Tuberculostearic Acid, a Potential Parameter for Scoring System Construction for Tuberculous Meningitis Diagnosis

Abstract

Background: This study aimed to validate the value of tuberculostearic acid (TBSA) whether it could implicate the existence of Mycobacterium tuberculosis (Mtb) and assist for the clinical diagnosis of tuberculous meningitis (TBM). Methods: The patient's cerebrospinal fluid (CSF) specimen was collected through the lumbar puncture and detected for TBSA with gas chromatography/mass spectrometry. At the same time, gold standard tests, i.e., CSF direct culture, CSF smear microscopy, or nucleic acid amplification tests, for Mtb were routinely performed. Furthermore, we evaluated all patients by the Lancet consensus scoring system, which classifies suspected patients to “Definite (depend on gold standard results only),” “Probable (>10 pts without imaging or >12 pts with imaging information),” “Possible (6–9 pts without imaging or 6–11 pts with imaging),” and “Not (<6 pts or with alternative diagnoses)” TBM. Results: In total, 140 patients were admitted for our study included 27 confirmed TBM patients and 50 TBSA-positive patients. Sensitivity (0.7407, confidence interval [CI] 95%: 0.5372–0.8889) and specificity (0.7345, CI 95%: 0.6432–0.8132) were calculated. The Lancet consensus scoring system was also applied to evaluate the possibility of TBM in suspected patients, finding that TBSA-positive patients showed a similar grouping distribution as the definite TBM patients. Conclusions: Our study implicates that the prospective utilization of TBSA is worth combining into a scoring system for characterizing the features of Mtb, showing a great potential of TBM diagnosis by TBSA in future.

Tuberculostearic acid incorporated predictive model contributes to the clinical diagnosis of tuberculous meningitis

“…However, positive TBSA demonstrated a similar distributive pattern of Definite TBM patients in the Lancet scoring system, which may indicate the highly suspected patients in the Possible or Not group. 15 TBSA model could find out 72% of Possible or Not TBM patients in CRA, in contrast to 59% using TBSA alone. Furthermore, the TBSA model was able to identify most (92.3%) of the Probable patients, showing this simpler and more practicable model with an improved capacity for TBM diagnosis.…”

“…We previously found that positive TBSA results shared a similar distribution of definite TBM patients classified by the Lancet scoring system, 15 suggesting a promising role of TBSA to improve TBM diagnosis. In the present study, we sought to integrate the benefits of TBSA with multiple diagnostic criteria to build up a diagnostic system, aiming to overcome the insufficient sensitivity of gold-standard tests.…”

“…In the past, while some indirect molecular indices such as Mtb-specific antibodies, and interferon-gamma release assays (IGRAs) level in cerebrospinal fluid (CSF) 13 , 14 have been reported to be valuable for TBM diagnosis, yet a near-direct biomarker, tuberculostearic acid (TBSA), was overlooked in the past two decades, 15 which showed 83.3%–100% sensitivity and 98.9%–100% specificity, and could be rapidly detected by gas-chromatography/mass spectrometry (GC/MS) in 1–2 days from last century’s studies. 16 , 17 , 18 , 19 It was known that TBSA, a mycobacterial origin lipid marker, is currently thought as a good candidate to indicate Mtb burden by lipidomics within a day in non-CNS infection, 20 but unexpectedly, not in TBM.…”