Diagnostic Accuracy of Cerebrospinal Fluid Multiplex Polymerase Chain Reaction Panel Testing in Patients With Suspected Central Nervous System Infections: A Multi-Center Study in the United Arab Emirates

Background Delays in diagnosis and treatment of central nervous system (CNS) infections can lead to significant morbidity and mortality among children and adults. Prior antibiotic treatment is a major hurdle to accurate diagnosis due to falsely negative cerebrospinal fluid (CSF) cultures in partially treated patients. Increasingly, molecular diagnostic methods using multiplex polymerase chain reaction (mPCR) testing on CSF samples are being utilized in clinical practice for timely and accurate diagnosis. However, there is no data regarding the diagnostic accuracy or clinical impact of CSF mPCR testing in the Middle East region. We sought to compare the diagnostic accuracy of an automated mPCR CSF panel with routine CSF culture, the current gold standard, in the United Arab Emirates (UAE). Methods This single-gated, multi-center, diagnostic accuracy study included patients from birth onwards who were admitted to any of the three participating hospitals with an initial diagnosis of meningitis or encephalitis, between January 2017 and March 2021, and had CSF samples collected for mPCR and culture. Sociodemographic, clinical, and molecular data were collected for all. Results A total of 353 CSF samples were collected from patients from 0-90 years old hospitalized for suspected CNS infection. Children constituted 51% of the study population, and males were slightly over-represented (55.2%). Pathogens were detected by mPCR in 78 (22%) CSF samples, of which 19 (24%) were bacteria and 59 (76%) were viruses. No fungal pathogens were detected. Enteroviruses were the most prevalent CNS pathogen among our cohort (40%), followed by herpes simplex virus type 2 (HSV-2) (12.5%). Children constituted 69% of positive samples for enterovirus, while HSV-2 was exclusively detected among adults. Using CSF culture as the diagnostic gold standard, the mPCR panel demonstrated high specificity (100%) and sensitivity (96.3%) in diagnosing CNS infection among all age groups. mPCR testing demonstrated a high overall percentage of agreement (OPA) with CSF culture (98.9%). Patients with bacterial meningitis had a significantly longer hospitalization (p=0.004) and duration of antibiotic therapy (p=0.001) compared to those with viral meningitis. Three CSF samples were negative on mPCR testing but positive on culture. These pathogens included: methicillin-sensitive Staphylococcus aureus (MSSA), Bacillus cereus, and Mycobacterium Tuberculosis (MTB). In addition, 13 patients had negative CSF cultures but tested positive on CSF mPCR. These pathogens included Streptococcus pneumoniae (seven patients), Haemophilus influenzae (three patients), Streptococcus agalactiae (two patients), and Escherichia coli (one patient). All discordant results were confirmed by reviewing the patient's clinical presentation, CSF analysis, clinical course, and final diagnosis. Conclusion CSF mPCR panel is a highly sensitive and specific diagnostic tool for the diagnosis of CNS infections among all age groups in the UAE. Routine use of CSF mPCR panels can decrease healthcare costs by reducing the length of stay and can also aid antibiotic stewardship efforts by reducing antibiotic overuse in patients with viral CSF infections. CSF culture and mPCR complement each other by identifying CNS pathogens in patients with prior antibiotic exposure who would otherwise be missed if relying on CSF culture alone. However, concomitant CSF culture samples should be sent to avoid missing unusual CNS pathogens.