In November 2023, the Dutch Pediatric Formulary recommended a once-daily intravenous gentamicin dose of 4 mg/kg for neonates and 7 mg/kg for children ≥1 month of age for the treatment of infections such as pneumonia, urinary tract infections, and sepsis.

Rapid changes in body composition and organ maturation may have an effect on the gentamicin plasma level across the pediatric age span. In addition, there is a shift in the therapeutic target above 1 month of age: in neonates it is aimed for a maximum plasma concentration between 8 - 12 mg/L and between 15 - 20 mg/L in older infants.

The following questions arise when looking at the dosing guideline:
·       Is the gentamicin plasma level adequate when 4 mg/kg is given to neonates and 7 mg/kg to children ≥1 month of age?
·       Is this sudden dose increase at the age of 1 month appropriate?
·       Is a more gradual increase in dosing more appropriate from a pharmacokinetic perspective?

These questions are addressed with physiologically-based pharmacokinetic (PBPK) computer model simulations. A PBPK model can be seen as a computerized pediatric body receiving a drug. The effects of development and maturation on the organs are described in the model. Also, physical and chemical properties of gentamicin are incorporated. Together, the model can simulate the distribution, metabolism and excretion of gentamicin in a pediatric virtual population with the aim to establish model-informed dosing recommendations for term neonates and infants up to 2 years of age.

Pediatric gentamicin dose recommendation in the Netherlands: link
Dose-finding simulations of gentamicin in pediatrics by De Hoop-Sommen and colleagues: link

The physiologically-based pharmacokinetic (PBPK) computer model that was used in this pediatric dose-finding study is constructed in the Simcyp™ Simulator (v21). The approach of this study can be seen as efficient and pragmatic because there was no need to create the models from scratch; the default pediatric population model from the Simcyp™ software was used and the gentamicin model was retrieved from literature (Abduljalil et al. 2020, PubMed ID: 31587145).

Model performance was assessed through verification activities. To this end, pharmacokinetic data from intravenous single and multi-dose clinical studies with adult (eight studies in total) and pediatric subjects (21 studies in total, of which 3 studies in preterm neonates) were used. Visual predictive checks reveal that the plasma concentration-time curve predicted by the model is comparable to what have been observed clinically, both for adult and pediatric subjects. For adults, 82% of all predicted pharmacokinetic parameters (e.g., volume of distribution and clearance) were between 2-fold lower and 2-fold higher than the observed value. For pediatric patients this percentage was as high as 91%.

Based on these results, PBPK model credibility is considered ‘satisfactory’.

Part 1 of the MID framework provides a more detailed description of model credibility assessment. 

Pediatric gentamicin dose recommendation in the Netherlands: link
Dose-finding simulations of gentamicin in pediatrics by De Hoop-Sommen and colleagues: link

What we aimed for
The most optimal dosing regimen had to fulfill the following requirements:
-   The maximum plasma concentration (C_max) should be between 8 - 12 mg/L in neonates and between 15 - 20 mg/L in infants up to 2 years of age
-   The concentration reached by a drug immediately before the next dose (trough plasma concentration, C_trough) should be below 1 mg/L in children of all ages 

How this is done
A physiologically-based pharmacokinetic (PBPK) computer model was used to first:
1)   predict gentamicin pharmacokinetics (C_max and C_trough) in pediatric subjects, with the following dosing regimens as provided by the Dutch Pediatric Formulary in November 2023:
          - neonates: 4 mg/kg every 24h (intravenous infusion over 30 min)
          - infants 1 month - 2 years of age: 7 mg/kg every 24h (intravenous infusion over 30 min)

and subsequently:
2)   simulate alternative dosing strategies to reach the intended gentamicin C_max and C_trough. We used this information to decide which dosing regimens will most likely result in the desired gentamicin plasma concentration.

What is found
Current dose advice – Term neonates
With the current dose advice (4 mg/kg every 24h), the PBPK computer model predictions show that an adequate C_max in term neonates up to 1 month of age is reached. This is visualized in Figure A below, where black squares (indicating 4 mg/kg every 24h) show that the C_max falls within the desired plasma concentration range for term neonates. A therapeutic effect can therefore be expected.

While the C_max falls within the desired range, the C_trough level is not optimal. In Figure B, it can be seen that the 24h C_trough level is too high (>1 mg/L) in the majority of the term neonates of 1 day (postnatal age) when 4 mg/kg is given every 24h (black squares). It should be noted that the risk of having such high C_trough levels decreases with increasing postnatal age (all neonates of 28 days have an adequate C_trough level).

Current dose advice – Infants up to 2 years of age
With the current dose advice (7 mg/kg every 24h), C_max levels are expected to be within the therapeutic range in infants up to 2 years of age. This is visualized in Figure E, where black squares represent the 7 mg/kg every 24h dosing schedule. Though, the 5^th percentile for infants ≥21 months of age just dipped below the lower limit, meaning that a few infants of this age may in fact have C_max levels below the lower limit of the effective range (<15 mg/L).

Figure F shows that a large proportion of the infants around 1 month of age have suboptimal C_trough levels (>1 mg/L). Adequate C_trough levels are seen in older age (infants ≥2 months).

Taken together, these predictions show that the gentamicin dosing recommendations for term neonates and infants up to 2 years of age can be improved.

Optimized dosing schedules
Subsequent PBPK computer model predictions with various alternative dosing strategies showed that the total dose does not need to be changed (4 mg/kg) but that the dosing interval for term neonates up to 6 weeks of age should be extended from once every 24h to 36 - 48h to keep C_trough levels below 1 mg/mL. Figure B-D show that a higher proportion of term neonates have adequate C_trough levels (<1mg/L) when the dosing interval is extended, as depicted by the trough levels at 35.5h and 47.5h as compared to 23.5h.

