Study objectives: Elderly patients with fever and altered mental status (AMS) are often pan-cultured to search for a source. Evaluations may be halted when an initial source is found. However, these patients may be harboring undiagnosed bacterial meningitis (BM; potentially translocated from the initial source) inadequately treated by initial antibiotic coverage. This study analyzes the contribution of lumbar puncture in these patients. A secondary objective is to analyze differences between nursing home and private residence patients. Methods: This was a multicenter (60,000 tertiary and 40,000 secondary center census) retrospective medical record review. All patients older than 65 years, with a temperature greater than 100.6°F, with new onset AMS, who had a fever evaluation (chest radiograph, urinalysis, blood cultures) inclusive of computed tomography (CT)–lumbar puncture in the emergency department were included. Excluded were patients in a known persistent vegetative state and diagnostic finding on CT (acute stroke, hemorrhage, abscess, mass). Final diagnosis was determined by composite endpoint of cerebrospinal fluid or blood culture, viral polymerase chain reaction, serology, discharge, or expiration diagnosis. A power analysis with α equal to 0.05 and β equal to 0.20 to detect a 20% difference in BM rates between positive pre-LP source and negative pre–lumbar puncture source stipulated a need for 98 subjects. Univariate analysis with appropriate tests for continuous and categoric data was used. Results: There have been 161 enrollees to date, 93 (58%) admitted from a private residence and 68 (42%) from a nursing home. Mean age was 77.6 years (range 65 to 97 years). Seventy-four (46%) patients had a non–central nervous system infection diagnosed pre–lumbar puncture: 38 pneumonias, 27 urinary tract infections, 9 other (skin, joint, intra-abdominal). Thirty-five (21%) patients had a source identified by lumbar puncture: 10 (6.2%) cases of BM, 21 (13%) cases of viral meningitis, and 4 (2.5%) cases of viral encephalitis (1 West Nile). Seventy-two percent (18/25) of patients with viral pathology on lumbar puncture had a negative pre–lumbar puncture evaluation. In patients with a positive pre–lumbar puncture source, there were 7 of 74 (9.5% confidence interval [CI] 6 to 18) cases of BM versus 3 of 87 (2.8% CI 1.5 to 7) cases of BM with a negative pre–lumbar puncture evaluation ( P P P =.37). The relative risk for nursing home patients to get lumbar puncture with a positive pre–lumbar puncture source was 0.44 (95% CI 0.14 to 0.63). This is a retrospective study of only those patients who underwent lumbar puncture. Therefore, it cannot take into account the clinical characteristics of fever/AMS patients who did not undergo lumbar puncture in nursing home patients with a positive pre–lumbar puncture source is due to the fact that many of those patients do not receive lumbar puncture. This piece of data is therefore partially inferential. A prospective study performing lumbar punctures on all appropriate patients with fever and AMS would be the next logical step. Conclusion: Patients with fever/AMS and a pre–lumbar puncture source of infection may have a higher rate (9.5% versus 2.8% in this study) of BM. When a pre–lumbar puncture source is found, nursing home patients appear less likely to receive lumbar puncture. Emergency physicians should perform lumbar punctures on all elderly patients with fever/AMS.