Sampling from covariate distribution may not always be necessary in PK/PD simulations: illustrative examples with antibiotics

Pharmacokinetics (PK)/pharmacodynamics (PD) modeling and simulation is crucial for optimizing antimicrobial dosing. This study assessed covariate impact on PK variability and identified scenarios where fixing the covariate with median value proves effective PK/PD simulations for antibiotics with population PK (popPK) model including only one covariate effect. Three published popPK models were employed, with creatinine clearance (CRCL) identified as a covariate on clearance (CL) for meropenem and tobramycin, and total body weight (WT) associated with the volume of distributions for daptomycin. Given a fixed dose for Meropenem (1000 mg), and a WT based dose for tobramycin (7 mg/kg) and daptomycin (6 mg/kg), PK/PD simulation outcomes (e.g., percentage of PK/PD target attainment (PTA) and toxicity risk) were compared between fixed covariate-based and covariate distribution-based approaches. Covariate impact on PK was assessed through deterministic simulation using outer bounds of covariate versus simulation using median covariate value, with an overlap ratio calculated the percentage of overlapped area under concentration-time curve (AUC) between these two simulation approaches. Meropenem and tobramycin simulations showed a broader PK profiles and distinct PTA distribution with sampled covariate distribution, while daptomycin simulations exhibited consistency in PK/PD characteristics. CRCL had a relative strong impact on meropenem and tobramycin PK, while a weak impact of WT on daptomycin PK was observed from extensive overlap in simulated PK curves, with an overlap ratio of 99.5%. Regarding a weak covariate impact on PK with high overlap ratio, sampling from covariate distribution may not significantly enhance simulation performance, fixing covariate with median value could be efficient for antibiotic PK/PD simulations.

The impact of extracorporeal support on antimicrobial pharmacokinetics in critically ill neonatal and paediatric patients: A systematic review

OBJECTIVES

Infections represent a major risk for critically ill neonatal and paediatric patients requiring extracorporeal life-saving support such as extracorporeal membrane oxygenation (ECMO) and/or continuous renal replacement therapies (CRRT). Patient outcomes rely on achieving target antimicrobial concentrations. In critically ill adults on extracorporeal support, suboptimal antimicrobial concentrations have been shown to be common. Our objective was to systematically review antimicrobial pharmacokinetic studies in critically ill term neonatal and paediatric patients receiving ECMO and/or CRRT and compare them to similar cohorts of patients not receiving ECMO or CRRT.

METHODS

Studies published between 1990 and 2022 were identified through systematic searches in PUBMED, Embase, Web of Science, Medline, Google Scholar and CINAHL. Studies were included which provided antimicrobial pharmacokinetic parameters (volume of distribution and clearance) in the neonatal and paediatric patients receiving ECMO and/or CRRT. Studies were excluded if no antimicrobial pharmacokinetic parameters were described or could be calculated.

RESULTS

Forty-four pharmacokinetic studies were identified describing 737 patients, with neonatal patients recruited in 70% of the ECMO studies and <1% of the CRRT studies. Of all the studies, 50% were case reports or case series. The pharmacokinetics were altered for gentamicin, daptomycin, ceftolozane, micafungin, voriconazole, cefepime, fluconazole, piperacillin, and vancomycin, although considerable patient variability was described.

CONCLUSION

Significant gaps remain in our understanding of the pharmacokinetic alterations in neonatal and paediatric patients receiving ECMO and CRRT support.

Population pharmacokinetics of intravenous daptomycin in critically ill patients: implications for selection of dosage regimens

Daptomycin is gaining prominence for the treatment of methicillin-resistant Staphylococcus aureus infections. However, the dosage selection for daptomycin in critically ill patients remains uncertain, especially in Chinese patients. This study aimed to establish the population pharmacokinetics of daptomycin in critically ill patients, optimize clinical administration plans, and recommend appropriate dosage for critically ill patients in China. The study included 64 critically ill patients. Blood samples were collected at the designated times. The blood daptomycin concentration was determined using validated liquid chromatography-tandem mass spectrometry. A nonlinear mixed-effects model was applied for the population pharmacokinetic analysis and Monte Carlo simulations of daptomycin. The results showed a two-compartment population pharmacokinetic model of daptomycin in critically ill adult Han Chinese patients. Monte Carlo simulations revealed that a daily dose of 400 mg of daptomycin was insufficient for the majority of critically ill adult patients to achieve the anti-infective target. For critically ill adult patients with normal renal function (creatinine clearance rate >90 mL/min), the probability of achieving the target only reached 90% when the daily dose was increased to 700 mg. For patients undergoing continuous renal replacement therapy (CRRT), 24 h administration of 500 mg met the pharmacodynamic goals and did not exceed the safety threshold in most patients. Therefore, considering its efficacy and safety, intravenous daptomycin doses are best scaled according to creatinine clearance, and an increased dose is recommended for critically ill patients with hyperrenalism. For patients receiving CRRT, medication is recommended at 24 h intervals.

