Association between hypoalbuminemia and mortality among subjects treated with ertapenem versus other carbapenems: prospective cohort study.

The aim of this study was to determine whether ertapenem, being highly protein bound, is less effective than other carbapenems in the presence of hypoalbuminemia. In a prospective cohort study, we included adults with clinically and microbiologically documented infections caused by carbapenem-susceptible Enterobacteriaceae who were hospitalized in a tertiary medical center from March 2010 to September 2012. We tested whether hypoalbuminemia (serum albumin <2.5 g/dL) had a larger effect on 30-day mortality in subjects treated with ertapenem compared to those treated with meropenem or imipenem (I/M). Logistic regression analysis was used to identify independent risk factors for death including the carbapenem drug and the interaction between albumin and the carbapenem. Of 279 individual subjects included, 173 were treated with ertapenem and 106 with I/M. The odds ratio (OR) for 30-day mortality with hypoalbuminemia was 4.6 (95% confidence interval (CI) 2.1-10.1) among subjects with ertapenem versus 1.2 (95% CI 0.5-2.70) with I/M (p = 0.02 for difference between groups). In the regression model, the interaction between carbapenem type and albumin levels was significant (p = 0.03); for ertapenem lower albumin levels were associated with increased 30-day mortality (OR 2.45, 95% CI 1.19-5.05), while for I/M the change was not significant (OR 0.67, 95% CI 0.31-1.41). The model suggests that the risk of death for ertapenem-treated subjects quintupled when albumin was 2 g/dL compared to 4 g/dL. Hypoalbuminemia was associated with mortality significantly more among subjects treated with ertapenem compared to subjects treated with I/M. The effectiveness of current dosing schemes of ertapenem in subjects with significant hypoalbuminemia should be revisited.

Factors associated with influenza vaccination among adult cancer patients: a case-control study.

Influenza vaccination is recommended for cancer patients; however, adherence is low. We aimed to identify predictive factors for vaccination among cancer patients. We conducted a case-control analysis of a patient cohort in the 2010-2011 influenza season. We included adult cancer patients with solid malignancies undergoing chemotherapy, and haematological patients with active disease. Patients who died between October and November 2010 (N = 43) were excluded from analysis. Cases received the 2011 seasonal influenza vaccine, and controls did not. Data were obtained from patients' records, and validated through personal interviews. We collected socio-demographic information, and data on the malignancy and co-morbidities and triggers for vaccination and non-vaccination. We performed bivariate and multivariable analyses, in which vaccination status was the dependent variable. Of 806 patients included in analysis, 387 (48%) were vaccinated. Variables associated with vaccination on bivariate analysis were older age, higher socio-economic status, lower crowding index, marital status (widowed > married > single), malignancy type (haematological > solid tumours) and time from diagnosis, low-risk malignancy, diabetes, past vaccination, country of birth (non-Russian origin), and physicians' recommendations. Predictive factors found to be independently associated with vaccination on multivariable analysis were past vaccinations, low-risk malignancy, and country of birth. In the analysis conducted among interviewees (N = 561), recommendations from the oncologist (OR 10.7, 95% CI 5.4-21.2) and from the primary-care physician (OR 3.35, 95% CI 2.05-5.49) were strong predictors for vaccination. We conclude that 'habitual vaccinees' continue influenza vaccinations when ill with cancer. Physicians' recommendations, especially the oncologist's, have a major influence on patients' compliance with influenza vaccination.

Macrolides vs. quinolones for community-acquired pneumonia: meta-analysis of randomized controlled trials.

The relative efficacy, safety and ecological implications of macrolides vs. quinolones in the treatment of community-acquired pneumonia (CAP) are debatable. We performed a systematic review and meta-analysis of randomized controlled trials comparing any macrolide vs. any quinolone for the treatment of CAP among adult inpatients or outpatients, as monotherapy or both in combination with a beta-lactam. We did not limit inclusion by pneumonia severity, publication status, language or date of publication. The primary outcomes assessed were 30-day all-cause mortality and treatment failure. Two authors independently extracted the data. Fixed effect meta-analysis of risk ratios (RRs) with 95% confidence intervals was performed. Sixteen trials (4989 patients) fulfilling inclusion criteria were identified, mostly assessing outpatients with mild to moderate CAP. All-cause mortality was not significantly different for macrolides vs. quinolones, RR 1.03 (0.63-1.68, seven trials), with a low event rate (2%). Treatment failure was significantly lower with quinolones, RR 0.78 (0.67-0.91, 16 trials). The definition of failure used in the primary studies was not clearly representative of patients' benefit. Microbiological failure was lower with quinolones, RR 0.63 (0.49-0.81, 13 trials). All adverse events, adverse events requiring discontinuation and any premature antibiotic discontinuation were significantly more frequent with macrolides, mainly on account of gastrointestinal adverse events. Resistance development was not assessed in the trials. Randomized controlled trials show an advantage of quinolones in the treatment of CAP with regard to clinical cure without need for antibiotic modification at end of treatment and gastrointestinal adverse events. The clinical significance of this advantage is unclear.