Abstract by Sara Thønnings

Bacteraemia is the presence of viable bacteria in the bloodstream as evidenced by growth in blood cultures where contamination has been ruled out. In the clinical setting bacteraemia is an invasion of the bloodstream by bacteria either due to a local infection where the host has not managed to control bacterial spread or due to breakage in the natural barriers. From population-based studies, the incidence of bacteraemia in Denmark is 166-216 cases pr. 100,000 person-years, which is similar to the incidence reported in other western countries. With an increase in both short-term and long-term mortality in bacteraemia patients compared to control cases, and an estimated 30-day mortality rate ranging from 16-22% in case of bacteraemia, it is still a major cause of death and a burden to Danish society. Beta-lactam antibiotics are the cornerstone in bacteraemia treatment. Hence, beta-lactam antibiotic therapy is an obvious target for optimisation to improve outcome of bacteraemia. Beta-lactam antibiotic dosing regimens can be investigated in well-defined patient categories using pharmacokinetic and pharmacodynamic models.

This thesis focuses on optimising the treatment of bacteraemia, specifically with beta-lactam antibiotics. The objectives were to 1) determine clinical characteristics and mortality in defined bacteraemia cohorts with different bacterial species and different patient populations (Paper I and II); 2) compare different beta-lactam antibiotic treatments of specific bacteria and their association with mortality (Paper I and II); and 3) investigate the pharmacokinetics and pharmacodynamics of a specific beta-lactam antibiotic and how these correlate to clinical use (Paper III).

This was investigated in three papers with the following methodology and findings:

Paper I investigates clinical characteristics and risk factors of death from Listeria monocytogenes bacteraemia/meningitis, especially focusing on empiric and definitive antibiotic therapy. The design was retrospective observational and included data from medical records, microbiology databases and the Danish Civil Registration System. The study included 229 patients with L. monocytogenes bacteraemia/meningitis. Significant risk factors for death were inadequate empiric antibiotic therapy, septic shock and altered mental state. Definitive antibiotic treatment with benzylpenicillin or aminopenicillin resulted in a lower mortality compared to meropenem.

Paper II investigates the difference in mortality between patients treated with piperacillin/tazobactam (PTZ) and cefuroxime empirically in an E. coli bacteraemia cohort. The design was retrospective observational and included data from the Danish Collaborative Bacteraemia Network (DACOBAN) and the Danish Civil Registration System. The study included 628 E. coli bacteraemia patients, and in the Cox proportional hazard model found that empiric antibiotic treatment with cefuroxime compared to PTZ was significantly associated with death. Other variables significantly associated with death were health care related infection, hospital-acquired infection, admission at intensive care unit and combination therapy with ciprofloxacin.

Paper III investigates the pharmacokinetics of cefuroxime in healthy volunteers. Monte Carlo simulation was used to evaluate dosage and effect on E. coli,K. pneumoniae,S. aureusandS. pneumoniae with PK/PD analysis. The study was an observational in vivo experiment in healthy volunteers where blood samples were collected after cefuroxime infusion. Based on these test subjects and their serum cefuroxime concentration a population pharmacokinetic model was made. Serum protein binding of cefuroxime and the bacteria’s MIC breakpoints were also established and included in the model. The study included 20 healthy volunteers studied on two separate days with two different doses of cefuroxime (750 mg and 1500 mg) with a target value of time above MIC (T>MIC) greater than 50%, S. pneumoniae is most likely treatable with a two-daily dose of 750 mg cefuroxime. Not treatable are K. pneumoniae and E. coli, where the PTA was <90% even with the highest, simulated dosage of 1500 mg q6h. For S. aureus, 1500 mg q8h constitutes an optimal dosing schedule.

In conclusion, we found that mortality in case of bacteraemia was high, likely due to inadequate effect of both the empiric and the definitive antibiotic treatment of bacteraemia. In case of L. monocytogenes bacteraemia/meningitis inadequate empiric treatment was associated with higher mortality as well as definitive antibiotic therapy with meropenem compared to benzylpenicillin or aminopenicillins. When it came to E. coli bacteraemia, empiric treatment with cefuroxime was associated with increased mortality compared to empiric treatment with PTZ. We demonstrated that cefuroxime appears not to cover E. coli and K. pneumoniae sufficiently with the standard dosage regime of 1500 mg q8h in healthy volunteers and conclude that cefuroxime should not be recommended for treatment of serious infections with E. coli when the renal function is normal or increased. Hence the importance of identifying patients at risk of bacteraemia and adjusting empiric therapy to risk factors is essential. Moreover, definitive antibiotic therapy needs to be adjusted specifically to the pathogen, patient category and the clinical situation. This also applies for broad-spectrum antibiotics. Increased awareness should be addressed to patients with preserved renal function when treating serious infections with Gram-negative pathogens such as E. coli and K. pneumoniae.