Removal of common antimicrobial agents by sustained low-efficiency dialysis

Adequate dosing of antimicrobials is paramount for treating infections in critically ill patients undergoing kidney replacement therapy; however, little is known about antimicrobial removal by sustained low-efficiency dialysis (SLED). The objective was to quantify the removal of cefepime, daptomycin, meropenem, piperacillin-tazobactam, and vancomycin in patients undergoing SLED. Adult patients ≥18 years with acute kidney injury (AKI) or end-stage kidney disease receiving one of the select antimicrobials and requiring SLED were included. Blood and dialysate flow rates were maintained at 250 and 100 mL/min, respectively. Simultaneous arterial and venous blood samples for the analysis of antibiotic concentrations were collected hourly for 8 hours during SLED (on-SLED). Arterial samples were collected every 2 hours for up to 6 hours while not receiving SLED (off-SLED) for the calculation of SLED clearance, half-life (t1/2) on-SLED and off-SLED, and the fraction of removal by SLED (fD). Twenty-one patients completed the study: 52% male, mean age (±SD) 53 ± 13 years, and mean weight of 98 ± 30 kg. Eighty-six percent had AKI, and 4 patients were receiving cefepime, 3 daptomycin, 10 meropenem, 6 piperacillin-tazobactam, and 13 vancomycin. The average SLED time was 7.3 ± 1.1 hours, and the mean ultrafiltration rate was 95 ± 52 mL/hour (range 10-211). The t1/2 on-SLED was substantially lower than the off-SLED t1/2 for all antimicrobials, and the SLED fD varied between 44% and 77%. An 8-hour SLED session led to significant elimination of most antimicrobials evaluated. If SLED is performed, modification of the dosing regimen is warranted to avoid subtherapeutic concentrations.

Bayesian prediction-based individualized dosing of anti-methicillin-resistant Staphylococcus aureus treatment: Recent advancements and prospects in therapeutic drug monitoring

As one of the efficient techniques for TDM, the population pharmacokinetic (popPK) model approach for dose individualization has been developed due to the rapidly growing innovative progress in computer technology and has recently been considered as a part of model-informed precision dosing (MIPD). Initial dose individualization and measurement followed by maximum a posteriori (MAP)-Bayesian prediction using a popPK model are the most classical and widely used approach among a class of MIPD strategies. MAP-Bayesian prediction offers the possibility of dose optimization based on measurement even before reaching a pharmacokinetically steady state, such as in an emergency, especially for infectious diseases requiring urgent antimicrobial treatment. As the pharmacokinetic processes in critically ill patients are affected and highly variable due to pathophysiological disturbances, the advantages offered by the popPK model approach make it highly recommended and required for effective and appropriate antimicrobial treatment. In this review, we focus on novel insights and beneficial aspects of the popPK model approach, especially in the treatment of infectious diseases with anti-methicillin-resistant Staphylococcus aureus agents represented by vancomycin, and discuss the recent advancements and prospects in TDM practice.

Population Pharmacokinetic Model for Unbound Concentrations of Daptomycin in Patients with MRSA Including Patients Undergoing Hemodialysis

BACKGROUND AND OBJECTIVE

Unbound daptomycin concentrations are responsible for pharmacologically beneficial and adverse effects, although most previous reports have been limited to the use of total concentrations. We developed a population pharmacokinetic model to predict both total and unbound daptomycin concentrations.

METHODS

Clinical data were collected from 58 patients with methicillin-resistant Staphylococcus aureus including patients undergoing hemodialysis. A total of 339 serum total and 329 unbound daptomycin concentrations were used for model construction.

