March 29, 2020 Journal article Open Access

Attributable mortality of antibiotic resistance in Gram-negative infections in the Netherlands: a parallel matched cohort study

J.W. Timotëus Deelen; Jan Kluytmans; Anton Buiting; Wouter C. Rottier; Marc J. M. Bonten; Bart J. M. Vlaminckx; Steven F. T. Thijsen; Paul D. van der Linden; J. Wendelien Dorigo-Zetsma; Heidi S.M. Ammerlaan; Annemarie J. L. Weersink; Giorgia Caruana; Giorgia Caruana

Abstract Objectives Antibiotic resistance in Gram-negative bacteria has been associated with increased mortality. This was demonstrated mostly for third-generation cephalosporin-resistant (3GC-R) Enterobacterales bacteraemia in international studies. Yet, the burden of resistance specifically in the Netherlands and created by all types of Gram-negative infection has not been quantified. We therefore investigated the attributable mortality of antibiotic resistance in Gram-negative infections in the Netherlands. Methods In eight hospitals, a sample of Gram-negative infections was identified between 2013 and 2016, and separated into resistant and susceptible infection cohorts. Both cohorts were matched 1:1 to non-infected control patients on hospital, length of stay at infection onset, and age. In this parallel matched cohort set-up, 30-day mortality was compared between infected and non-infected patients. The impact of resistance was then assessed by dividing the two separate risk ratios (RRs) for mortality attributable to Gram-negative infection. Results We identified 1,954 Gram-negative infections, of which 1,190 (61%) involved Escherichia coli, 210 (11%) Pseudomonas aeruginosa, and 758 (39%) bacteraemia. Resistant Gram-negatives caused 243 infections (12%; 189 (78%) 3GC-R Enterobacterales, 9 (4%) multidrug-resistant P. aeruginosa, no carbapenemase-producing Enterobacterales). Subsequently, we matched 1,941 non-infected controls. After adjustment, point estimates for RRs comparing mortality between infections and controls were similarly higher than 1 in case of resistant infections and susceptible infections (1.42 (95% confidence interval 0.66-3.09) and 1.32 (1.06-1.65), respectively). By dividing these, the RR reflecting attributable mortality of resistance was calculated as 1.08 (0.48-2.41). Conclusions In the Netherlands, antibiotic resistance did not increase 30-day mortality in Gram-negative infections.