Oseltamivir Dosing for Influenza Infection in Premature Neonates

Oseltamivir is approved for the treatment and prevention of influenza in patients aged ≥1 year, although few published data are available regarding the pharmacokinetics of oseltamivir in 1-year-old children [1]. In response to the 2009 novel H1N1 (2009 H1N1) pandemic, the United States Food and Drug Administration issued an Emergency Use Authorization for oseltamivir use in infants aged <1 year [2]. Initially under the Emergency Use Authorization, children aged 3 months were recommended to receive 12 mg twice daily for 5 days, regardless of weight, for the treatment of influenza infection [3, 4]. These guidelines subsequently have been changed to a weight-based dosing approach (3.0 mg/kg/dose orally twice daily for children aged 0 to <12 months) on the basis of CASG 114 interim study results [5]. The CASG 114 protocol is enrolling infants and children aged <2 years in an oseltamivir pediatric dose-finding trial. For inclusion into CASG 114, participants must fall into 1 of 5 chronological age strata (12–23, 9–11, 6–8, 3–5, and 0–2 months), have confirmed diagnosis of influenza, and have duration of symptoms ≤96 h. Steady-state pharmacokinetic data are collected in each cohort to determine the appropriate oseltamivir dosage with use of an area-under-the-curve (AUC) targeted design. However, no data exist on the dosing of oseltamivir in premature babies, and it was for this reason that the sampling study reported herein (CASG 119) was undertaken.

The prodrug oseltamivir phosphate is rapidly and efficiently converted to its active metabolite, oseltamivir carboxylate, which is then cleared unchanged by the kidney [6]. It has been well documented that infants and young children exhibit diminished renal capacity, and drugs or drug metabolites eliminated by the kidneys require dosage adjustments to account for this decrease in clearance [7]. Consequently, premature babies likely will require an even lower dose than term infants and young children, because their renal function is severely underdeveloped. Other factors, such as altered enzymatic conversion and changes in oral bioavailability, may also contribute to determination of neonatal dosing.

Developing a prospective, dose-finding trial of a drug to treat influenza in premature babies is exceedingly difficult. Because of blood volume restraints, few samples for pharmacokinetic analyses can be collected in any 1 baby, thereby limiting the trial design and robustness of the dataset. Importantly, most premature babies will be in a neonatal intensive care unit (NICU). The primary method of these babies contracting influenza is by exposure from an adult health care worker or NICU visitor, most of whom will not realize that they have influenza until after the exposure has occurred. For this reason, a prospective study design to evaluate oseltamivir dosing in premature neonates is impractical. Because it would be unethical to perform a dose-finding trial in otherwise healthy premature babies, the only options for collecting these important data are to open a trial whereby multiple NICUs have a standing protocol to collect blood samples for pharmacokinetic analysis in case an exposure occurs and treatment or prophylaxis ensues or to rapidly implement a study to collect these specimens when an exposure and subsequent therapy actually does occur. The latter situation forms the basis for this report.

Methods. A hospital health care worker treated an adult patient with documented 2009 H1N1 several days before going to work in the hospital's NICU. The health care worker became ill, including fever, during her NICU shift caring for 32 babies. She remained at work, however, and did not wear gloves or masks. Following her shift, she tested positive for influenza and was subsequently confirmed to have 2009 H1N1 infection. Following the NICU exposure, the treating neonatologist elected to administer oseltamivir prophylactically to these neonates at a scheduled dose of 1.5 mg/kg/dose twice daily for 10 days. This dose was an educated guess on the basis of data already collected from the CASG 114 study and knowledge of developmental pharmacology, with premature babies generally having diminished renal function relative to term babies. All doses were administered using a standard milliliter-marked oral syringe. The CASG 119 protocol was then drafted expeditiously to collect plasma samples for drug measurement and to determine whether the selected dose produced similar oseltamivir carboxylate exposures (AUC) relative to those observed in infants and young children in the ongoing CASG 114 protocol [5].