RESULTS

Total and unbound daptomycin concentration was explained by a model that assumed first-order distribution with two compartments, and first-order elimination. Normal fat body mass was identified as covariates. Renal function was incorporated as a linear function of renal clearance and independent non-renal clearance. The unbound fraction was estimated to be 0.066 with a standard albumin of 45 g/L and standard creatinine clearance of 100 mL/min. Simulated unbound daptomycin concentration was compared with minimum inhibitory concentration as a measure of clinical effectiveness and exposure-level-related induction of creatine phosphokinase elevation. The recommended doses were 4 mg/kg for patients with severe renal function [creatinine clearance (CLcr) ≤ 30 mL/min] and 6 mg/kg for patients with mild to moderate renal function (CLcr > 30 and ≤ 60 mL/min). A simulation indicated that dose adjusted by body weight and renal function improved target attainment.

CONCLUSIONS

This population pharmacokinetics model for unbound daptomycin could help clinicians to select the appropriate dose regimen for patients undergoing daptomycin treatment and reduce associated adverse effects.

Population Pharmacokinetic/Pharmacodynamic Modelling of Daptomycin for Schedule Optimization in Patients with Renal Impairment

The aims of this study are (i) to develop a population pharmacokinetic/pharmacodynamic model of daptomycin in patients with normal and impaired renal function, and (ii) to establish the optimal dose recommendation of daptomycin in clinical practice. Several structural PK models including linear and non-linear binding kinetics were evaluated. Monte Carlo simulations were conducted with a fixed combination of creatinine clearance (30-90 mL/min/1.73 m2) and body weight (50-100 kg). The final dataset included 46 patients and 157 daptomycin observations. A two-compartment model with first-order peripheral distribution and elimination kinetics assuming non-linear protein-binding kinetics was selected. The bactericidal effect for Gram+ strains with MIC ≤ 0.5 mg/L could be achieved with 5-12 mg/kg daily daptomycin based on body weight and renal function. The administration of 10-17 mg/kg q48 h daptomycin allows to achieve bactericidal effect for Gram+ strains with MIC ≤ 1 mg/L. Four PK samples were selected as the optimal sampling strategy for an accurate AUC estimation. A quantitative framework has served to characterize the non-linear binding kinetics of daptomycin in patients with normal and impaired renal function. The impact of different dosing regimens on the efficacy and safety outcomes of daptomycin treatment based on the unbound exposure of daptomycin and individual patient characteristics has been evaluated.

Physiologically based pharmacokinetic modeling of daptomycin dose optimization in pediatric patients with renal impairment

Background and Objective: Daptomycin is used to treat Gram-positive infections in adults and children and its dosing varies among different age groups. We focused on the pharmacokinetics of daptomycin in children with renal impairment, which has not been evaluated.

Methods: A physiologically based pharmacokinetic (PBPK) model of daptomycin was established and validated to simulate its disposition in healthy populations and adults with renal impairment, along with a daptomycin exposure simulated in pediatric patients with renal impairment.

Results: The simulated PBPK modeling results for various regimens of intravenously administered daptomycin were consistent with observed data according to the fold error below the threshold of 2. The Cmax and AUC of daptomycin did not differ significantly between children with mild-to-moderate renal impairment and healthy children. The AUC increased by an average of 1.55-fold and 1.85-fold in severe renal impairment and end-stage renal disease, respectively. The changes were more significant in younger children and could reach a more than 2-fold change. This scenario necessitates further daptomycin dose adjustments.

Conclusion: Dose adjustments take into account the efficacy and safety of the drug; however, the steady-state Cmin of daptomycin may be above 24.3 mg/L in a few instances. We recommend monitoring creatine phosphokinase more than once a week when using daptomycin in children with renal impairment.

