Nosocomial pneumonia in 27 ICUs in Europe: perspectives from the EU-VAP/CAP study.

We report on intensive care nosocomial pneumonia (NP) in Europe through a review of EU-VAP/CAP manuscripts: a prospective observational study, enrolling patients from 27 ICUs in nine European countries. From 2,436 eligible ICU patients, 827 cases presented NP, with 18.3 episodes of VAP per 1000 ventilator-days. Most common findings were worsening oxygenation, purulent respiratory secretions and temperature increase. At least three criteria from Clinical Pulmonary Infection score (CPIS) were present in 77.9 % of episodes, but only 0.2 % met six CPIS criteria. Diagnosis was confirmed mainly noninvasively (74.8 %), with half qualitative and quantitative cultures. The dominant isolate was S. aureus in Spain, France, Belgium and Ireland, P. aeruginosa in Italy and Portugal, Acinetobacter in Greece and Turkey, but Escherichia coli in Germany. NP resulted in 6 % higher mortality, longer ICU stay and duration of mechanical ventilation (12 and 10 days). COPD and age ≥45 years were not associated with higher VAP incidence but did correlate with increased mortality. Trauma had higher VAP incidence but lower mortality. Bacteremia (led by MRSA and Acinetobacter baumannii) was documented in 14.6 %, being associated with extra ICU stay and mortality. Vasopressors and ICUs with above 25 % prevalence of Potential Resistant Organisms (PRM) were independently associated with PRM, being documented in 50.7 % of patients with early-onset VAP without known risk factors. Most patients initially received combination therapy. Delay in appropriate antimicrobial choice significantly increased mortality, and LOS in survivors was six days longer (p < 0.05). In conclusion, NP management in Europe presents local differences and major shifts when compared to reports from North America, outcomes of randomized trials and general guidelines.

COPD patients with ventilator-associated pneumonia: implications for management.

Data on the occurrence and outcome of patients with chronic obstructive pulmonary disease (COPD) and ventilator-associated pneumonia (VAP) are quite limited. The aim of this study was to determine if COPD intensive care unit (ICU) patients have a higher rate of VAP development, different microbiological aetiology or have worse outcomes than other patients without VAP. A secondary analysis of a large prospective, observational study conducted in 27 European ICUs was carried out. Trauma patients were excluded. Of 2082 intubated patients included in the study, 397 (19.1%) had COPD; 79 (19.9%) patients with COPD and 332 (19.7%) patients without COPD developed VAP. ICU mortality increased by 17% (p < 0.05) when COPD patients developed VAP, remaining an independent predictor of mortality [odds ratio (OR) 2.28; 95% confidence interval (CI) 1.35-3.87]. The development of VAP in COPD patients was associated with a median increase of 12 days in the duration of mechanical ventilation and >13 days in ICU stay (p < 0.05). Pseudomonas aeruginosa was more common in VAP when COPD was present (29.1% vs. 18.7%, p = 0.04) and was the most frequent isolate in COPD patients with early-onset VAP, with a frequency 2.5 times higher than in patients without early-onset VAP (33.3% vs. 13.3%, p = 0.03). COPD patients are not more predisposed to VAP than other ICU patients, but if COPD patients develop VAP, they have a worse outcome. Antibiotic coverage for non-fermenters needs to be included in the empiric therapy of all COPD patients, even in early-onset VAP.

Point prevalence of surgical checklist use in Europe: relationship with hospital mortality.