For infants, a dose of 7.5 mg/kg every 24 hours will result in adequate C_max levels. This suggestion is based on modeling predictions visualized in Figure E, showing a more adequate C_max with 7.5 mg/kg every 24h (red squares) as compared to 5 mg/kg every 24h (black squares).

Predicted gentamicin plasma levels in different age groups upon various dosing strategies. Predicted C_max and C_trough levels and corresponding 5^th and 95^th percentiles are shown. Graph A shows C_max levels after doses of 3, 3.5, 4, 4.5, and 5 mg/kg for neonates and graph E shows C_max levels after doses of 6, 7, 7.5, 8, and 9 mg/kg for infants. Graph B - D show C_trough levels at 23.5, 35.5, and 47.5 hours after start of infusion for neonates and graph F shows C_trough levels for infants until the age of 6 months; older infants all have C_trough levels below 1 mg/L and are therefore not depicted. Black squares (⯀) in graphs A - D and black diamonds (◆) in graphs E - F represent the Dutch Pediatric Formulary dose of November 2023. Abbreviations: C_max, maximum plasma concentration; C_trough, trough plasma concentration; PNA, postnatal age. 

Predicted pharmacokinetic parameter values for defined age groups are shown in the table below. 

Abbreviations: CL, clearance; C_max, maximum plasma concentration; tmax, time to maximum plasma concentration; Vd, volume of distribution; t_1/2, elimination half-life.

Part 2 of the MID framework provides a more detailed description of model credibility assessment. 

Note that dosing regimens as suggested by modeling do not necessarily represent final dosing recommendations for clinical practice. Final dosing guidelines are the result of a careful assessment of the benefits and risks of the model-informed dose in the context of all available clinical evidence and experience of healthcare providers. Consult the website of the Dutch Pediatric Formulary for gentamicin dosing recommendations for clinical practice.  

Pediatric gentamicin dose recommendation in the Netherlands: link
Dose-finding simulations of gentamicin in pediatrics by De Hoop-Sommen and colleagues: link

With physiologically-based pharmacokinetic (PBPK) computer modeling, we explored gentamicin plasma levels over time with various dosing strategies. We used this information to decide which dosing regimens are most appropriate for the age groups of interest.

We propose age-range appropriate dosing regimens, which are depicted in the table below. More information on how these model-informed dosing regimens are selected can be found in the section ‘Dose-finding simulations’.

Note that the dose regimens displayed here concern an intravenous infusion over 30 minutes. Abbreviations: DPF, Dutch Pediatric Formulary, PBPK, physiologically-based pharmacokinetic. 

Note that dosing regimens as suggested by modeling do not necessarily represent final dosing recommendations for clinical practice. Final dosing guidelines are the result of a careful assessment of the benefits and risks of the model-informed dose in the context of all available clinical evidence and experience of healthcare providers. Consult the website of the Dutch Pediatric Formulary for gentamicin dosing recommendations for clinical practice.  

Pediatric gentamicin dose recommendation in the Netherlands: link
Dose-finding simulations of gentamicin in pediatrics by De Hoop-Sommen and colleagues: link

Implementing the model-informed dose (see section 'Model-informed dose') in clinical practice is not straightforward; the following aspects should be taken into consideration.

1.    Translation of model-informed dose to practical dose

The model-informed dosing recommendations for gentamicin found in this study are practical, but entail dosing recommendations for four pediatric age groups (<3 weeks, 3 to <4 weeks, 4 to <6 weeks, 6 weeks to 2 years) instead of two (term neonate vs. children 1 month to 12 years of age). There are no difficulties expected regarding dose calculation and there are also no issues with respect to drug formulation or excipient safety.

2.    A different susceptibility of the micro-organism involved may require a different dose

The model-informed dosing regimens for gentamicin are based on the assumptions 1) that the micro-organism causing the infection has a certain minimum inhibitory concentration (MIC) and 2) that a certain ratio of the maximum plasma concentration (C_max)/MIC should be reached during the dosing period for the drug to be effective. The MIC as well as the C_max/MIC ratio are derived from literature (high level of evidence supporting this). The model-informed dose is appropriate for the treatment of neonatal infections if the MIC of the micro-organism involved is <1 mg/L and targeting a C_max between 8 - 12 mg/L. For infants up to 2 years of age, the model-informed dose is considered appropriate when the infection is caused by a micro-organism with an MIC of ≤2 mg/L and targeting a C_max of 15 - 20 mg/L. As such, the model-informed dose may not be appropriate when the MIC value is different (e.g., other type of micro-organism involved) and/or when the target C_max/MIC ratio is different (e.g., because of novel insights. The most optimal dosing regimen may in that case deviate from the model-informed dosing regimen.

Part 3 of the MID framework provides a more detailed description of model credibility assessment. 

Pediatric gentamicin dose recommendation in the Netherlands: link
Dose-finding simulations of gentamicin in pediatrics by De Hoop-Sommen and colleages: link

Pediatric gentamicin dosing recommendation for clinical practice are provided by the European Pediatric Formularies:
The Netherlands:   https://www.kinderformularium.nl/geneesmiddel/23/gentamicine          link
Germany:                   https://kinderformularium.de/monographie/5173/gentamicin                   link
Austria:                       https://kindermedika.at/monographie/10520/gentamicin                             link
Norway:                     https://www.koble.info/legemiddel/8672/gentamicin-parenteral               link
 
Publication of this study by De Hoop-Sommen et al. 2023: “Pragmatic physiologically-based pharmacokinetic modeling to support clinical implementation of optimized gentamicin dosing in term neonates and infants: proof-of-concept”. Front Pediatr. 2023. Nov 21:11:1288376. doi: 10.3389/fped.2023.1288376.

Version 1. Nov 2023.