Third-generation cephalosporin resistant gram-negative bacteraemia in patients with haematological malignancy; an 11-year multi-centre retrospective study.

OBJECTIVES: Among patients with haematological malignancy, bacteraemia is a common complication during chemotherapy-induced neutropenia. Resistance of gram-negative bacteria (GNB) to third-generation cephalosporins (3GC) is increasing. In order to explore the value of using surveillance cultures to guide empirical treatment e.g. choosing between carbapenem versus ceftazidime- we aimed to assess the distribution of pathogens causing bacteraemia in patients with haematological malignancy, and the proportion of 3GC-resistant GNB (3GC-R GNB) bacteraemia that was preceded by 3GC-R GNB colonization. METHODS: Using 11 years of data (2008-2018) from the Dutch national antimicrobial resistance surveillance system, we assessed the prevalence of 3GC-R GNB in episodes of bacteraemia, and the proportion of 3GC-R GNB bacteraemia that was preceded by 3GC-R GNB colonization. Colonization was defined as availability of any GNB surveillance isolate in the year before, independent of the causative micro-organism (time-paired isolates). RESULTS: We included 3887 patients, representing 4142 episodes of bacteraemia. GNB were identified in 715/4142 (17.3%), of which 221 (30.9%) were 3GC-R GNB. In 139 of these 221 patients a time-paired surveillance culture was available. In 76.2% (106/139) of patients these surveillance cultures already showed 3GC-R GNB isolates in the year prior to the culture date of the 3GC-R GNB positive blood isolate. CONCLUSIONS: This multi-centre study shows that in patients with haematological malignancy, the majority of 3GC-R GNB bacteraemia is preceded by 3GC-R GNB colonization. Prospective clinical studies are needed to assess the safety and benefits of the use of surveillance-cultures to guide empirical therapy to restrict the empirical use of carbapenems in this population.

Structural Antibiotic Surveillance and Stewardship via Indication-Linked Quality Indicators: Pilot in Dutch Primary Care.

Insight into antibiotic prescribing quality is key to general practitioners (GPs) to improve their prescribing behavior and to national antibiotic surveillance and stewardship programs. Additionally to numbers of prescribed antibiotics, quality indicators (QIs) linked to the clinical indication for prescribing are urgently needed. The aim of this proof of concept study was to define indication-linked QIs which can be easily implemented in Dutch primary care by collaborating with data-extraction/processing companies that routinely process patient data for GP practices. An expert group of academic and practicing GPs defined indication-linked QIs for which outcomes can be derived from routine care data. QI outcomes were calculated and fed back to GPs from 44 practices, associations between QI outcomes were determined, and GPs' opinions and suggestions with respect to the new set were captured using an online questionnaire. The new set comprises: (1) total number of prescribed antibiotics per 1000 registered patients and percentages of generally non-1st choice antibiotics; (2) prescribing percentages for episodes of upper and lower respiratory tract infection; (3) 1st choice prescribing for episodes of tonsillitis, pneumonia and cystitis in women. Large inter-practice variation in QI outcomes was found. The validity of the QI outcomes was confirmed by associations that were expected. The new set was highly appreciated by GPs and additional QIs were suggested. We conclude that it proved feasible to provide GPs with informative, indication-linked feedback of their antibiotic prescribing quality by collaborating with established data extraction/processing companies. Based on GPs' suggestions the set will be refined and extended and used in the near future as yearly feedback with benchmarking for GPs and for national surveillance and stewardship purposes.