Demographic data, such as gestational age, chronological age, weight, serum creatinine level, and dosing information, were obtained on enrollment. A single whole blood sample (0.5 mL) from each patient was collected after the fifth dose, to measure steady-state plasma oseltamivir and oseltamivir carboxylate concentrations. This single sample was scheduled to be obtained from each baby during a specific time window to fully encompass the 12-h dosing interval. The sample collection time windows were 0–3 h, 4–6 h, 7–9 h, and 10–12 h. The goal of this design was to have samples evenly distributed across the collection windows. Mass spectrometry was used to quantitate oseltamivir and oseltamivir carboxylate from plasma samples; lower limits of detection for the assay were 1 and 10 ng/mL, respectively [8].

Several approaches were undertaken to optimally describe the CASG 119 concentration-time data. First, data were modeled using the ADAPT 5.0 systems analysis software [9]. The CASG 114 dataset was used to establish a combined parent-metabolite model. A 2-compartment model for oseltamivir phosphate and 1-compartment model for oseltamivir carboxylate was applied, and maximum likelihood estimation maximization was used to conduct a population pharmacokinetic analysis of a portion of the CASG 114 dataset (n = 43) and the CASG 119 data. CASG 119 data were modeled as if 1 subject had received the oseltamivir dose at steady-state, and the mean of all available concentrations at each time point constituted the concentration-time curve. Absorption, metabolite formation, and clearance processes were assumed to be linear. Second, a noncompartmental analysis (WinNonlin 5.2.1) of the mean concentrations at each time point collected in CASG 119 was also performed. Lastly, raw oseltamivir carboxylate concentrations were averaged for each dataset to estimate the mean steady-state concentration.

Results. Twenty of the 32 exposed babies were enrolled on this pharmacokinetic sampling study. Nine babies were discharged prior to sampling, 2 refused oseltamivir prophylaxis, and 1 declined study participation. Subject demographic characteristics are presented in Table 1. The median gestational age was 27.5 weeks; 1 subject was term at delivery (38 weeks). The median weight and chronological age at time of pharmacokinetic sampling were 1684 grams and 2.5 weeks, respectively. Seven subjects were extremely premature, with gestational ages ≤26 weeks. Eight subjects weighed <1000 grams at delivery, of whom 2 weighed <500 grams. The median dose received and number of doses administered prior to pharmacokinetic sampling were 1.79 mg/kg/dose (range, 1.33–2.55 mg/kg/dose) and 11 (range, 9–13), respectively. The range of actual doses received by the premature neonates was 1.3–6.9 mg twice daily. None of the babies developed influenza infection or experienced drug-related adverse effects over the course of therapy.

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

Figure 1

Top panel, Measured oseltamivir phosphate concentrations from all cohorts in CASG 114 (blue triangles) and from premature neonates in the current study (red circles). Bottom panel, Measured oseltamivir carboxylate concentrations from all cohorts in CASG 114 (blue triangles)and from premature neonates in the current study (red circles).

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

Table 1

Participant Demographic Characteristics

One baby was treated with oseltamivir on a once-daily basis, and these data were not included in the analysis. A second baby had an oseltamivir concentration 23 times that of the mean concentration of the remaining babies (335 ng/mL vs 14.3 ng/ mL for the remaining babies). This subject also had the highest serum creatinine level of 1.31 mg/dL. Because of the low robustness of the dataset (1 sample per subject), this subject's results were not used because they strongly influenced the best model fit. The oseltamivir result for another subject was below the limit of assay quantitation. Therefore, 17 oseltamivir and 18 oseltamivir carboxylate concentrations were available for analyses.

