ABSTRACT
Adults with cystic fibrosis (CF) frequently harbor Staphylococcus aureus, which is increasingly antibiotic resistant. Telavancin is a once-daily rapidly bactericidal antibiotic active against methicillin-, linezolid-, and ceftaroline-resistant S. aureus. Because CF patients experience alterations in pharmacokinetics, the optimal dose of telavancin in this population is unknown. Adult CF patients (n = 18) admitted for exacerbations received 3 doses of telavancin 7.5 mg/kg of body weight (first 6 patients) or 10 mg/kg (final 12 patients) every 24 h (q24h). Population pharmacokinetic models with and without covariates were fitted using the nonparametric adaptive grid algorithm in Pmetrics. The final model was used to perform 5,000-patient Monte Carlo simulations for multiple telavancin doses. The best fit was a 2-compartment model describing the volume of distribution of the central compartment (Vc) as a multiple of total body weight (TBW) and the volume of distribution of the central compartment scaled to total body weight (Vθ) normalized by the median observed value (Vc = Vθ × TBW/52.1) and total body clearance (CL) as a linear function of creatinine clearance (CRCL) (CL = CLNR + CLθ × CRCL), where CLNR represents nonrenal clearance and CLθ represents the slope term on CRCL to estimate renal clearance. The mean population parameters were as follows: Vθ, 4.92 ± 0.76 liters · kg−1; CLNR, 0.59 ± 0.30 liters · h−1; CLθ, 5.97 × 10−3 ± 1.24 × 10−3; Vp (volume of the peripheral compartment), 3.77 ± 1.41 liters; Q (intercompartmental clearance), 4.08 ± 2.17 liters · h−1. The free area under the concentration-time curve (fAUC) values for 7.5 and 10 mg/kg were 30 ± 4.6 and 52 ± 12 mg · h/liter, respectively. Doses of 7.5 mg/kg and 10 mg/kg achieved 76.5% and 100% probability of target attainment (PTA) at a fAUC/MIC threshold of >215, respectively, for MIC of ≤0.12 mg/liter. The probabilities of reaching the acute kidney injury (AKI) threshold AUC (763 mg · h · liter−1) for these doses were 0% and 0.96%, respectively. No serious adverse events occurred. Telavancin 10 mg/kg yielded optimal PTA and minimal risk of AKI, suggesting that this FDA-approved dose is appropriate to treat acute pulmonary exacerbations in CF adults. (The clinical trial discussed in this study has been registered at ClinicalTrials.gov under identifier NCT03172793.)
TEXT
Infection and colonization by Staphylococcus aureus is common in patients with cystic fibrosis (CF). In children, S. aureus is the most prevalent bacterial pathogen identified. During adulthood, it remains the second most common pathogen isolated from the CF lung and is present in approximately 70% of respiratory cultures in CF patients across all ages. The incidence of methicillin-resistant S. aureus (MRSA) in CF patients, however, has been steadily increasing over the past 2 decades. According to the Cystic Fibrosis Foundation 2014 patient registry annual report, 9.2% of positive respiratory specimen cultures were due to MRSA in 2002, but as of 2017, this value had increased to 25.9% (1). The highest prevalence of MRSA appears to occur in patients aged 10 to 30 years; however, these rates are expected to shift as CF patients live longer. Until recently, it was unknown what influence this increased incidence had on clinical outcomes. Pediatric patients with respiratory specimen cultures persistently positive for MRSA have significant reductions in lung function (2). Furthermore, the chronic presence of MRSA-positive cultures for CF patients of any age is associated with decreased survival, especially in coinfections with Pseudomonas aeruginosa (3, 4).
Current treatment options for MRSA exacerbations in CF patients predominantly include vancomycin, linezolid, trimethoprim-sulfamethoxazole, doxycycline, minocycline, and, more recently, ceftaroline fosamil (5, 6). All of the intravenous (i.v.) antibiotic options must be dosed multiple times daily, making outpatient treatment challenging. Additionally, resistance to linezolid and ceftaroline fosamil has been documented in CF patients during treatment (7–9), and vancomycin can be difficult to tolerate due to infusion-related adverse events (10). Thus, alternative intravenous treatment options are needed.
