Reduction on antibiotics (ab) consumption and costs through a physicians educational program in a general hospital

F. Paganin, F. Chan-ou-teung, D. Graber, M. Masanovic, N. Lugagne, Y. Rateau, MC. MC. Jaffar, Th. Béraud, B-A. Gaüzere and the CLIN. Hôpital Félix Guyon, Saint-Denis-Réunion. France.

Background : In most hospitals both antibiotics consumption and emergence of resistant bacterial strains rise. Based on a large and costly empirical prescription, the use of broad spectrum antibiotics is controversial and deleterious for the bacteriological ecology.

Aim and Methods : Since January 1996, a multidisciplinary committee (physicians, bacteriologists, pharmacists) is assisting clinicians prescribe AB according to good practices. The local bacterial ecology is beeing reviewed by bacteriologists (ie % of MRSA, broad spectrum enterobacterias). A prescription form is required to obtain "restricted anti-infective agents": IV broad spectrum penicillins, cephalosporins and quinolones, fosfomicin, amikacin, glycopeptides, rifampicin, fucidic acid, anti-fungal drugs. Not restricted are: oral penicillins, quinolones, macrolides, gentamicin, cefotaxime and ceftriaxone. The initial prescription covers a 3 days period pending the bacteriological results. The committee reviews the prescription forms on a weekly basis. Whenever abnormalities are detected, informations are seeked from the prescribing physicians and advice is provided, in order to foster the use of small spectrum or oral AB, according to the clinical and bacteriological situations.

Results : A dramatic decrease on some AB consumptions is noted: ceftazidime 70%, cefepime 56%, amikacin 53%, IV quinolones 47%, aztreonam 46%, fosfomicin 30%, vancomicin 7,7%, piperilin 10%, IV fluconazole 31%, oral fluconazole 12%. Conversely, there is an increase in the consumption of cefotaxime +29%, oral quinolones + 18%, ceftriaxone + 17%, and imipenem + 13%. The savings reaches 17,6% ($108,700). Physicians compliance is good, with only 2 out of 33 departments that resist the recommendations. Both mortality and emergence of resistant strains remain stable.

Conclusions : An educational program based on good prescription practices, is helpfull in reducing the AB consumption. The long-term impact on the ecology is to be evaluated.

EFFICACY OF CEFOTAXIME VERSUS DOUBLE REGIMENS IN THE MANAGEMENT OF SEVERE COMMUNITY-ACQUIRED PNEUMONIAE (CAP) IN ICU. F PAGANIN, D MORAU, E SPILMANN, J SOUK-ALOUN, JL YVIN, A STECIUK, Y DJARDEM, T GARNAUD, F LILIENTHAL*, B-A GAUZERE. Service de Réanimation, Médecine Interne, CHD Félix Guyon. Saint Denis-Réunion. FRANCE, *Collegeville, PA, USA

Background: CAP requiring admission in ICU are usualy linked with a high mortality rate (22 to 54%), particularly in patients undergoing mechanical ventilation (MV). Causative agent identification does not influence the overall survival. Local bacteriological epidemiology guides the appropriate empirical antibiotherapy (AB). Concensus conferences usually recommend the use of an AB combination in the management of severe CAP.

Aim and Methods: 82 patients (54 +/-15 y) were admitted in our ICU from July 1994 to December 1996. Epidemological profile was assessed through fiberbronchoscopy (FB) with bronchoalveolar lavage and/or protected brush. Empirical AB was started according to the direct examination (DE) and the physician’ experience. Monotherapy (cefotaxime 3 to 4 gr/day) was used whenever DE showed gram + cocci (likely pneumococcus) or gram - bacili (haemophilus or klebsiella). Double regimen (b lactam + macrolide or quinolone or aminoglycoside) was used in case of negative DE or FB not feasible.

Results: 80 % required MV. Identification was obtained in 80% (S. Pneumoniae: 41%, Klebsiella: 20%, Haemophilus: 5%, Miscellaneous: 17%). In 14 patients (17 %) no agent was identified. Cefotaxime alone was administered in 15 cases (group 1) and multiple therapy in 65 (group 2). The 2 groups were identical for: demographics data, shock and acidosis, positive blood cultures, % of MV. The overall mortality was 38 %, 33.3% in group 1 and 38.5 % in group 2 (NS). Mortality rates were identical among patients with or without bacteriological identification

Conclusion: Monotherapy based on cefotaxime is safe in the management of severe CAP in ICU. However, the implementation of such a regimen requires a bacterial identification and a knowledge of the local epidemiology.

MANAGEMENT OF SEVERE COMMUNITY-ACQUIRED PNEUMONIAE (CAP) IN ICU: PROGNOSIS AND RISK FACTORS. T GARNAUD, E SPILMANN, D MORAU, J SOUK-ALOUN, F LILIENTHAL**, B-A GAUZERE, H HUBERT*, F PAGANIN,. Service de Réanimation, CHD Félix Guyon. Saint Denis-Réunion. FRANCE,* CHU Lille, **Collegeville, PA, USA

Background and Aim: CAP requiring admission in ICU are usualy linked with a high mortality rate (22-54%), particularly in patients requiring mechanical ventilation (MV). Causative agent identification does not influence the overall survival. However, we hypothized that a local epidemiology specificity (high incidence of klebsiella) could be correlated with a fatal outcome.

Methods: 82 patients (54 +/-15 y) were admited in our ICU from July 1994 to December 1996. Epidemological profile was assessed through fiberbronchoscopic bronchoalveolar lavage and/or protected brush (FB). All demographic and usual data used in ICU were recorded. Univariate and multivariate analysis (logistic regession) were performed in order to assess prognosis factors.

Results: MV (81 %), alcoholism ( 60%), smoking (44%), COPD (29%) and age over 65 (26%) of patients. Bacterial identification was obtained in 80% (S. Pneumoniae: 41%, Klebsiella: 20%, Haemophilus: 5%, miscellaneous: 17%). No agent was identified in 14 patients (17 %). SAPS II was 44.76+/-20. Shock on admission (40%), overall mortality (38%). Hemodynamic monitoring was performed and revealed a septic shock profile in all 16 patients (CI: 4.2+/-1.8, SVRI: 949+/-301). Univariate analysis concluded as poor prognosis factors: alcoolism, absence of fever, shock, MV, leukopenia, hypoxia, acidosis and number of lobes involved. But the multivariate analysis model revealed only 3 significant poor prognosis factors: shock, MV and radiographic involvement. The probability of death calculated with the multivariate analysis model is 42% (observed 38%) versus the one predicted by SAPS II (40%).

Conclusion: In this study, the high incidence of Klebsiella is not correlated with a higher mortality rate. Most of the poor prognosis factors, revealed by univariate analysis, are related to shock and multiple organ failure. SAPS II appears a good predictive tool for CAP outcome in ICU. Howhever, epidemiological data are deemed necessary both to initiate empiric antibiotherapy and to monitor the evolution of resistance patterns.