Precision Dosing of Meropenem in a Neonate on CARPEDIEM^® Dialysis: A Grand Round

Abstract

Background:

Optimal antibiotic dosing is challenging in critically ill neonates because of the substantial pharmacokinetic variability, which is influenced by factors such as immature renal function, body composition, and critical illness. The use of extracorporeal therapies adds complexity, making predictions difficult in many cases. Meropenem, a broad-spectrum antibiotic, is commonly used due to resistant gram-negative organisms in neonates; however, dosing guidelines for neonates on CARPEDIEM^® dialysis are lacking.

Case Report:

This is a case of a neonate with liver failure of unclear etiology, who was on CARPEDIEM^® dialysis and started on meropenem for sepsis due to extended-spectrum beta-lactamase-producing Escherichia coli and suspected meningitis. Blood samples were sent to an external laboratory for meropenem concentration measurements, and model-informed precision dosing (MIPD) was used to guide the dosing adjustments. Initially, meropenem was administered at 40 mg/kg every 8 h with a 30-min infusion, resulting in exposures that exceeded those required to achieve free concentrations above four times the minimum inhibitory concentration for the entire dosing interval (100% fT>4xMIC). The dosing interval was adjusted to every 12 h to avoid unnecessarily high exposure. The regimen was continued without further complications, and the patient underwent successful liver transplantation.

Conclusion:

This case highlights the successful application of MIPD to individualize meropenem therapy in a critically ill neonate with liver failure on CARPEDIEM^® dialysis. MIPD is a valuable tool for dose adjustment in patients with unique and unpredictable pharmacokinetics.

CLINICIAN

A four-day-old ex-full-term female neonate was admitted for liver failure of unclear etiology. Prenatal history was notable for intrauterine growth restriction and positive maternal group B Streptococcus testing. The patient had progressive direct hyperbilirubinemia, hyperammonemia, and coagulopathy refractory to vitamin K.

On day four of admission, the patient developed fever, prompting peripheral blood and urine cultures and empiric nafcillin and gentamicin. Lumbar puncture was not performed because of coagulopathy. Blood culture revealed Escherichia coli resistant to ampicillin, ceftriaxone, ceftazidime, and cefepime. The BioFire^® FilmArray^® Blood Culture Identification Panel detected CTX-M type β-lactamase. Given the susceptibility profile and inability to rule out meningitis, the patient was transitioned from nafcillin and gentamicin to monotherapy with meropenem (minimum inhibitory concentration [MIC] <0.25 μg/mL) at the standard meningitis dose of 40 mg/kg/dose (102.8 mg based on birth weight of 2.57 kg) every 8 h infused over 30 min. Discussions on liver transplantation were ongoing.

On the following days, the patient had worsened hyperammonemia and severe acute kidney injury, undergoing ultrafiltration (aquapheresis; Smartflow^™; Nuwellis, Inc., Minneapolis, MN, USA) and escalation to continuous renal replacement therapy (CRRT). CARPEDIEM^® (Mozarc Medical, Minneapolis, MN, USA) was initiated as continuous venovenous hemodialysis with a dialysis flow rate of 10 mL/min, corresponding to a prescribed effluent flow rate of 5,700 mL/h/1.73 m² (higher than our AKI standard of 2,000 for ammonia clearance). Although the patient had sustained oliguria, serum creatinine levels and urine output improved before initiating CRRT (Table 1). No drug interactions affecting meropenem concentrations were identified after reviewing the patient’s medication list. We then attempted to determine the optimal meropenem dose for this patient.

Table 1.

Timeline of key clinical events, laboratory parameters, and dialysis-related metrics

Parameter	Day 0	Day 2	Day 4 Meropenem started	Day 5	Day 6 Aquapheresis	Day 7 CRRT started	Day 8	Day 9 PK Consult	Day 10	Day 12	Day 14	Day 16	Day 18
Serum creatinine (mg/dL)	0.42	-	-	0.45	0.83	0.37	0.23†	<0.15†	<0.15†	<0.15†	<0.15†	<0.15†	Liver Transplant
Ammonia (mcmol/L)	80	140	156	171	144	196	135	57	60	44	35	39
Urine output (mL/kg/h)	-	1.59	0.29	0.18	1.23	0.79	0.61	0.48	0.9	1.98	1.56	2.17
Dialysis flow rate (mL/min)	-	-	-	-	-	10	10	10	10	10	10	10
WBC x103/mcL	13.35	10.12	9.89	17.97	25.58	14.82	10.1	10.7	9.15	11.03	12.37	9.69

Abbreviations: CRRT, continuous renal replacement therapy; PK, pharmacokinetics; WBC, white blood cell count.

^†Creatinine levels are expected to be reduced after CRRT initiation and may not reflect renal function.