Telavancin is a rapidly bactericidal, once-daily, injectable antibiotic with activity against clinically important Gram-positive pathogens, including methicillin-susceptible and methicillin-resistant S. aureus. It is approved in the United States for treatment of adults with complicated skin and skin structure infections and hospital-acquired bacterial pneumonia and ventilator-associated bacterial pneumonia (HABP and VABP, respectively). Importantly, telavancin retains activity against linezolid- and ceftaroline-resistant S. aureus and is largely unaffected by mechanisms that produce heterogeneous vancomycin-intermediate S. aureus (hVISA) and vancomycin-intermediate S. aureus (VISA) phenotypes (11).
Telavancin has linear, predictable pharmacokinetics (PK) in the adult population. Its half-life is approximately 8 h in adults with normal kidney function, which permits once-daily dosing (12). In addition, epithelial lining fluid (ELF) penetration is 73%, which is appropriate for treatment of pulmonary infections (13, 14). The pharmacokinetics in adults with CF, however, is unknown. CF patients have been shown to eliminate some antibiotics more rapidly than non-CF patients, presumably due to increases in glomerular filtration secondary to a higher resting energy expenditure or a primary renal tubular transport defect (15). Telavancin dosing is also based on total body weight (TBW), and CF patients are typically smaller in stature and have lower fat mass than non-CF patients due to nutritional deficiencies.
Here, the plasma pharmacokinetics characteristics of telavancin in adult patients with cystic fibrosis currently experiencing an acute pulmonary exacerbation were investigated. A dose escalation design was used to determine the dose of telavancin which maximizes the probability of target attainment (PTA) while minimizing the potential of reaching a threshold of exposure associated with nephrotoxicity, which is the major dose-limiting toxicity of telavancin.
RESULTS
Twenty adults with CF admitted for an acute pulmonary exacerbation were enrolled in this multicenter, open-label, dose escalation, pharmacokinetic study. Two patients withdrew consent prior to receiving the first dose of telavancin and were excluded from the analysis. The remaining 18 patients had a median age of 24 years and TBW of 52.1 kg; additional patient characteristics are reported in Table 1. The majority of the participants received three intravenous doses (1-h infusion) of telavancin based on TBW every 24 h to reach steady state before intensive blood sampling. Two patients included in the analysis were discharged from the hospital before receiving all three telavancin doses and therefore contributed only peak concentrations or trough concentrations or both following the initial doses.
Baseline characteristics for 18 adult participants with CFa
Population pharmacokinetic analysis.Plasma samples were drawn at the peak and trough of each dose and at 6 additional time points after the final dose, yielding 179 samples for analysis (Fig. 1). The first cohort of 6 patients received doses of 7.5 mg/kg of body weight and had a free area under the concentration-time curve (fAUC) value of 30 ± 4.6 mg · h · liter−1 (mean ± standard deviation), whereas the final cohort of 12 patients received doses of 10 mg/kg and had fAUC of 52 ± 12 mg · h · liter−1.
Observed total telavancin concentrations in 18 adult patients with CF after doses of 7.5 mg/kg (n = 6) or 10 mg/kg (n = 12) and associated fAUC, assuming 90% protein binding, from 48 to 72 h.