BACKGROUND: The prevalence of use of the World Health Organization surgical checklist is unknown. The clinical effectiveness of this intervention in improving postoperative outcomes is debated. METHODS: We undertook a retrospective analysis of data describing surgical checklist use from a 7 day cohort study of surgical outcomes in 28 European nations (European Surgical Outcomes Study, EuSOS). The analysis included hospitals recruiting >10 patients and excluding outlier hospitals above the 95th centile for mortality. Multivariate logistic regression and three-level hierarchical generalized mixed models were constructed to explore the relationship between surgical checklist use and hospital mortality. Findings are presented as crude and adjusted odds ratios (ORs) with 95% confidence intervals (CIs). RESULTS: A total of 45 591 patients from 426 hospitals were included in the analysis. A surgical checklist was used in 67.5% patients, with marked variation across countries (0-99.6% of patients). Surgical checklist exposure was associated with lower crude hospital mortality (OR 0.84, CI 0.75-0.94; P=0.002). This effect remained after adjustment for baseline risk factors in a multivariate model (adjusted OR 0.81, CI 0.70-0.94; P<0.005) and strengthened after adjusting for variations within countries and hospitals in a three-level generalized mixed model (adjusted OR 0.71, CI 0.58-0.85; P<0.001). CONCLUSIONS: The use of surgical checklists varies across European nations. Reported use of a checklist was associated with lower mortality. This observation may represent a protective effect of the surgical checklist itself, or alternatively, may be an indirect indicator of the quality of perioperative care. CLINICAL TRIAL REGISTRATION: The European Surgical Outcomes Study is registered with ClinicalTrials.gov, number NCT01203605.

Determinants of prescription and choice of empirical therapy for hospital-acquired and ventilator-associated pneumonia.

The objectives of this study were to assess the determinants of empirical antibiotic choice, prescription patterns and outcomes in patients with hospital-acquired pneumonia (HAP)/ventilator-associated pneumonia (VAP) in Europe. We performed a prospective, observational cohort study in 27 intensive care units (ICUs) from nine European countries. 100 consecutive patients on mechanical ventilation for HAP, on mechanical ventilation>48 h or with VAP were enrolled per ICU. Admission category, sickness severity and Acinetobacter spp. prevalence>10% in pneumonia episodes determined antibiotic empirical choice. Trauma patients were more often prescribed non-anti-Pseudomonas cephalosporins (OR 2.68, 95% CI 1.50-4.78). Surgical patients received less aminoglycosides (OR 0.26, 95% CI 0.14-0.49). A significant correlation (p<0.01) was found between Simplified Acute Physiology Score II score and carbapenem prescription. Basal Acinetobacter spp. prevalence>10% dramatically increased the prescription of carbapenems (OR 3.5, 95% CI 2.0-6.1) and colistin (OR 115.7, 95% CI 6.9-1,930.9). Appropriate empirical antibiotics decreased ICU length of stay by 6 days (26.3±19.8 days versus 32.8±29.4 days; p=0.04). The antibiotics that were prescribed most were carbapenems, piperacillin/tazobactam and quinolones. Median (interquartile range) duration of antibiotic therapy was 9 (6-12) days. Anti-methicillin-resistant Staphylococcus aureus agents were prescribed in 38.4% of VAP episodes. Admission category, sickness severity and basal Acinetobacter prevalence>10% in pneumonia episodes were the major determinants of antibiotic choice at the bedside. Across Europe, carbapenems were the antibiotic most prescribed for HAP/VAP.

Bloodstream infections in ICU with increased resistance: epidemiology and outcomes.

BACKGROUND: Aim of this study was to evaluate the epidemiology and outcomes of hospital-acquired bloodstream infections (HA-BSI) in Greek intensive care units (ICU). METHODS: Secondary analysis of data from 29 ICU collected during the EUROBACT study, a large prospective, observational, multination survey of HA-BSI. First episodes of HA-BSI acquired in the ICU or within 48 hours prior to admission were recorded. RESULTS: Gram-negative bacteria predominated namely Acinetobacter sp, Klebsiella sp, Pseudomonas sp (73.3% of monomicrobial infections) followed by Gram-positive cocci (18.3%); fungi (7.6%) and anaerobes (0.8%). Overall 73.3% of isolates were multidrug resistant (MDR), 47.1% extensively resistant (XDR) and 1.2% pan-drug resistant (PDR). Carbapenems were the most frequent empirically prescribed antibiotics, while colistin was the most frequently adequate; for both, calculated mean total daily doses were suboptimal. Overall 28-day all-cause mortality was 33.3%. In the multivariate analysis, factors adversely affecting outcome were higher SOFA score at HA-BSI onset (OR 1.19; 95% CI 1.08-1.31, P=0.0006), need for renal supportive therapy (OR 2.75; 95% CI 1.35-5.59, P=0.0053), and for vasopressors/inotropes (OR 2.68; CI 1.18-6.12, P=0.02); adequate empirical treatment had a protective effect (OR 0.48; CI 0.24-0.95, P=0.03). CONCLUSION: TIMELY administration of adequately dosed treatment regimens and early ICU admission of critically ill patients could help in improving outcomes.

