Antibiotic prophylaxis in intensive care units: meta-analyses versus clinical practice

Objective: At least 7 meta-analyses (MA) have been published since 1991 on the effectiveness of antibiotic prophylaxis in Intensive care units (ICU) p atients, but controversy still remains about the overall effectiveness and risk-benefits profile of the treatment. This paper aims to summarise availa ble data on effectiveness and discuss reasons why the controversy is still open and possible directions for future research. Design: Review of available published MA on the effectiveness of various re gimens of antibiotic prophylaxis with particular emphasis on the results of the individual patient data analysis published in 1998. Setting: MA or randomised control trials (RCTs) published and unpublished, conducted anywhere in the world. Patients and participants: Unselected adult ICU populations included in stu dies, published and unpublished, comparing different forms of antibiotic pr ophylaxis. Main outcome measure: Respiratory tract infections (RTIs) - however defined in individual studies - and total mortality. Data sources: General information from the 7 MAs published between 1991 and 1999 and detailed information from the MA published in the British Medical Journal in 1998 that reported data on 5727 patients enrolled in 33 RCTs; a ccess to individual patients data could be obtained from 25 of 33 RCTs and allowed a confirmatory individual patient MA on 4343 patients. Results: Pooled estimates from 16 RCTs (including 3361 patients) testing th e effect of the topical and systemic antibiotic combination indicates a sig nificant reduction of both RTIs (OR = 0.35, 95 % CI = 0.29-0.41) and total mortality (OR = 0.80, 95 % CI = 0.69-0.93). Five and 23 patients need to be treated to prevent one infection and one death, respectively, using this t reatment. Pooled data from the 17 RCTs (including 2366 patients) testing th e effect of a regimen based on topical antimicrobials indicated a statistic ally significant reduction in RTIs (OR = 0.57, 95 % CI = 0.46-0.69) but not in total mortality (OR = 1.01; 95 % CI = 0.84-1.22). Individual patient da ta analyses confirmed these results. Conclusions: After over 30 RCTs and seven MAs, there is strong, evidence th at antibiotic prophylaxis can reduce both RTIs and total mortality in ICUs patients in a statistically and clinically significant way. Concerns about the possible occurrence of antimicrobial resistance are not supported by av ailable data but cannot, at the same time, be ruled out due to methodologic inadequacies of the studies carried out so far. Whether new trials are nee ded, and how they should be designed to answer the question of the potentia l for antibiotic resistance following widespread use of the treatment, are now the main issues to be settled. Convening an international panel of clin ical experts and methodologists could be appropriate, in order to explore t he best way to resolve the controversy that seems to be preventing the wide spread use of a treatment that the best analysis of available data now indi cates is effective.