Date of Award


Document Type


Degree Name

Master of Medicine (MMed)

First Supervisor/Advisor

Professor Gerald Yonga

Second Supervisor/Advisor

Dr. Peter Waweru

Third Supervisor/Advisor

Professor Gunturu Revathi


Internal Medicine (East Africa)


Background: TB remains a key global health challenge; particularly Extra pulmonary Tuberculosis (EPTB) and is also a major cause of morbidity and mortality in sub-Saharan Africa. The increasing incidence of EPTB compounded by difficulties in making a timely diagnosis portends to poor prognosis and increased mortality. Delays in diagnosis and misdiagnosis of EPTB in suspected patients account for mortality as well. Challenges in diagnosis of EPTB are mainly due to disease related factors, the paucibacillary nature of disease and challenges with sample collection and processing. Reference standard test such as TB culture, geneXpert and histology which are used in the diagnosis of EPTB are insensitive (depending upon organ sites suspected), take time and specialized laboratory equipment. Emerging diagnostics utilizing pathogen derived biomarkers such as the urine LAM antigen which are rapid, unsophisticated, require minimal laboratory set-up and highly adaptable both in the field and outpatient departments thus become an appealing option. The LAM antigen has been studied in the diagnosis of PTB with varying sensitivities and specificities. This study evaluated the diagnostic utility of the LAM antigen in the diagnosis of EPTB compared to the current reference standard tests.

Methods: This prospective validation study was carried out between June 2015-April 2016 at The Aga Khan University Hospital, Nairobi and its satellite clinics. The main objective was to determine the sensitivity, specificity and predictive values of the LAM antigen compared to reference standard tests which included TB culture (solid and liquid media), histology and geneXpert suspected to have EPTB. Patients with at least 1 positive reference standard test were defined as cases of EPTB. We enrolled participants aged > 14 years who were clinically suspected to have EPTB by the primary attending physician. Participants with active or evidence of PTB and those having concomitant PTB and EPTB, as well as those who were unable to provide a urine specimen for LAM antigen testing or sample specimen for reference standard testing were excluded. Urine sample was obtained and tested with the LAM antigen. A grade 2 and above cut-off point was defined as a positive LAM antigen test. Laboratory personnel carrying out the reference test were blinded to results of LAM antigen. In addition, an independent reader reported 24 LAM strips was blinded to the results of both the reference tests and results of LAM antigen test done by the principal investigator. Sensitivity (Sn), specificity (Sp), positive and negative predictive values (NPV, PPV) and positive and negative likelihood ratios (LR+, LR-) were calculated for the LAM antigen.

Results: 99 participants were enrolled during the study period. In patients suspected to have EPTB, the LAM antigen had a sensitivity of 65.52% (95% CI; 45.67-82.06%) and a specificity of 94.29% (95% CI; 86.01-98.42%). Positive predictive value was 82.61% (95% CI; 61.22-95.05%) while the negative predictive value was 86.84% (95% CI: 77.13-93.51%). A positive LAM antigen test increased the likelihood of having EPTB, LR+= 11.47(95%CI; 4.27-30.78), while a negative LAM antigen decreased the probability of having EPTB, LR-=0.37(95% CI; 0.22-0.61). The accuracy of LAM antigen was 85.8%.Baseline and clinical demographics, particularly HIV status, CD4 count and prior history of TB were not found to be statistically significant with LAM antigen positivity in this study.

Conclusion: Overall, our findings suggest that the urine LAM antigen, an easy to use test and readily applicable point of care test has moderate sensitivity and high specificity in rapid detection of EPTB. It is a low cost alternate test to current EPTB diagnostic tests. Clinicians treating patients with suspected EPTB who have a positive LAM antigen, the high specificity of the LAM antigen test, together with high predictive values confidently allows them to make a more accurate diagnosis thereby initiating rapid and appropriate treatment. There is need to investigate if diagnostic accuracy can be improved if suspected specimens (fluids, tissue) were directly tested with the LAM antigen. Additionally, clinical impact and cost-effectiveness of this assay compared to reference standard tests should also be assessed.