Daptomycin Population Pharmacokinetics in Patients Affected by Severe Gram-Positive Infections: An Update

Daptomycin pharmacokinetics may not depend on renal function only and it significantly differs between healthy volunteers and severely ill patients. Herein, we propose a population pharmacokinetics model based on 424 plasma daptomycin concentrations collected from 156 patients affected by severe Gram-positive infections during a routine therapeutic drug monitoring protocol. Model building and validation were performed using NONMEM 7.2 (ICON plc), Xpose4 and Perl-speaks-to-NONMEM. The final pop-PK model was a one-compartment first-order elimination model, with a 2.7% IIV for drug clearance (Cl), influence of creatinine clearance on drug clearance and of sex on distribution volume. After model validation, we simulated 10,000 patients with the Monte-Carlo method to predict the efficacy and tolerability of different daptomycin daily dosages. For the most common 6 mg/kg daily dose, the simulated probability of overcoming the toxic minimum concentration (24.3 mg/L) was 14.8% and the efficacy (expressed as a cumulative fraction of response) against methicillin-resistant S. aureus, S. pneumoniae and E. faecium was 95.77%, 99.99% and 68%, respectively. According to the model-informed precision dosing paradigm, pharmacokinetic models such as ours could help clinicians to perform patient-tailored antimicrobial dosing and maximize the odds of therapy success without neglecting toxicity risks.

Optimizing Antimicrobial Dosing for Critically Ill Patients with MRSA Infections: A New Paradigm for Improving Efficacy during Continuous Renal Replacement Therapy

The dosage regimen of vancomycin, teicoplanin and daptomycin remains controversial for critically ill patients undergoing continuous renal replacement therapy (CRRT). Monte Carlo simulation was applied to identify the optimal regimens of antimicrobial agents in patients with methicillin-resistant Staphylococcus aureus (MRSA) infections based on the mechanisms of different CRRT modalities on drug clearance. The optimal vancomycin dosage for patients received a CRRT doses ≤ 30 mL/kg/h was 20 mg/kg loading dose followed by 500 mg every 8 h, while 1 g every 12 h was appropriate when 35 mL/kg/h was prescribed. The optimal teicoplanin dosage under a CRRT dose ≤ 25 mL/kg/h was four loading doses of 10 mg/kg every 12 h followed by 10 mg/kg every 48 h, 8 mg/kg every 24 h and 6 mg/kg every 24 h for continuous veno-venous hemofiltration, continuous veno-venous hemodialysis and continuous veno-venous hemodiafiltration, respectively. When the CRRT dose increased to 30–35 mL/kg/h, the teicoplanin dosage should be increased by 30%. The recommended regimen for daptomycin was 6–8 mg/kg every 24 h under a CRRT dose ≤ 25 mL/kg/h, while 8–10 mg/kg every 24 h was optimal under 30–35 mg/kg/h. The CRRT dose has an impact on probability of target attainment and CRRT modality only influences teicoplanin.

Population Pharmacokinetics and Dosing Optimization of Gentamicin in Critically Ill Patients Undergoing Continuous Renal Replacement Therapy

Purpose

Appropriate gentamicin dosing in continuous renal replacement therapy (CRRT) patients remains undefined. This study aimed to develop a population pharmacokinetic (PK) model of gentamicin in CRRT patients and to infer the optimal dosing regimen for gentamicin.

Methods

Fourteen CRRT patients dosed with gentamicin were included to establish a population PK model to characterize the variabilities and influential covariates of gentamicin. The pharmacokinetic/pharmacodynamic (PK/PD) target attainment and risk of toxicity for different combinations of gentamicin regimens (3–7 mg/kg q24h) and CRRT effluent doses (30–50 mL/h/kg) were evaluated by Monte Carlo simulation. The probability of target attainment (PTA) was determined for the PK/PD indices of the ratio of drug peak concentration/minimum inhibitory concentration (C_max/MIC > 10) and the ratio of area under the drug concentration–time curve/MIC over 24 h (AUC_0-24h/MIC > 100), and the risk of toxicity was estimated by drug trough concentration thresholds (1 and 2 mg/L).

Results

A one-compartment model adequately described the PK characteristics of gentamicin. Covariates including body weight, age, gender, and CRRT modality did not influence the PK parameters of gentamicin based on our dataset. All studied gentamicin regimens failed to achieve satisfactory PTAs for pathogens with an MIC ≥2 mg/L. A good balance of PK/PD target attainment and risk of toxicity (>2 mg/L) was achieved under 7 mg/kg gentamicin q24h and 40 mL/kg/h CRRT dose for an MIC ≤1 mg/L. CRRT dose intensity had a significant impact on the target attainment of AUC_0-24h/MIC >100 and risk of toxicity.

Conclusion

A combination of 7 mg/kg gentamicin q24h and 40 mL/kg/h CRRT dose might be considered as a starting treatment option for CRRT patients, and drug monitoring is required to manage toxicity.

