The proposed study had been performed in order to explore metabolic profiling in CSF, serum and urine for differentiation of BM, TBM (meningitis cases) and controls in adults using NMR spectroscopy.Meningitis is a serious disease accompanied by acute inflammation of the meninges, associated with raised intracranial pressure linked with serious neurological sequelae. BM is the most common form of meningitis constituting about 80%¹ of all infected cases while, TBM is the most severe form of extra pulmonary tuberculosis. The differential diagnosis is quandary with low sensitivity and specificity as it depends on clinical/microbiological (CSF)/pathological/radiological parameters and hence it results to high risk of morbidity and mortality. This study therefore intends to explore metabolic profiling of serum and urines samples also for differentiating meningitis.Patients with confirmed clinico- radiological/ pathological and/ or microbiological evidence of bacterial and tubercular meningitis (n=51) were admitted in hospital with a prior approval from Institutional Ethical Committee. The CSF, serum and urine were collected using standard protocol and stored in -80ºC. The NMR spectra were acquired using Avance III 800 MHz NMR spectrometer (BrukerGmBH, Germany). One-dimensional NOESY-preset and CPMG experiments were recorded in all the samples. The metabolites were quantified using QUANTAS (QUANTification by Artificial Signal)² and subjected to multivariate stepwise Discriminant Functional Analysis (DFA) using SPSS version 21 software. The case vs control study was performed using healthy controls (n=32) for serum and urine samples and neurological disease controls for CSF samples. The Orthogonal Signal Correction using two components followed by Principal Component Analysis (OSC-PCA) of significant metabolites of CSF, serum and urine samples were carried out using ‘The Unscrambler X’ Software package (Version 10.0.1, Camo ASA, Norway).83 subjects were enrolled for the study, out of which there were 51 cases of meningitis and remaining 32 were controls. Among 51 cases BM (pyogenic meningitis) and TBM were 21 and 30 respectively. Twenty one metabolites were quantified in CSF (Figure1c), showed varied alterations along with the upregulation of lactate, pyruvate, and citrate, and downregulation of glucose in cases (Figure 1a). Similarly among BM and TBM group significant increase of 2-hydroxyisovalerate, isobutyrate and formate in BM cases was observed (Figure 4a). DFA provided correct classification of 78.0% in both the groups viz. case vs. control group and BM vs. TBM. Among the twenty three metabolites quantified in serum, valine and acetate were found to be elevated among case vs control groups (Figure 2c). Metabolites viz. ethanol, acetoacetate and succinate were observed in serum samples of cases only (Figure 2a). Comparatively, BM and TBM showed a depletion of alanine in BM cases (Figure 4b). The DFA of cases and controls afforded correct classification of 93.5%, and of BM and TBM 60.0%. Twenty six metabolites were quantified in the spectra of urine samples (Figure 3c). Quantified metabolites resulted into clear distinction among cases and controls. Elevated metabolites in meningitis cases were ketonic bodies and GPC with depletion of creatine/creatinine and malonate (Figure 3a). Decrease in valine and acetone levels in BM, differentiated BM vs TBM group (Figure 4c). The multivariate DFA afforded correct classification, 89.7% in case vs control and 89.1% in BM vs TBM group. The 3D score plot of PCA using significant metabolites of DFA showed clear distinction among case vs control groups.

The OSC-PCA of significant metabolites obtained through DFA in CSF samples could not clearly differentiate meningitis cases into separate clusters. However, clear classification between case vs. control group was obtained in alternative bodyfluids i.e. serum and urine samples respectively. This may be because neurological disease control subjects were used instead of healthy controls for the comparison of metabolic profile in CSF samples.

The metabolic profile of serum and urine correlates with infection and inflammation. The elevation of acetate in both the biofluids of cases, can be accounted for due to infection in which plasma free fatty acids concentration have been reported to be variably increase or decrease in human and experimental animals. The increase of alanine levels in serum samples of TBM cases may represent changes in the source of excitatory amino acid synthesis^3,4. The elevation of choline levels in urine may be due to the breakdown of membrane phospholipids leading to neuro-degeneration. Depletion of malonate may be due to its utilization by mycobacterial species for their growth as reported earlier during tissue hypoxia in children with tuberculosis⁵.