One- and two-compartment population pharmacokinetic models were fitted using the nonparametric adaptive grid (NPAG) algorithm in Pmetrics for R (LAPK, Los Angeles, CA). A base population pharmacokinetic model with two compartments (Akaike information criterion [AIC] of 818.0) fit the data better than a one-compartment model (AIC of 967.1). The relationships between individual predicted parameters from the base model and multiple covariates were explored with linear regression, and significant correlations were found between the volume of distribution of the central compartment (Vc) and TBW (r2 = 0.45, P = 0.002), lean body weight (LBW) (r2 = 0.36, P = 0.007), and body surface area (BSA) (r2 = 0.44, P < 0.002). Total body clearance (CL) was significantly correlated with TBW (r2 = 0.30, P = 0.020), LBW (r2 = 0.47, P = 0.002), BSA (r2 = 0.42, P = 0.004), and creatinine clearance (CRCL) (Cockcroft-Gault equation; r2 = 0.24, P = 0.039). Preliminary covariate models included covariate relationships for only one parameter at a time (Vc or CL) using linear or allometric scaling; the best individual fits for each parameter based on AIC were combined in the final model. The final covariate model (AIC of 811.9) scaled Vc by TBW normalized by the median observed value (Vc = Vθ × TBW/52.1) and CL as a linear function of CRCL (CL = CLNR + CLθ × CRCL), with CLNR representing nonrenal clearance and CLθ × CRCL representing renal clearance. Population parameter estimates are presented in Table 2. Population (r2 = 0.925) and individual (r2 = 0.985) Bayesian posterior predicted versus observed plots showed an excellent fit with negligible bias (Fig. 2); weighted residual errors are presented in Fig. S1 in the supplemental material. The final estimate for gamma was 3.99, indicating low environmental noise.
Population Bayesian posterior parameter estimates for the final telavancin population PK model of 18 adults with CFa
Predicted versus observed plots for the final covariate model. (Left) Population Bayesian posterior predicted versus observed telavancin concentrations. (Right) Individual Bayesian posterior predicted versus observed telavancin concentrations.
Probability of target attainment.The results of a 5,000-patient Monte Carlo simulation using a telavancin pharmacodynamic target fAUC/MIC of 215 demonstrated that a 10-mg/kg dose of telavancin had 100% PTA at the MIC breakpoint of 0.12 mg/liter for S. aureus (16), whereas 7.5 mg/kg had suboptimal PTA of 76.5% at this MIC. The probability of attaining a toxicodynamic target total AUC of 763, which has been associated with an increased risk of acute kidney injury (AKI) from telavancin (17), was 0.96% for the 10-mg/kg dose and 0% for the 7.5-mg/kg dose. A higher dose of 12.5 mg/kg also achieved 100% PTA at the MIC breakpoint, but PTA was 38.7% at the next highest doubling dilution of 0.25 mg/liter; furthermore, toxicodynamic target attainment probability increased to 14.3%. A dosing strategy using a fixed 750-mg dose (that is, an entire commercial vial) was also unable to produce optimal PTA above the breakpoint and yielded the highest attainment of the toxicodynamic target at 21%. Full results of the Monte Carlo simulations are presented in Fig. 3.
Probability of target attainment assessed at a target fAUC/MIC of 215 for a range of telavancin doses and MICs.
Safety and tolerability.Telavancin was well tolerated in the adult patients with CF. All adverse events were mild, and no significant increases in serum creatinine or significant decreases in CRCL were observed. Adverse events were observed in 14 participants and included metallic taste in mouth (n = 8), foamy urine (n = 5), nausea (n = 5), anemia (n = 4), emesis (n = 3), headache (n = 2), low serum creatinine (n = 2), electrolyte abnormalities (n = 1), rash (n = 1), elevated aspartate aminotransferase (AST) (n = 1), hypotension (n = 1), proteinuria (n = 1), dry mouth (n = 1), hypoalbuminemia (n = 1), and low neutrophil count (n = 1). No participants discontinued telavancin due to adverse events.
DISCUSSION
In this study, the plasma pharmacokinetics of telavancin were described in adult patients with CF experiencing an acute pulmonary exacerbation and the ability of various dosing strategies to attain a pharmacodynamic target predictive of clinical success was investigated. Prior investigations of telavancin plasma pharmacokinetics have been conducted in healthy subjects (12, 18), patients with complicated skin and skin structure infections and hospital-acquired pneumonia (HAP) (19, 20), patients with renal dysfunction (up to and including end-stage renal disease) (20, 21), the elderly (22), and patients with obesity class I to III (23). Adults with CF, however, may experience alterations in pharmacokinetics that are distinct from these groups which necessitate specific study of this population. Clinical experience using telavancin in this population is limited; to our knowledge, only a single case series, including 3 patients with CF, exists in the literature (24). The present report thus adds greatly to the literature informing the optimal dosing regimen and tolerability of telavancin for acute pulmonary exacerbations of CF. An important aspect of the present study is that only patients experiencing an acute pulmonary exacerbation of CF were eligible to enroll. In patients with HAP, increased severity was associated with decreasing CL and increasing Vc (19). Given the effect that infection can have on pharmacokinetics of telavancin, this pharmacokinetic model in patients with a CF pulmonary exacerbation is more clinically applicable.