Risk factors for target non-attainment during empirical treatment with ß-lactam antibiotics in critically ill patients.

PURPOSE: Risk factors for ß-lactam antibiotic underdosing in critically ill patients have not been described in large-scale studies. The objective of this study was to describe pharmacokinetic/pharmacodynamic (PK/PD) target non-attainment envisioning empirical dosing in critically ill patients and considering a worst-case scenario as well as to identify patient characteristics that are associated with target non-attainment. METHODS: This analysis uses data from the DALI study, a prospective, multi-centre pharmacokinetic point-prevalence study. For this analysis, we assumed that these were the concentrations that would be reached during empirical dosing, and calculated target attainment using a hypothetical target minimum inhibitory concentration (MIC), namely the susceptibility breakpoint of the least susceptible organism for which that antibiotic is commonly used. PK/PD targets were free drug concentration maintained above the MIC of the suspected pathogen for at least 50 % and 100 % of the dosing interval respectively (50 % and 100 % f T (>MIC)). Multivariable analysis was performed to identify factors associated with inadequate antibiotic exposure. RESULTS: A total of 343 critically ill patients receiving eight different ß-lactam antibiotics were included. The median (interquartile range) age was 60 (47-73) years, APACHE II score was 18 (13-24). In the hypothetical situation of empirical dosing, antibiotic concentrations remained below the MIC during 50 % and 100 % of the dosing interval in 66 (19.2 %) and 142 (41.4 %) patients respectively. The use of intermittent infusion was significantly associated with increased risk of non-attainment for both targets; creatinine clearance was independently associated with not reaching the 100 % f T( >MIC) target. CONCLUSIONS: This study found that-in empirical dosing and considering a worst--case scenario--19 % and 41 % of the patients would not achieve antibiotic concentrations above the MIC during 50 % and 100 % of the dosing interval. The use of intermittent infusion (compared to extended and continuous infusion) was the main determinant of non-attainment for both targets; increasing creatinine clearance was also associated with not attaining concentrations above the MIC for the whole dosing interval. In the light of this study from 68 ICUs across ten countries, we believe current empiric dosing recommendations for ICU patients are inadequate to effectively cover a broad range of susceptible organisms and need to be reconsidered.

Variability in protein binding of teicoplanin and achievement of therapeutic drug monitoring targets in critically ill patients: lessons from the DALI Study.

The aims of this study were to describe the variability in protein binding of teicoplanin in critically ill patients as well as the number of patients achieving therapeutic target concentrations. This report is part of the multinational pharmacokinetic DALI Study. Patients were sampled on a single day, with blood samples taken both at the midpoint and the end of the dosing interval. Total and unbound teicoplanin concentrations were assayed using validated chromatographic methods. The lower therapeutic range of teicoplanin was defined as total trough concentrations from 10 to 20 mg/L and the higher range as 10-30 mg/L. Thirteen critically ill patients were available for analysis. The following are the median (interquartile range) total and free concentrations (mg/L): midpoint, total 13.6 (11.2-26.0) and free 1.5 (0.7-2.5); trough, total 11.9 (10.2-22.7) and free 1.8 (0.6-2.6). The percentage free teicoplanin for the mid-dose and trough time points was 6.9% (4.5-15.6%) and 8.2% (5.5-16.4%), respectively. The correlation between total and free antibiotic concentrations was moderate for both the midpoint (ρ = 0.79, P = 0.0021) and trough (ρ = 0.63, P = 0.027). Only 42% and 58% of patients were in the lower and higher therapeutic ranges, respectively. In conclusion, use of standard dosing for teicoplanin leads to inappropriate concentrations in a high proportion of critically ill patients. Variability in teicoplanin protein binding is very high, placing significant doubt on the validity of total concentrations for therapeutic drug monitoring in critically ill patients.