different populations, and has performed best in studies of hospitalised patients with HIV infection and advanced immunosuppression (CD4 count below 150 cells/mm3). The test is not considered suitable for outpatient screening. (Lawn 2012)
Speeding up diagnosis by using a point-of-care test in addition to laboratory diagnostic tests such as TB culture or Xpert MTB/RIF has the potential to speed up the initiation of treatment and to increase the number of people receiving treatment.
The randomised controlled trial studied whether the LAM strip test, in combination with current available TB diagnostics, would improve TB-related mortality, morbidity and length of hospital stay in HIV-positive people. TB mortality is high, especially in people living with HIV with extensive immune damage, where a TB diagnosis is already difficult.
The study was conducted in South Africa, Tanzania, Zambia and Zimbabwe.
Of the 8728 patients with presumed HIV/TB co-infection from secondary and tertiary hospitals 2659 were randomised to be included in the study, made up of 1336 people in the LAM arm and 1323 in the non-LAM arm. Patients in the LAM arm were allocated LAM in addition to routine TB diagnostics (minimum of two sputum smear specimens and culture and one Xpert MTB/RIF, if available).
A total of 1257 patients were included in the modified intention to treat analysis and 59 were lost to follow-up.1198 patients were analysed per protocol analysis, after 70 did not receive the LAM test, 47 did not meet the inclusion criteria and nine were excluded from the analysis due to withdrawing consent and missing hard copy documents.
A total of 1323 people were allocated to the non-LAM arm where routine TB diagnostics were used to detect TB. 1271 patients were included in the modified intention to treat analysis after 38 did not receive diagnostic tests, 23 did not meet the inclusion criteria and 14 were excluded from the analysis. An additional 58 people were lost to follow-up resulting in 1213 patients being included in the per protocol analysis.
Of the 2528 people included in the study, the median age was 37 (IQR: 30-44) and 51% (n = 1300) were women. 27 % (n = 682) had previously had TB. The median CD4 count was 84 cells/mm3 (IQR: 26 – 208 cells/mm3), with 2% (n = 59) at WHO clinical Stage 1, 27%(n = 671) at Stage 2, 53% (n = 1331) at Stage 3 and 17 % (n = 438) at Stage 4. Almost half (48%, n = 1224) were on ART at study entry and 73% (n = 615/848) of eligible patients were initiated on ART at 8-week follow up.
Of the total number of people started on anti-TB treatment, 52% (n = 648) of people who were diagnosed using LAM compared to 48% (598) in the non-LAM arm (p = 0.024). The time-to-treatment analysis showed significant differences in the cumulative proportion of people receiving TB treatment and the number of deaths for patients who started treatment by the end of Day 1 (55.1 % vs 40.3%); end of Day 2 (71.1% vs 60.5%); end of Day 3 (79.2% vs 69.1%) and the end of Day 4 (84.0 % vs 75.6%) between the LAM and non-LAM arm respectively (p < 0.001 for all of these).
The study also showed that there was no significant difference between the two groups in the hospital length of stay from enrolment or change in TB morbidity as measured by Karnofsky score in patients on TB treatment.
Despite overall suboptimal test sensitivity even for people with low CD4 counts, as shown in a study published in July this year by the same group, the ability of LAM to identify patients at high risk of death or lost-to-follow-up offers important prognostic value.
Sensitivity measures the percentage of results that will be (correctly) positive when a disease is actually present. Lower rates of sensitivity will produce more false negative results.
This previous study of outpatients with HIV showed that among 583 participants with signs and symptoms of TB that had HIV or refused testing, the overall LAM sensitivity was 22.7 % (95% CI: 16.6 - 28.7; n = 41/181) and 30.4 % (95% CI: 17.1 - 43.7; n = 14/46) in the CD4 ≤ 100 cells/mm3 sub-group. Among culture-positive TB cases, adjunctive LAM testing did not improve the sensitivity of either sputum Xpert-MTB/RIF [78.2 % (95% CI: 69.8 - 86.7) versus 76.1 % (95% CI: 67.4-84.8, p = 0.7] or smear-microscopy [56.2 % (95% CI: 45.9 - 66.5) versus 43.8 % (95% CI: 33.5 - 54.1, p = 0.1).
Lawn S. Point-of-care detection of lipoarabinomannan (LAM) in urine for diagnosis of HIV-associated tuberculosis: a state of the art review. BMC Infectious Diseases, 12:103, 2012. (Full text available here).
Peter J et al. Rapid point-of-care urine-based testing for tuberculosis and its impact on mortality: a multi-centre randomised controlled trial. 46th Union World Conference on Lung Health, Cape Town, 2015
Peter J et al. Test characteristics and potential impact of the urine LAM lateral flow assay in HIV-infected outpatients under investigation for TB and able to self-expectorate sputum for diagnostic testing. BMC Infect Dis. 2015; 15: 262. Published online 2015 Jul 9. doi: 10.1186/s12879-015-0967-z