Similarly to previous studies of telavancin plasma pharmacokinetics, a two-compartment model best described telavancin plasma concentration profiles in these CF adults. However, telavancin exposure was lower in CF adults than in non-CF adults in several other studies. At a dose of 10 mg/kg of body weight, and factoring in 90% plasma protein binding, these 12 CF participants had an observed mean steady-state fAUC of 52 mg · h · liter−1 compared with 76 mg · h · liter−1 in healthy volunteers (18), 78 mg · h · liter−1 in adults in phase I to III clinical trials with normal renal function (20), and 53 mg · h · liter−1 in adults with normal renal function after a single dose, that is, with no accumulation (21). Additionally, after multiple doses of 7.5 mg/kg, telavancin fAUC in healthy volunteers (60 mg · h · liter−1) was twice that of the six CF adults (30 mg · h · liter−1) who received this dosing regimen, and yet the mean TBW of the healthy volunteers (70 kg) was only 1.3 times greater than that of the CF adults receiving this dose (54 kg) (12). The fact that patients with CF had lower exposures agrees with prior observations that CF patients have greater clearance of renally eliminated drugs than adults without CF (25).
Among the factors contributing to altered pharmacokinetics in patients with CF are altered body composition and overall smaller stature due to reduced absorption of nutrients. Therefore, the effects of several descriptors of body composition on Vc were compared. Scaling Vc by TBW normalized by the observed median value produced the best fit in the final covariate model, and comparable fits were obtained using BSA and LBW (data not shown), both of which incorporate TBW. TBW is the most frequent covariate affecting telavancin Vc in non-CF populations (17, 18), and although this CF population had a TBW lower than the general population, its Vc (5.10) was similar to the Vc of healthy volunteers (5.01 liters) and of patients with HAP (5.00 liters) (19). In contrast, obese volunteers had a higher Vc (6.4 liters), and adjusted body weight produced a better covariate model fit than TBW (23). These results are unsurprising because, due to modern improvements in CF treatment, these patients tend to more closely resemble the general population than admitted CF patients did a few decades ago (26).
Clearance of renally eliminated drugs such as telavancin in adults with CF is typically enhanced, which was observed in the present study. Participants with CF had higher mean CL (1.25 liters · h−1) than healthy adult volunteers (1.06 liters · h−1) and adults with HAP (0.928 liters · h−1) (19), whereas obese volunteers had higher CL (1.58 liters · h−1) than these CF adults, which could have been due to higher Vc in the obese patients. In this population of relatively young CF adults without chronic kidney disease, CL was correlated with CRCL as well as with measures of body size and/or composition. The best-fit covariate model described CL as a function of CRCL; however, the improvement in fit over measures of body size and composition alone was modest (data not shown). This is likely due to the fact that the lowest CRCL observed in these participants was 73 ml · min−1, which is above the threshold for a renal dose adjustment in the package insert, and CRCL, itself, contains a measure of body size (27). Similarly, in other pharmacokinetic studies that included patients with impaired renal function, CRCL was utilized as a covariate to explain interindividual variability in telavancin CL (19, 20, 22).
