SESSION TITLE: Critical Care Posters SESSION TYPE: Original Investigation Posters PRESENTED ON: October 18-21, 2020 PURPOSE: ICU-acquired delirium frequently complicates critical illness with the elderly and those with pre-existing neurocognitive disorders at greatest risk. Primary central nervous system (CNS) diseases frequently present similarly to ICU-acquired delirium. Whether to perform lumbar puncture (LP) in all cases of diagnostic uncertainty remains unknown. Previous retrospective studies of hospitalized patients have shown that nosocomial meningitis is rare. In these cohorts, LP yield depended on indication and timing of the procedure and pretest probability of neurological disease processes. These studies are largely descriptive and are hampered by small sample sizes. The aims of this study were to establish the diagnostic yield of LP among the critically ill and to determine therapeutic efficacy i.e. how often management changed based on LP results. METHODS: This retrospective cohort study examined patients admitted to a medical ICU service from 2012 to 2018 at a tertiary academic center. Patients were included if they had altered mental status (or synonymous term) documented and underwent LP in the medical ICU.We gathered clinical variables including patient demographics, duration of hospitalization and ICU stay prior to LP, and duration of index ICU admission. We also collected patient variables such as immunocompromised status, underlying CNS disorders, and receipt of EEG, vasopressors, sedatives, or antimicrobials. LP-specific variables included indication and cerebral spinal fluid (CSF) analyses.The primary outcome was LP diagnostic yield as defined by CSF pleocytosis, elevated CSF protein, positive cytology, or microbiological evidence of infection. The secondary outcome of interest was the therapeutic efficacy on a Likert scale of 1-5 which was subsequently dichotomized.We performed chi-square analyses for individual variables and multivariable logistic regression of all variables. To assess the reliability of therapeutic efficacy, a second reviewer rated a subset of cases on the same Likert scale. RESULTS: We identified 104 records of patients who met inclusion criteria. While 48% of LPs yielded a "positive result” (50/104), only 16% (17/104) demonstrated abnormal cytology (3 cases) or evidence of CNS infection (14 cases). Moreover, only 13% (14/104) of LPs had therapeutic efficacy.In both chi-square and multivariate logistic regression analyses, we did not identify any variables that were significantly associated with diagnostic yield or therapeutic efficacy of LP. Cohen's kappa was 0.64 for dichotomized ratings of therapeutic efficacy (2 reviewers, 40 charts), indicating modest to strong agreement. CONCLUSIONS: Diagnostic yield and therapeutic efficacy of LP were low among critically ill medical ICU patients and likely cannot be predicted from clinical variables alone. CLINICAL IMPLICATIONS: Physicians must rely on clinical judgement to determine the necessity of LP in medical ICU patients. DISCLOSURES: No relevant relationships by Paul Bergl, source=Web Response No relevant relationships by Rahul Nanchal, source=Web Response No relevant relationships by Meghan Nothem, source=Web Response No relevant relationships by Alan Salazar, source=Web Response No relevant relationships by Shama Sharwani, source=Web Response
Read full abstract