Evaluation of the fluorescent-thin layer chromatography (f-TLC) for the diagnosis of Buruli ulcer disease in Ghana

Abstract

Background Buruli ulcer is a tissue necrosis infection caused by an environmental mycobacterium called Mycobacterium ulcerans (MU). The disease is most prevalent in rural areas with the highest rates in West and Central African countries. The bacterium produces a toxin called mycolactone which can lead to the destruction of the skin, resulting in incapacitating deformities with an enormous economic and social burden on patients and their caregivers. Even though there is an effective antibiotic treatment for BU, the control and management rely on early case detection and rapid diagnosis to avert morbidities. The diagnosis of Mycobacterium ulcerans relies on smear microscopy, culture histopathology, and PCR. Unfortunately, all the current laboratory diagnostics have various limitations and are not available in endemic communities. Consequently, there is a need for a rapid diagnostic tool for use at the community health centre level to enable diagnosis and confirmation of suspected cases for early treatment. The present study corroborated the diagnostic performance and utility of fluorescent-thin layer chromatography (f-TLC) for the diagnosis of Buruli ulcer. Methodology/Principal findings The f-TLC method was evaluated for the diagnosis of Buruli ulcer in larger clinical samples than previously reported in an earlier preliminary study Wadagni et al. (2015). A total of 449 patients suspected of BU were included in the final data analysis out of which 122 (27.2%) were positive by f-TLC and 128 (28.5%) by PCR. Using a composite reference method generated from the two diagnostic methods, 85 (18.9%) patients were found to be truly infected with M. ulcerans, 284 (63.3%) were uninfected, while 80 (17.8%) were misidentified as infected or noninfected by the two methods. The data obtained was used to determine the discriminatory accuracy of the f-TLC against the gold standard IS2404 PCR through the analysis of its sensitivity, specificity, positive (+LR), and negative (–LR) likelihood ratio. The positive (PPV) and negative (NPV) predictive values, area under the receiver operating characteristic curve Azevedo et al. (2014), and diagnostic odds ratio were used to assess the predictive accuracy of the f-TLC method. The sensitivity of f-TLC was 66.4% (85/128), specificity was 88.5% (284/321), while the diagnostic accuracy was 82.2% (369/449). The AUC stood at 0.774 while the PPV, NPV, +LR, and–LR were 69.7% (85/122), 86.9% (284/327), 5.76, and 0.38, respectively. The use of the rule-of-thumb interpretation of diagnostic tests suggests that the method is good for use as a diagnostic tool. Conclusions/Significance Larger clinical samples than previously reported had been used to evaluate the f-TLC method for the diagnosis of Buruli ulcer. A sensitivity of 66.4%, a specificity of 88.5%, and diagnostic accuracy of 82.2% were obtained. The method is good for diagnosis and will help in making early clinical decisions about the patients as well as patient management and facilitating treatment decisions. However, it requires a slight modification to address the challenge of background interference and lack of automatic readout to become an excellent diagnostic tool.