The predicted PTA of televancin 10 mg/kg in adults with CF was 100% for MICs at or below the susceptibility breakpoint of 0.12 mg/liter (16), indicating the superiority of this dose to 7.5 mg/kg, which achieved only 76.5% PTA at this MIC. It should be noted that there currently are no pharmacodynamic threshold targets established specifically for acute pulmonary exacerbations of CF; therefore, the threshold resulting in at least 1 log reduction in CFU from the immunocompromised murine thigh infection model (fAUC/MIC ≥ 215) was substituted as a conservative target. A dose of 10 mg/kg was also sufficient to meet the same translational murine-derived pharmacodynamic target in obese adults (23) and in those with normal renal function (20); yet, in these studies, the PTA was also optimal for MICs up to 0.5 mg/liter and 1 mg/liter, respectively. Although neither 10 mg/kg nor higher doses of 12.5 mg/kg or 750 mg achieved optimal PTA for those MICs above the susceptibility breakpoint in adults with CF, this is unlikely to be clinically relevant. In a collection of 333 S. aureus strains cultured from patients with CF, 100% had a telavancin MIC of less than or equal to 0.06 mg/liter (28), which a 10-mg/kg dose is predicted to adequately treat even if any of these MICs were underestimated by 1 doubling dilution. Attempts to overcome telavancin resistance (MIC > 0.12 mg/liter) by using doses above 12.5 mg/kg would necessarily increase the chances of reaching the toxicity threshold AUC of 763 mg · h · liter−1 associated with AKI beyond 12.2% (17), creating an unacceptable risk of a serious adverse event while providing little to no benefit. The risk is further amplified by using fixed doses without regard to weight. If the participant with the lowest weight in this study (38.9 kg) received 750 mg (one vial), it would represent approximately 20 mg/kg.
Telavancin was well tolerated by all participants in this study, with the most common adverse events—metallic taste, foamy urine, and nausea—matching those described in the package insert (27). Safety was further validated by the low proportion of patients reaching a toxicodynamic AKI threshold; only 3.2% of simulated CF patients achieved this exposure at the 10-mg/kg dose, which was concordant with our observation of no significant increases in serum creatinine or CRCL in study participants.
Limitations of the present study included the small population and lack of determining drug exposure at the site of infection. Although pharmacokinetics in the epithelial lining fluid (ELF) better describes bacterial drug exposure during lung infections than in plasma, telavancin ELF concentrations in non-CF healthy volunteers were similar to free plasma concentrations, with a mean penetration ratio estimated at 1.01 ± 0.96 (14), and may thus serve as a proxy for drug exposure in the lung. Another limitation was that estimates of telavancin protein binding were obtained from the package insert rather than measured specifically in this CF population; however, all participants had normal serum albumin, and alterations in protein binding compared with the general population were therefore not expected. Finally, this was not a treatment study, and thus the efficacy of telavancin cannot be extrapolated from these results, which should instead be interpreted as supportive of clinical investigation of a telavancin dose of 10 mg/kg in this population.
In summary, a two-compartment population pharmacokinetic model of telavancin was developed for 18 adult patients with CF that included CRCL and TBW as significant covariates. This model was used to predict optimal PTA from a dose of 10 mg/kg/day against S. aureus at or below the MIC breakpoint with a low probability of nephrotoxicity. Despite decreased overall telavancin exposure in CF adults compared with the general population, currently approved dosing is predicted to be safe and effective against susceptible bacteria during a pulmonary exacerbation in this population.
MATERIALS AND METHODS
Study design and participants.This prospective, multicenter, open-label, clinical pharmacokinetic trial enrolled adult patients with CF admitted to the hospital for treatment of an acute pulmonary exacerbation (ClinicalTrials registration no. NCT03172793). All study procedures were approved by the institutional review board (IRB) at Hartford Hospital, the primary study center, and at all 3 additional study centers: University of Indiana/Riley Hospital for Children (Indianapolis, IN), UPMC Presbyterian Shadyside Hospital (Pittsburgh, PA), and St. Christopher’s Hospital for Children (Philadelphia, PA). All participants provided informed consent. Inclusion criteria were (i) age of ≥18 years; (ii) a documented diagnosis of CF; (iii) acute pulmonary exacerbation as the primary reason for hospital admission with the requirement to receive systemic antibiotic treatment; and (iv) if female, nonpregnant, nonbreastfeeding, and either not of childbearing potential or using an acceptable method of birth control, including abstinence from sexual intercourse, oral/parenteral contraceptives, or a barrier method. Exclusion criteria were (i) a history of any moderate or severe hypersensitivity or allergic reaction to telavancin or any component of telavancin or any glycopeptide antibiotic, except red man syndrome with vancomycin; (ii) solid-organ transplantation within the last 12 months; (iii) moderate to severe renal dysfunction, defined as CRCL <50 ml/min (Cockcroft-Gault calculation) or requirement for renal replacement therapy; (iv) oliguria or significant alterations in fluid/electrolyte status with a history of renal compromise within 72 h of enrollment; (v) hemoglobin <8 g/dl; (vi) anticipated length of stay <4 days; (vii) receiving or anticipation of receiving intravenous vancomycin; (viii) receiving an anticoagulant and requiring coagulation testing; (ix) concomitant administration of a medication with a cyclodextrin stabilizer; (x) rapidly progressing or immediately life-threatening illness; (xi) any condition or circumstance that, in the opinion of the investigator, would compromise the safety of the patient or the quality of the data; and (xii) planned or prior participation in any other interventional drug study within 30 days.
