A prospective study of the diagnostic utility of sputum Gram stain in pneumonia.

INTRODUCTION: Sputum Gram stain and culture have been said to be unreliable indicators of the microbiological diagnosis of bacterial pneumonia. The etiological diagnosis of pneumonia is surrounded by great degree of uncertainty. This uncertainty should be and can be calculated and incorporated in the diagnosis and treatment. STUDY OBJECTIVES: To determine the diagnostic accuracy and diagnostic value of sputum Gram stain in etiological diagnosis and initial selection of antimicrobial therapy of bacterial community acquired pneumonia (CAP). DESIGN-METHOD: Prospective study of 1390 patients with CAP admitted January 2002-June 2008, to our institutions. Of the 1390 patients, 178 (12.8%) fulfilled the criteria for inclusion into this study (good-quality sputa and presence of the same microorganism in blood and sputum cultures which was used as gold standard for assessing the diagnostic accuracy and diagnostic value of sputum Gram stain). RESULTS: The sensitivity of sputum Gram stain was 0.82 for Pneumococcal pneumonia, 0.76 for Staphylococcal pneumonia, 0.79 for Haemophilus influenzae pneumonia and 0.78 for Gram-negative bacilli pneumonia. The specificity of sputum Gram stain was 0.93 for Pneumococcal pneumonia, 0.96 for Staphylococcal pneumonia, 0.96 for H. influenzae pneumonia and 0.95 for Gram-negative bacilli pneumonia. The positive likelihood ratio (LR+) was 11.58 for Pneumococcal pneumonia, 19.38 for Staphylococcal pneumonia, 16.84 for H. influenzae pneumonia, 14.26 for Gram-negative bacilli pneumonia. The negative likelihood ratio (LR-) was 0.20 for Pneumococcal pneumonia, 0.25 for Staphylococcal pneumonia, 0.22 for H. influenzae pneumonia, and 0.23 for Gram-negative bacilli pneumonia. CONCLUSIONS: Sputum Gram stain is a dependable diagnostic test for the early etiological diagnosis of bacterial CAP that helps in choosing orthological and appropriate initial antimicrobial therapy.

Prevalence of nosocomial infections after surgery in Greek hospitals: results of two nationwide surveys.

OBJECTIVE: To determine the frequency and type of nosocomial infections (NIs) (especially surgical-site infections [SSIs]), risk factors, and the type and duration of antibiotic use among surgical patients in Greek hospitals. DESIGN: Two point-prevalence studies. SETTING: Fourteen Greek hospitals. PATIENTS: Those in the hospitals during two prevalence surveys undergoing surgery during their stay. RESULTS: In the 1999 survey, 129 of 1,037 surgical patients had developed 148 NIs (14.3%). A total of 1,093 operations were registered, and 49 SSIs (4.5%) were found. In the 2000 survey, 82 of 868 surgical patients had developed 88 NIs (10.1%). A total of 902 operations were registered, and 38 SSIs were detected (4.2%). The median length of stay (LOS) for surgical patients without SSI was 10.0 days (range, 1-19 days); for patients who developed SSI it was 30 days (range, 1-52 days; P < .001). The median LOS prior to surgery for patients without SSI was 1 day (range, 0-4 days); for patients who developed SSI it was 3 days (range, 0-7.5 days; P < .001). Among 30 possible risk factors studied, wound class, LOS prior to surgery, and central venous catheterization were independent predictors of SSI. Median durations of prophylactic antibiotic therapy were 4 days (range, 1-14 days) and 6 days (range, 1-16 days) in the 1999 and 2000 surveys, respectively. CONCLUSION: Surgical patients in Greek hospitals suffered higher rates of SSI than did surgical patients in other developed countries while prophylactic antibiotics were used excessively.