A successful daptomycin and micafungin dosing strategy in veno-venous ECMO and continuous renal replacement

WHAT IS KNOWN AND OBJECTIVE

Studies have demonstrated that ECMO leads to pharmacokinetic changes, with alterations in volume of distribution, clearance and drug sequestration by the circuit. We describe a successful dosing approach for daptomycin and micafungin for the treatment of VRE faecium bacteremia and C. glabrata fungemia in a patient receiving veno-venous ECMO and CRRT.

CASE SUMMARY

We report a case of a patient with ARDS on veno-venous ECMO complicated by VRE faecium bacteremia and C. glabrata fungemia. The patient was treated with daptomycin 10 mg/kg every 24 h and micafungin 150 mg every 24 h for 14 days. Key observations included the documented bacteremia and fungemia clearance without the need for ECMO circuit exchange.

WHAT IS NEW AND CONCLUSION

This case report demonstrates successful bacteremia and fungemia clearance in an adult without the need for ECMO circuit exchange. It also highlights the need for more research to identify optimal antimicrobial dosing strategies in similar scenarios.

The impact of continuous renal replacement therapy on antibiotic pharmacokinetics in critically ill patients

Introduction: Mortality due to severe infections in critically ill patients undergoing continuous renal replacement therapy (CRRT) remains high. Nevertheless, rapid administration of adequate antibiotic therapy can improve survival. Delivering optimized antibiotic therapy can be a challenge, as standard drug regimens often result in insufficient or excessive serum concentrations due to significant changes in the volume of distribution and/or drug clearance in these patients. Insufficient drug concentrations can be responsible for therapeutic failure and death, while excessive concentrations can cause toxic adverse events.Areas covered: We performed a narrative review of the impact of CRRT on the pharmacokinetics of the most frequently used antibiotics in critically ill patients. We have provided explanations for the changes in the PKs of antibiotics observed and suggestions to optimize dosage regimens in these patients.Expert opinion: Despite considerable efforts to identify optimal antibiotic dosage regimens for critically ill patients receiving CRRT, adequate target achievement remains too low for hydrophilic antibiotics in many patients. Whenever possible, individualized therapy based on results from therapeutic drug monitoring must be given to avoid undertreatment or toxicity.

Population pharmacokinetics and individual analysis of daptomycin in kidney transplant recipients

BACKGROUND

Little is known about the population pharmacokinetics (PPK) of daptomycin in kidney transplant patients. The present study established a pharmacokinetic model for daptomycin in kidney transplant patients in China and examinee the important factors affecting the pharmacokinetic parameters of daptomycin.

METHODS

The study population included 49 kidney transplant patients with 537 daptomycin concentrations. The PPK model of daptomycin was developed using a nonlinear mixed-effects model, a two-compartment structural model, and a mixed residual error model. The stability and predictive ability of the final model were evaluated based on bootstrapping, visual prediction checks and normalized prediction distribution errors.

RESULTS

Glomerular filtration rate (GFR) and total body weight significantly affected clearance, and body weight influenced the central volume of distribution. The average clearance of the population was 0.316 L/h, the central volume of distribution was 6.04 L, the intercompartmental clearance was 2.31 L/h, and the peripheral volume of distribution was 2.46 L. Based on the established model and the target of area under curve (AUC_0-24h)/minimum inhibition concentration (MIC) ≥666, we developed a recommended dose regimen for kidney transplant patients according to their renal function and weight. The daily doses were 4.0±0.31, 4.7±0.36, 5.1±0.40, 5.5±0.43, 5.8±0.45, and 6.1±0.48 mg/kg when the GFRs were 15, 30, 45, 60, 75, and 90 ml/min/1.73 m², respectively.

CONCLUSION

This study provides a reference for individualized daptomycin administration in kidney transplant recipients, and it is a valuable resource for improving the treatment effect and reducing the toxic effects of daptomycin.

High-Dose Daptomycin and Clinical Applications

OBJECTIVE

To evaluate evidence for high-dose daptomycin (doses ≥ 8 mg/kg/d).

DATA SOURCES

A PubMed/MEDLINE literature search was performed (January 2000 to December 2020) using the search terms daptomycin, high dose, and dosing. Review article references and society guidelines were reviewed.