Antibiotic administration and blood sampling.Participants received 3 doses of telavancin based on TBW as a 1-h i.v. infusion every 24 h. A dose escalation study design was used where the first cohort of 6 participants received doses of 7.5 mg/kg of body weight and the second cohort of 6 participants received doses of 10 mg/kg. The dose for the third cohort of 6 participants, 10 mg/kg, was chosen after preliminary analysis of the first 12 patients. Telavancin 750 mg vials were provided by the study sponsor (Cumberland Pharmaceuticals, Nashville, TN) and stored refrigerated per manufacturer instructions. Each dose was prepared from a single vial reconstituted with 45 ml of 0.9% sodium chloride (normal saline [NS]) according to the manufacturer’s instructions and further diluted in NS to 100 ml total volume. All doses were required to be infused through a peripheral intravenous catheter, a peripherally inserted central catheter, or a port-a-cath, and no other drugs were simultaneously coadministered through the same access site.
Eleven blood samples were collected in sodium heparin-containing tubes: 1 sample immediately after the completion of each of the three infusions (peaks; drawn via venipuncture in arm contralateral to the infusion site), 1 sample immediately before the beginning of the second and third infusions (troughs), and 1 sample at the following time ranges after the third and final infusion: 15 to 30 min, 1 to 2 h, 3 to 4 h, 6 to 7 h, 8 to 12 h, and 24 h. Blood samples were immediately centrifuged at 4,000 × g for 15 min, and then plasma was separated and frozen at –80°C until analysis.
Analytical procedures.Telavancin plasma concentrations were assayed using a validated liquid chromatography/tandem mass spectrometry (LC-MS/MS) method (Q2 Solutions, Ithaca, NY). Plasma samples were extracted with ion-exchange solid-phase extraction. The extracted samples were separated in a high-performance LC (HPLC) system with a Cadenza HTC18 column using two mobile-phase gradients controlled by 5 Shimadzu LC-10AD pumps, 2 SCL-10 controllers, and a CTC PAL autosampler at 4°C. The retention time of telavancin and metabolite THRX-146241 was 8.1 ± 1.6 min and that of metabolite THRX-651540 was 6.5 ± 1.2 min; total run time per sample was 14 min. A Sciex API 5000 tandem mass spectrometer was used with electrospray ionization and a 13-minute acquisition time and 15.5-min cycle time. Selective reaction monitoring was performed for transitions at m/z 586.2 to 112.0 (telavancin), 591.5 to 112.0 (THRX-651540), and 592.9 to 112.0 (THRX-146241). The calibration curve was linear from the lower limit of quantification of 100 ng/ml to the upper limit of quantification of 25,000 ng/ml (r2 = 0.999). Mean interday coefficients of variation for high-quality (18,800 ng/ml) and low-quality (300 ng/ml) control samples were 2.4% and 5.1%, respectively. Mean intraday coefficients of variation for high-quality and low-quality control samples were 2.1% and 1.7%, respectively.
