III-53 Jérémy Seurat

Rifampin co-administration with clindamycin can result in failure to reach the PKPD targets in staphylococcal osteoarticular infections

Tiphaine Goulenok (1), Jérémy Seurat (2), Andréa De la Selle (3), Vincent Jullien (4), Véronique Leflon-Guibout (5), Nathalie Grall (2), François-Xavier Lescure (2), Raphaël Lepeule (6), Julie Bertrand (2), Bruno Fantin (2), Charles Burdet (2)*, Agnès Lefort (2)*

(1) Département de Médecine Interne, Hôpital Bichat, Assistance Publique Hôpitaux de Paris, Université de Paris, France (2) Université de Paris, INSERM, IAME, Paris F-75006, France (3) Service de Médecine Interne, Centre Hospitalier Universitaire Henri-Mondor, Assistance Publique-Hôpitaux de Paris, Université Paris-Est Créteil (UPEC), Créteil, France (4) Département de Pharmacologie, Hôpital Avicenne, Assistance Publique Hôpitaux de Paris, Bobigny, France (5) Département de Bactériologie, Hôpital Beaujon, Assistance Publique Hôpitaux de Paris, Clichy, France (6) Département de Microbiologie, Centre Hospitalier Universitaire Henri-Mondor, Assistance Publique-Hôpitaux de Paris, Université Paris-Est Créteil (UPEC), Créteil, France *: equal contribution

Objectives:

In staphylococcal osteoarticular infections (OAI), the oral combination of clindamycin and rifampin, is a relevant choice [1,2]. However, rifampin induces CYP3A4, which suggests a pharmacokinetic interaction with clindamycin [3]. The aim of the study was to describe the pharmacokinetics of clindamycin before and during rifampin co-administration, to estimate the rifampin effect on pharmacokinetics parameters and on the probability to reach the minimal inhibitory concentrations (MIC)-based PKPD targets.

Methods:

A clinical trial (NCT02782078) was conducted in three French hospitals between 2017 and 2018, in adult patients with staphylococcal OAI. After an initial IV antibiotic therapy, oral antibiotic therapy was started with clindamycin (1800mg, low dose, or 2400mg, high dose, in 3 daily administrations), followed by the addition of rifampin 36 hours after clindamycin introduction. Clindamycin plasmatic concentrations were measured on day 2 (without rifampin) and on day 14 (with rifampin): just before (0) and at 1, 2, 4, 6 hours after administration, each patient being his own control.

Population pharmacokinetic analysis was performed using the Stochastic Approximation Expectation Minimization algorithm of Monolix v2020R1 [4]. We developed a model of clindamycin pharmacokinetics. Model selection (structural, random and error models) was based on the corrected Bayesian information criteria (BICc) and the goodness of fits plots. Rifampin was investigated as a binary and varying across occasion covariate effect, using a backward selection based on the Wald test (p < 0.05). Evaluation of the final model was conducted using NPDE and prediction-corrected VPC.

Using empirical Bayes estimates and individual MIC, we derived PKPD markers: , residual concentrations , along with the fraction of time for which the concentration is above the MIC between two administrations %fT>MIC. These were compared with and without rifampin co-administration using paired Wilcoxon test (p < 0.05).

Then, we simulated clindamycin plasma concentration for 1000 individuals, for each dosing regimen (low or high), with and without rifampin. For various MIC values, we determined the proportion of individuals achieving the following PK/PD targets:  above 2, %fT>MIC = 100%, and  above 60 or above 120 [5].

Results:

Clindamycin concentration data were available for 19 patients. Median (min; max) observed residual concentrations were 2.7 (0.3; 8.9) mg/L before rifampin administration and <0.05, i.e. BLQ, mg/L (<0.05; 0.3) during rifampin administration.

Clindamycin kinetics in plasma at steady state (t1/2 = 3h) was adequately described by a two-compartment model, with a first order absorption and elimination. Rifampin co-administration increased the clindamycin clearance by a factor 15 (relative standard error (rse): 5%) and the intercompartmental clearance (Q) by 3 (rse: 36%). Rifampin co-administration reduced the clindamycin  by a factor 14 (p = 4e-6). Moreover,  and %fT>MIC were significantly lowered by rifampin co-administration (medians of 65.7 vs. 0.3 and 100% vs. 83% respectively).

From simulations, we determined that against a susceptible strain (MIC = 0.25 mg/L), more than 90% of individuals reached all the proposed PKPD targets without rifampin co-administration, even at the low dose. Against the same strain, with rifampin co-administration, for a low/high dose of clindamycin, the proportion reaching %fT>MIC = 100% and AUC24/MIC > 60 dropped to 2%/4% and 15%/25%, respectively. Without rifampin, a high dose of clindamycin increased the probability to reach PKPD targets against an intermediate strain (MIC = 1 mg/L), with  > 2 for 58% of the simulated patients vs. 46% for a low dose. For the same strain, 0% of individuals reached any of the 4 proposed PKPD targets when co-administered with rifampin.

Conclusion:

Our study quantifies the pharmacokinetic interaction of rifampin with clindamycin in OAI. However, the sample size was low and we cannot propose an optimal clindamycin dose in combination with rifampin. Our results nevertheless suggest that a high dose of clindamycin would be preferable in case of intermediary strain. Another study showed that IV clindamycin helps to reach PKPD when co-administered with rifampin [6]. This work is a first step, the PK data of rifampin will also be exploited in a semi-mechanistic model to develop a global pharmacodynamic interaction model [7].

References:

  1. Courjon J, Demonchy E, Cua E, Bernard E, Roger P-M. Efficacy and safety of clindamycin-based treatment for bone and joint infections: a cohort study. Eur J Clin Microbiol Infect Dis. 2017;36:2513–8.
  2. Muller-Serieys C, Saleh Mghir A, Massias L, Fantin B. Bactericidal activity of the combination of levofloxacin with rifampin in experimental prosthetic knee infection in rabbits due to methicillin-susceptible Staphylococcus aureus. Antimicrob Agents Chemother. 2009;53:2145–8.
  3. Niemi M, Backman JT, Fromm MF, Neuvonen PJ, Kivistö KT. Pharmacokinetic interactions with rifampicin : clinical relevance. Clin Pharmacokinet. 2003;42:819–50.
  4. Kuhn E, Lavielle M. Maximum likelihood estimation in nonlinear mixed effects models. Comput Stat Data Anal. 2005;49:1020–38.
  5. Curis E, Pestre V, Jullien V, Eyrolle L, Archambeau D, Morand P, et al. Pharmacokinetic variability of clindamycin and influence of rifampicin on clindamycin concentration in patients with bone and joint infections. Infection. 2015;43:473–81.
  6. Zeller V, Magreault S, Heym B, Salmon D, Kitzis M-D, Billaud E, et al. Influence of the clindamycin administration route on the magnitude of clindamycin-rifampicin interaction: a prospective pharmacokinetic study. Clin Microbiol Infect. 2021;27:1857.e1-e7.
  7. Wicha SG, Chen C, Clewe O, Simonsson USH. A general pharmacodynamic interaction model identifies perpetrators and victims in drug interactions. Nat Commun. 2017;8:2129.

Reference: PAGE 30 (2022) Abstr 10185 [www.page-meeting.org/?abstract=10185]

Poster: Drug/Disease Modelling - Infection

PDF poster / presentation (click to open)