Multi-Center Study of Diffusion-Weighted Imaging in Coma After Cardiac Arrest.

BACKGROUND: The ability to predict outcomes in acutely comatose cardiac arrest survivors is limited. Brain diffusion-weighted magnetic resonance imaging (DWI MRI) has been shown in initial studies to be a simple and effective prognostic tool. This study aimed to determine the predictive value of previously defined DWI MRI thresholds in a multi-center cohort. METHODS: DWI MRIs of comatose post-cardiac arrest patients were analyzed in this multi-center retrospective observational study. Poor outcome was defined as failure to regain consciousness within 14 days and/or death during the hospitalization. The apparent diffusion coefficient (ADC) value of each brain voxel was determined. ADC thresholds and brain volumes below each threshold were analyzed for their correlation with outcome. RESULTS: 125 patients were included in the analysis. 33 patients (26%) had a good outcome. An ADC value of less than 650 × 10(-6) mm(2)/s in ≥10% of brain volume was highly specific [91% (95% CI 75-98)] and had a good sensitivity [72% (95% CI 61-80)] for predicting poor outcome. This threshold remained an independent predictor of poor outcome in multivariable analysis (p = 0.002). An ADC value of less than 650 × 10(-6) mm(2)/s in >22% of brain volume was needed to achieve 100% specificity for poor outcome. CONCLUSIONS: In patients who remain comatose after cardiac arrest, quantitative DWI MRI findings correlate with early recovery of consciousness. A DWI MRI threshold of 650 × 10(-6) mm(2)/s in ≥10% of brain volume can differentiate patients with good versus poor outcome, though in this patient population the threshold was not 100% specific for poor outcome.

Characteristics of left-sided endocarditis due to Pseudomonas aeruginosa in the Detroit Medical Center.

Mortality due to left-sided endocarditis caused by Pseudomonas aeruginosa remains high despite the therapeutic regimen of high doses of an aminoglycoside plus a beta-lactam antibiotic. In this series, left-sided pseudomonal endocarditis presented as an acute illness refractory to optimal antibiotic therapy. Complications associated with active valvular infection, such as neurologic sequelae, ring and annular abscesses, congestive heart failure, and splenic abscesses, are frequent. The overall morbidity and mortality remain high; however, outcome appears to improve with surgery. Our present data suggest that early valve replacement accompanied by a 6-week course of high doses of a combination of drugs may be the optimal therapy for left-sided pseudomonal endocarditis. This approach not only may prevent serious and potentially lethal complications of the disease but also may cure them.

Fungemia caused by Candida species and Torulopsis glabrata in the hospitalized patient: frequency, characteristics, and evaluation of factors influencing outcome.

We reviewed 135 cases of candidemia occurring between 1983 and 1986 to examine oncologic and nononcologic populations and assess factors for survival. Candida albicans was the most common species (51%); Candida tropicalis occurred most frequently in leukemia patients (57%), whereas Candida parapsilosis and Torulopsis glabrata were associated with solid tumors and nononcologic diseases. Risk factors identified were: preceding surgery, antibiotics, cannulas, and steroids in solid tumor and nononcologic diseases; and chemotherapy and neutropenia with hematologic malignancies. Even transient cannula-associated candidemia was not a benign process. Intravenous cannulas were common portals of entry (39%) in debilitated patients without cancer (59%) and were associated with high mortality (55%). Overall mortality was 59%, candidemia directly contributing to death in 75% of cases. In patients with candidemia, failure to initiate therapy with amphotericin B had a negative influence on outcome, whereas analysis of the entire group identified severity of underlying illness as the dominant cofactor influencing outcome.