STUDY SELECTION AND DATA EXTRACTION

Clinical trials, observational studies, retrospective studies, meta-analyses, and systematic reviews reporting on high-dose daptomycin were included.

DATA SYNTHESIS

Experimentally, daptomycin outperforms other antimicrobials for high inoculum and biofilm-associated infections. Clinically, high-dose daptomycin is supported as salvage and first-line therapy for endocarditis and bacteremia, primarily when caused by methicillin-resistant Staphylococcus aureus (when vancomycin minimum inhibitory concentration is >1 mg/L) and Enterococcus. High-dose daptomycin appears effective for osteomyelitis and central nervous system infections, although comparative studies are lacking. High dosing in renal replacement therapy requires considering clearance modality to achieve exposures like normal renal function. Weight-based dosing in obesity draws concern for elevated exposures, although high doses have not been evaluated kinetically in obesity. Some data show benefits of high doses in overweight populations. Burn patients clear daptomycin more rapidly, and high doses may only achieve drug exposures similar to standard doses (6 mg/kg).

RELEVANCE TO PATIENT CARE AND CLINICAL PRACTICE

This review analyzes the efficacy and safety of high-dose daptomycin in serious gram-positive infections. Discussion of specific infectious etiologies and patient populations should encourage clinicians to evaluate their daptomycin dosing standards.

CONCLUSIONS

The efficacy of high-dose daptomycin and limited safety concerns encourage clinicians to consider high-dose daptomycin more liberally in severe gram-positive infections.

Clinical Pharmacokinetics of Daptomycin

Due to the low level of resistance observed with daptomycin, this antibiotic has an important place in the treatment of severe Gram-positive infections. It is the first-in-class of the group of calcium-dependent, membrane-binding lipopeptides, and is a cyclic peptide constituted of 13 amino acids and an n-decanoyl fatty acid chain. The antibacterial action of daptomycin requires its complexation with calcium. Daptomycin is not absorbed from the gastrointestinal tract and needs to be administered parenterally. The distribution of daptomycin is limited (volume of distribution of 0.1 L/kg in healthy volunteers) due to its negative charge at physiological pH and its high binding to plasma proteins (about 90%). Its elimination is mainly renal, with about 50% of the dose excreted unchanged in the urine, justifying dosage adjustment for patients with renal insufficiency. The pharmacokinetics of daptomycin are altered under certain pathophysiological conditions, resulting in high interindividual variability. As a result, therapeutic drug monitoring of daptomycin may be of interest for certain patients, such as intensive care unit patients, patients with renal or hepatic insufficiency, dialysis patients, obese patients, or children. A target for the ratio of the area under the curve to the minimum inhibitory concentration > 666 is usually recommended for clinical efficacy, whereas in order to limit the risk of undesirable muscular effects the residual concentration should not exceed 24.3 mg/L.

Recommendation of Antimicrobial Dosing Optimization During Continuous Renal Replacement Therapy

Continuous Renal Replacement Therapy (CRRT) is more and more widely used in patients for various indications recent years. It is still intricate for clinicians to decide a suitable empiric antimicrobial dosing for patients receiving CRRT. Inappropriate doses of antimicrobial agents may lead to treatment failure or drug resistance of pathogens. CRRT factors, patient individual conditions and drug pharmacokinetics/pharmacodynamics are the main elements effecting the antimicrobial dosing adjustment. With the development of CRRT techniques, some antimicrobial dosing recommendations in earlier studies were no longer appropriate for clinical use now. Here, we reviewed the literatures involving in new progresses of antimicrobial dosages, and complied the updated empirical dosing strategies based on CRRT modalities and effluent flow rates. The following antimicrobial agents were included for review: flucloxacillin, piperacillin/tazobactam, ceftriaxone, ceftazidime/avibactam, cefepime, ceftolozane/tazobactam, sulbactam, meropenem, imipenem, panipenem, biapenem, ertapenem, doripenem, amikacin, ciprofloxacin, levofloxacin, moxifloxacin, clindamycin, azithromycin, tigecycline, polymyxin B, colistin, vancomycin, teicoplanin, linezolid, daptomycin, sulfamethoxazole/trimethoprim, fluconazole, voriconazole, posaconzole, caspofungin, micafungin, amphotericin B, acyclovir, ganciclovir, oseltamivir, and peramivir.