Diagnosis of tuberculosis in a high burden resource limited setting

Dang, Thi Minh Ha (2012). Diagnosis of tuberculosis in a high burden resource limited setting. PhD thesis The Open University.

DOI: https://doi.org/10.21954/ou.ro.0000ee27

Abstract

It is estimated that over nine million new tuberculosis (TB) cases worldwide occurred in 2009. However, only 63% of these cases were notified, which is lower than the 75% target set by the DOTS strategy. The lack of ideal diagnostic tests is one of the major barriers to effective TB reduction. The aim of this thesis is to evaluate the efficiency of a novel liquid culture based technique (Microscopic Observation Drug Susceptibility Assay-MODS) and a new generation fluorescence microscope (Light Emitting Diodes fluorescence microscope - LED fluorescence microscope) in diagnosis of TB and multi drug resistant tuberculosis (MDR- TB) in Viet Nam, a resource limited, high burden setting. In the studies described in this thesis, MODS had a higher sensitivity than homogenous smear microscopy, detecting an additional 13% of cases, in diagnosis of TB. Importantly, the pooled sensitivity of MODS (53.0%) was comparable to that of Mycobacteria Growth Indicator Tube technique (MGIT) (57.8%) with clinical presentation as the gold standard. In terms of multi drug resistant tuberculosis (MDR- TB) diagnosis, MODS had rather low sensitivity in detection of Isoniazid (INH) or Rifampicin (RIF) resistant and MDR isolates, 72.6%, 72.7% and 77.8%, respectively against the proportional drug susceptibility testing (DST) method. The low sensitivity of DST-MODS in this study was probably due to low bacterial load samples and use of a high INH critical concentration (O.4~g/ml). For LED microscopy, the detection rate, sensitivity, specificity and false negative of LED 40X were comparable to those of conventional fluorescence microscope and light microscope. The false positive rate • of LED 40X increased in comparison with Ziehl-Neelsen (ZN) reading in the early phases of the evaluation (2.6% and 2.7% vs 0% in the validation and implementation phases vs continuation phase, respectively), probably due to lack of experience among technicians, but no significant difference was found (P>0.05). In the last phase, no false positive result was recorded. Although the reading time of fluorescence-LED (FM-LED) and ZN reading were less than 2 minutes, ZN-light microscope (ZN-LM) reading time was shorter than that of FM-LED40X, especially in positive smears (P=0.007) in the early study phase. By the end of the study, FM-LED reading time was shorter than ZN-LM which was recorded in the early phase. With high experienced technicians there may be little or no additional benefit to case detection if a light microscope is replaced by a fluorescence microscope. In general, MODS and LED microscopy are reliable methods for use in diagnosis of TB in resource limited settings due to their accuracy, reliability and low costs. Large scale operational projects should be conducted to evaluate the feasibility and cost-effectiveness of these methods in countries where these techniques are to be implemented.