Pharmacokinetic analyses.Two phases of pharmacokinetic analysis were performed. The first phase modeled concentration data from the first two cohorts (12 participants) to select a dose for the final cohort of 6 participants. Doses of 7.5, 10, 12.5, and 15 mg/kg were considered. The lowest dose achieving optimal PTA (>90%) at the telavancin MIC breakpoint of 0.12 mg/liter without serious adverse events, 10 mg/kg, was selected for the final 6 participants. The second phase of analysis was performed using concentration data from all 3 cohorts (18 participants). Noncompartmental analysis (Phoenix Winnonlin; Certara, Princeton, NJ) was performed first to estimate pharmacokinetic parameter ranges and to calculate observed AUC (48 to 72 h). Thereafter, the NPAG algorithm in Pmetrics for R (Laboratory of Applied Pharmacokinetics, Los Angeles, CA) was used to fit one- and two-compartment structural models with zero-order intravenous infusion, first-order elimination, and first-order intercompartmental clearance (two-compartment models only). For each fit, a multiplicative error model was employed as follows: error = standard deviation (SD) × γ, where SD represents the equation of the regression line of best fit for the standard deviation versus observed concentrations from the telavancin assay (SD = 0.0179 + 0.0227 × observed) and γ represents environmental noise (with its value determined by the NPAG algorithm). A model was deemed superior based on visual fit of observed versus predicted concentration plots and minimization of AIC. Initially, a base pharmacokinetic model which did not incorporate covariate effects was generated. Linear regressions were used to measure the correlation between base model parameters Vc and CL versus covariates describing body size/composition (TBW [kg], BSA [m2], BMI [kg · m−2], LBW [kg]) and renal function (CRCL [ml · min−1]). Significantly correlated covariates were then explored for Vc and CL in a stepwise fashion. The effect of covariates on parameters influencing drug concentration in the peripheral compartment (Vp and Q) was not tested.
Monte Carlo simulation.Four 5,000-patient Monte Carlo simulations were performed using the SIMrun function in Pmetrics. Pharmacokinetic parameter and covariate values for the simulated population were sampled from normal distributions around the model support points for the observed population (29). Covariate values were restricted to fall within the range of values in the observed population. As a result, the simulated population had mean PK parameter and covariate values similar to those seen with the observed population. All simulations of different doses were performed using the same simulated population with identical PK parameter and covariate values. Three weight-based doses of telavancin (7.5 mg/kg, 10 mg/kg, and 12.5 mg/kg) or a fixed 750 mg at 0, 24, and 48 h were simulated, and concentrations were calculated from 48 h to 72 h in 15-min increments. AUC from 48 to 72 h was calculated using the trapezoidal rule and multiplied by 0.1 to obtain fAUC, accounting for 90% protein binding in humans (27). The PTA was calculated as the percentage of simulated patients reaching a fAUC/MIC of 215, which has been associated with 1 log10 CFU killing of S. aureus in a murine thigh infection model (30). Additionally, the percentage of patients reaching a total AUC/MIC of 763, previously associated with a higher risk of acute kidney injury (23), was determined.
Safety and tolerability.Participants were monitored daily for any adverse events, including any clinically significant changes in laboratory values. Laboratory data, including data from basic metabolic panel, complete blood count with differential, and liver function tests, were obtained at baseline and at end of study. Serum creatinine and blood urea nitrogen data were obtained daily during study procedures. An increase in serum creatinine of 50% and over 1.5 mg/dl was considered significant. After each cohort of 6 patients completed participation, case report forms were reviewed by an unblind independent data monitoring committee comprised of two physicians not associated with the study who provided recommendations for moving to the next dose outlined in the protocol.
ACKNOWLEDGMENTS
We acknowledge Elizabeth Martin, Debora Santini, Dee Rendock, J. Samuel Pope, John McArdle, Laurie Varlotta, Elizabeth Hartigan, Joel Weinberg, Joseph Pilewski, and all CF participants and their families for assistance with the conduct of study.
This study was funded by a grant from Cumberland Pharmaceuticals (Nashville, TN) and by Cystic Fibrosis Foundation Center grant C101-12 FY19.
We have no conflicts of interest to disclose.
FOOTNOTES
- Received 19 September 2019.
- Returned for modification 21 October 2019.
- Accepted 25 October 2019.
- Accepted manuscript posted online 4 November 2019.
Supplemental material is available online only.
- Copyright © 2019 American Society for Microbiology.
