Flucloxacillin concentrations and clinical outcomes in Staphylococcus aureus bacteraemia, and the impact of renal impairment on flucloxacillin pharmacokinetics

Abstract

Background: Methicillin sensitive staphylococcus aureus bacteraemia (SAB) is a common cause of bacteraemia with substantial associated mortality. Flucloxacillin is one of the first line antibiotics used in the treatment of this condition. The relationship between serum flucloxacillin concentrations below the minimum inhibitory concentration (MIC) of the pathogen and clearance of bacteraemia based on blood cultures remains uncertain.

In addition, flucloxacillin is renally cleared by active tubular secretion, however the current Australian dosing guidelines do not recommend dose adjustment until the eGFR is <10 ml/min/1.73m2. We investigated the impact of estimated glomerular filtration rate (eGFR) on the serum flucloxacillin concentration.

Methods: Two retrospective observational studies are presented. In the first study, 138 episodes of SAB between January 2015 and December 2018 were analysed. Inclusion criteria were age > 16 years and a positive blood culture for methicillin sensitive staphylococcus aureus. Of these, there were 30 and 41 episodes in which appropriately timed trough and mid-dosage estimated free flucloxacillin concentrations [FLX]efree, respectively, were obtained within 7 days of the initial MSSA blood culture. Differences between the persistent (blood cultures positive >72 hours) and non-persistent (<72 hours) bacteraemia groups in age, sex, focus of infection and risk factors for complicated clinical course were determined.

In the second study, we analysed 396 serum flucloxacillin concentrations [FLX]total, restricted to the first per hospital admission or outpatient encounter and those simultaneously co-measured with creatinine. Inclusion criteria were age > 16 years and a serum flucloxacillin concentration measured at the study institution between May 2015 and June 2020. Free flucloxacillin concentrations [FLX]efree were estimated from total concentrations using a Michaelis-Menten model for serum albumin binding. We analysed the proportion of episodes with [FLX]efree below the upper limit oxacillin MIC for methicillin-susceptible Staphylococcus aureus (1mg/L) in this cohort, and [FLX]total>125.1mg/L - a previously identified neurotoxicity threshold, respectively, by GFR quintile.

Results: In the first study, the persistent and non-persistent bacteraemia groups were found to be similar in age, sex and types of infection. Sub-MIC trough concentrations were identified in 3/17 cases of persistent bacteraemia, as compared with 0/13 cases of non-persistent bacteraemia (OR 0.0; 95% CI 0.0 to 1.4; p=0.24). Sub-MIC mid-dosage concentrations were identified in 3/22 cases of persistent bacteraemia, as compared with 0/19 cases of non-persistent bacteraemia (OR 0.0; 95% CI 0.0 to 1.28; p=0.24).

In the second study, of the 396 [FLX]total samples, 242 [64.7%]) were from hospital ward inpatients, 70 (18.7%) samples were obtained from patients in ICU or HDU care, while 62 (16.6%) were from outpatients. In a multivariable regression analysis, [FLXfree] negatively correlated with eGFR. Flucloxacillin concentrations stratified by GFR quintiles were determined. The proportion of episodes with [FLXfree]<1 mg/L was significantly lower in the lower GFR quintiles when compared with the higher GFR quintile (p<0.001 for trend). Conversely, [FLX] above a previously identified neurotoxicity threshold was found to be significantly more common in lower GFR quintiles as compared with higher GFR quintiles (p<0.001 for trend).

Conclusions: Sub-MIC trough or mid-dosage estimated free flucloxacillin concentrations were not associated with a lower rate of persistent bacteraemia, however the wide confidence intervals include a clinically important effect, and larger studies are warranted.

Flucloxacillin concentrations were significantly negatively correlated with eGFR. Potentially subtherapeutic concentrations were more common in patients in higher eGFR quintiles, while potentially neurotoxic concentrations were more common in patients in lower eGFR quintiles but above the current threshold for dose adjustment.