Our study shows a HD TGC (200 mg LD, then 100 mg q12) time-curve concentration with mean peak and trough levels of 0.65 mg/mL and 0.25 mg/mL, respectively (Fig. 1). AUC0 − 24/MIC targets for nosocomial pneumonia (≥ 4.5) and complicated intra-abdominal infections (≥ 6.96) were obtained in the majority of cases in presence of bacteria with MIC values ≤ 0.25. Otherwise lower MIC values (≤ 0.12 mcg/mL) were required to have satisfactory AUC0 − 24/MIC results (78%), while treating a skin/soft tissue infection (Fig. 2, Table 2). Similar to plasma 1 h and 12 h, pulmonary concentrations (0.78 mg/L and 0.36 mg/L, respectively) were observed with a good median ELF/plasma ratio of 152.9% (Table 2, Fig. 3). This high-dose regimen was associated with a 65.6% of treatment success rate in a normal weight population including 60% of VAP, 31% of cIAI and 9% of SSTI. TGC was used in half of the cases as targeted regimen for a median duration of 12 days. The rate of septic shock, acute respiratory failure requiring MV and acute kidney injury requiring CRRT was also high, with a mortality rate of 28.1% (Table 1).
The pharmacokinetics/pharmacodynamics and tissue penetration of tigecycline have been extensively studied in various in vitro and human models [17]. However, these studies were generally carried out in healthy volunteers, and few pharmacokinetic data concerning infected patients are available, which may present pathophysiologic conditions influencing the pharmacokinetic profile of this molecule. Additionally, the majority of available data in infected patients derive from studies where normal doses are used, although for severe nosocomial infections a double-dose regimen is warranted [18, 19]
Recently, standard dose TGC pharmacokinetics in ten critically ill patients have been studied [6]. The authors observed that a larger body mass index was associated with increased TGC Cl, but standard doses produced satisfactory plasmatic levels for VAP and cIAI treatment due to Enterobacter cloacae, Esherichia coli, Klebsiella pneumoniae and methicillin-resistant Staphylococcus aureus. However higher dosages were required for the treatment of SSTI, especially in obese patients.
Eleven out of 32 patients in our cohort were receiving CRRT while being treated with high-dose TGC. Interestingly, in a recent paper, Broeker and cow. [20] described the PK/PD of standard dose TGC in eleven patients on continuous veno-venous hemodialysis (CVVHD) or hemodiafiltration (CVVHDF). TGC dialysability, as expressed by saturation coefficients (0.79 and 0.9 for CVVHD and CVVHDF, respectively) was very high, but the contribution of CRRT TGC clearance was minimal (about 2 L/h), compared with the total body clearance (18.3L/h). Peak drug concentrations were below 1 mg/L and trough levels about 0.2 mg/L. The authors, considering the AUC0-24/MIC referral value for cIAI (6.96), observed that such target was accomplished in 88% of the case if MIC was ≤ 0.5.
Indeed, our results are in line with current available data, underlying the plus-value of increased dosages while treating critically ill patients especially with severe cIAI and SSTI. In addition, there is a high need of PK/PD data on TGC administered at higher than approved dosages, in light of the wide spread of increased resistance to TGC among Gram-negative rods and Acinetobacter spp. The first investigation on PK/PD of HD TGC derives from Ramirez et al who conducted a randomized phase 2 trial to evaluate the clinical efficacy of two high-dosage regimen of TGC (75 mg bid and 100 mg bid) versus imipenem-cilastatin for the treatment of nosocomial pneumonia[8]. In the clinically evaluable population, clinical cure with TGC 100 mg bid was higher than with 75 bid and imipenem-cilastatin (85% vs. 69.6% vs. 75%). Mean peak TGC concentration was about 1 mg/mL, declining to less than 0.5 mg/ml after 8 hours, observing a safety profile comparable to that one known for the approved those. The only other study investigating the PK/PD of HD TGC profile was conducted by Borsuk-De Moor et al in 37 ICU patients with severe infections [21]. The time-concentration curve was similar to our data, displaying a peak concentration about 1 mg/L and 12 h level below 0.5 mg/L. Interestingly, the authors developed a model which showed that no individual covariates may influence target concentrations, advising to modify TGC daily dosage according to pathogens type, susceptibility pattern and PK targets.
Tissue concentrations of antibiotics at the target site contribute to therapeutic effects: using plasma concentrations may frequently overestimate the target site concentrations and therefore clinical efficacy. This is the first study to report steady-state ELF percentage penetration of TGC administered 100 mg q12 after 200 mg LD. Considering the AUC0 − 24/MIC target of 4.96, our data shows satisfactory pulmonary concentrations with potential clinical success in 100% − 94% − 75% − 41% of the cases treating bacteria with MIC of 0.12–0.25 mg/L – 0.5–1 mg/L, respectively [Fig. 2]. These data confirm the results observed in healthy subjects by Conte et al., where the Cmax/MIC90, AUC/MIC90 ratios, T > MIC90 and extended serum and intrapulmonary half-lives following the standard regimen are favourable for the treatment of TGC-susceptible pulmonary infections [22]. Penetration ratio may be even higher when in presence of infected lungs. Crandon et al. demonstrated in infected and non-infected mice lungs that the baseline penetration ratio of 8.1 is incremented to 23.3 in case of Acinetobacter pneumonia [23]. Conversely, the majority of lung penetration occurs in alveolar cells, than in ELF, as suggested by Welte et al. in three cases of MDR lung infections [24]. Finally in a recent study on 58 healthy subjects treated with standard TGC dose, the ratio of ELF and AUC to total plasma concentration of tigecycline was 1.71 and 20.8, respectively [25].
Our study has several limitations. First we adopted a single high-dose of tigecycline and we do not know if even higher dosages may result in better PK/PD profiles. Second we measured only pulmonary tissue concentration trough ELF collection and we can only postulate the real tissue/plasma ratio for cIAI and SSTI. Third, our analysis focused on total TGC concentration rather unbound AUC0 − 24, due to the lack of clinical reliable breakpoint of fAUC0 − 24/MIC90. Finally the sample size may be likely responsible of an under-estimated interindividual variability in the observed PK/PD profile.