Figure 1 illustrates the raw concentration-time data from premature babies in the current study, compared with those from all cohorts in the ongoing CASG 114 study. The mean dose (± standard deviation) administered in the current study was 1.73 ± 0.17 mg/kg. In CASG 114, the current mean dose and AUC₁₂ (± standard deviation) across all cohorts is 3.0 ± 0.25 mg/kg and 4326 ± 1878 ng · h/mL, respectively. The modeled CASG 119 AUC₁₂ was 9250 ng · h/mL. A linearly adjusted dose of 0.81 mg/kg twice daily would be needed to achieve a similar AUC₁₂ in the premature babies. Similarly, a noncompartmental analysis of the mean oseltamivir carboxylate concentration-time data at each time point yielded an AUC₁₂ of 8079 ng · h/mL at a mean dose of 1.73 mg/kg. This analysis suggests a linearly adjusted dose of 0.93 mg/kg twice daily in premature babies would produce an AUC₁₂ similar to that in the CASG 114 cohort. The final analysis included simply averaging the oseltamivir carboxylate concentrations in both studies. For CASG 119, the mean of all raw oseltamivir carboxylate concentrations was 728 ng/mL, compared with 346 ng/mL in CASG 114. Again using a linear dose adjustment, 0.82 mg/kg/dose twice daily would be needed to produce a similar mean steady-state concentration in premature babies.

Discussion. Analysis of these data suggest that metabolic differences occur between premature neonates and term infants and young children with regard to oseltamivir and oseltamivir carboxylate disposition following oral administration. The average oseltamivir dose in the youngest age cohort in CASG 114 (term babies aged 0–2 months; n = 18) is 3.02 mg/kg/dose twice daily (12.4 mg/dose twice daily), whereas premature neonates from the current study ( n = 18) received a mean oseltamivir dose of 1.73 mg/kg/dose twice daily (2.97 mg/dose twice daily). Although the premature neonates received an actual dose 4-fold lower than the term babies on average, the oseltamivir carboxylate exposures were approximately 2-fold higher, compared with term babies (Figure 1). Data from both studies imply that dosing oseltamivir in premature babies, infants, and young children with use of a mg/kg approach may be more accurate in terms of achieving desired exposures relative to an age-based fixed dose regimen. Comparatively, adults with normal renal function receiving 75 mg twice daily achieve an average carboxylate AUC₁₂ of 2719 ng · h/mL [10].

This dataset is limited by a relative lack of robustness, because only 1 sample could be collected from each participant, and the mean concentration at each time point had to be modeled as if 1 subject had received the oseltamivir dose at steady-state. From Figure 1, it is clear the premature babies had similarly shaped concentration-time curves but increased oseltamivir carboxylate exposure relative to the older children. Several different approaches were taken to quantify this difference. We first applied a population pharmacokinetic approach to maximize prior knowledge of oseltamivir and oseltamivir carboxylate absorption and disposition in infants and young children. Because only 1 sample was collected per subject, obtaining individual post hoc pharmacokinetic parameter estimates was not feasible. We also used a noncompartmental analysis of the mean oseltamivir carboxylate data to compare with the CASG 114 data, and we estimated the mean steady-state concentration in both studies by simply averaging the raw oseltamivir carboxylate concentrations. All 3 approaches led to the similar conclusion that premature neonates achieved a mean oseltamivir carboxylate exposure approximately 2-fold higher than infants and young children, even though their dose was nearly 50% lower (1.73 vs 3.0 mg/kg/dose).

Oseltamivir is converted to its active metabolite, oseltamivir carboxylate, primarily by hepatic esterases. Oseltamivir carboxylate is then renally eliminated through both glomerular filtration and tubular secretion processes. Both processes are diminished in neonates and young children and do not reach adult capacity until 6–12 months of age [7]. Results from the current study are consistent with known aspects of developmental pharmacology and the ontogeny of drug disposition in that lower doses are required in this population to achieve exposures similar to that in older children and adults.

These modeled results, noncompartmental analysis, and mean of raw concentrations suggest an oseltamivir dose of approximately 1.0 mg/kg/dose (twice daily) should achieve oseltamivir carboxylate exposures in premature neonates (<38 weeks) comparable to that in infants and young children receiving 3 mg/kg/dose twice daily. Additional pharmacokinetic data from this age group are necessary, however, to more precisely define the optimal oseltamivir dose. Given that further 2009 H1N1 exposures may occur in NICU settings around the world as the pandemic continues or as other influenza viruses circulate, these data provide initial oseltamivir dosing guidance in premature babies, which is vital to the global public health response.

• Received January 7, 2010.
• Accepted March 17, 2010.

• © 2010 by the Infectious Diseases Society of America