MODEL-INFORMED PRECISION DOSING (MIPD) CONSULTANT

Dosing meropenem in neonates is challenging because of its substantial pharmacokinetic (PK) variability due to factors such as immature renal function, diverse body composition, and pathological changes related to critical illness.¹

The use of CARPEDIEM^® increases complexity. This specialized CRRT system, designed for neonates and small infants, allows for precise ultrafiltration and small-solute clearance.² As a small and hydrophilic molecule with low protein binding (<2%), meropenem is susceptible to extracorporeal removal by CRRT.³ As there are no established guidelines for meropenem dosing in neonates on CARPEDIEM^®, clinicians often rely on empirical approaches or extrapolate from limited pediatric and adult data, which suggest that CRRT increases the risk of beta-lactam underexposure.^4-8

The patient presented with a unique clinical condition and unpredictable PK. Given these challenges and gaps in literature, we recommend measuring meropenem concentrations. Clinicians at our institution have access to a recently launched precision dosage consulting service for beta-lactam antibiotics, which can guide individualized treatment strategies.

CLINICIAN

Therefore, the clinical team decided to measure these concentrations. The decision was made collaboratively and involved the attending physician, infectious diseases and nephrology teams, bedside nurses, pharmacists, and the patient’s family. There was concern about the number of line entries and blood volume required from this neonate for PK modeling. Furthermore, there were not enough access points to administer extended infusions given the incompatibility of meropenem with the patient’s other medications. Thus, we determined that the optimal sampling strategy for meropenem is to accurately assess the patient’s meropenem exposure and guide appropriate dosing adjustments based on the above limitations.

MIPD CONSULTANT

We recommend collecting at least two 0.5 mL blood samples within the same dosing interval to capture the PK profile, allowing for a reasonable estimate of the volume of distribution and clearance: one sample was collected 15–30 min after the end of the infusion (near peak) and the other toward the end of the dosing interval (near trough).⁹ While multiple samples are ideal for modeling and simulating drug exposure, obtaining only one is acceptable.

The samples were packed in a temperature-controlled container and sent overnight to an external laboratory where meropenem concentrations were measured using liquid chromatography–tandem mass spectrometry.

CLINICIAN

We agreed that collecting two samples to measure meropenem concentrations was feasible. The patient had adequate venous access, and the required blood volume was not detrimental to the condition. Both the peak and trough samples were collected after the next meropenem dose.

MIPD CONSULTANT

The concentration results were obtained from the external laboratory within 36 h after the samples were sent. To prevent misinterpretation of the isolated concentrations, they were not immediately reported in the patient’s chart.

A comprehensive PK analysis (Table 2) was performed using a Bayesian software, MwPharm++ (Mediware, Prague, Czech Republic), which integrates the measured drug concentrations with patient-specific factors to estimate the drug concentration-time profile.¹⁰ The population PK model described by Saito et al. was used as a prior.⁶ Because the patient continued to urinate while on CARPEDIEM^® (Table 1), we used the creatinine value immediately prior to starting CARPEDIEM^® (0.37 mg/dL) to account for intrinsic kidney clearance.

Table 2.

Pharmacokinetic parameters estimated in a neonate receiving meropenem at 40 mg/kg every 8 h with 30 mi-infusion while on CARPEDIEM^® dialysis

Parameter	Estimated value
Peak concentration	40.8 mg/L
Trough concentration	6.1 mg/L
Volume of distribution	0.38 L/kg
Clearance	0.21 L/kg/h
AUC24h	324.5 mg*h/L
%fT>MIC of 0.25 mg/L	100%

Abbreviations: AUC_24h: Area under the concentration-time curve over a 24-h period; %fT>MIC: Percentage of time that the free drug concentration remained above the minimum inhibitory concentration

The current dose regimen of 102.8 mg (40 mg/kg) every 8 h ensured concentrations significantly above 4xMIC for 100% of the dosing interval, given the low MIC of the identified bacteria (<0.25 mg/L). The simulations suggested that reducing the dosing frequency to every 12 h would maintain 100% fT>4xMIC (Figure 1). Although there is no established toxicity threshold for meropenem concentrations, our dosing strategy prioritizes achieving effective exposure while avoiding unnecessarily high drug levels. The proposed regimen retains a standard 30-min infusion duration, as longer infusions improve plasma %fT>MIC but reduce %fT>MIC in the cerebrospinal fluid (CSF).¹¹ Although CSF meropenem concentrations could not be measured and are difficult to predict, data from the NeoMero study suggest that a 40 mg/kg dose achieving 100% fT>4xMIC in plasma for MICs up to 0.5 mg/L also achieves 100% fT>MIC in CSF.¹¹ This supports our recommendation because even after considering a safety margin, where the true MIC may be as high as 0.5 mg/L, the proposed regimen still ensures target attainment. We recognize that the patients in the NeoMero study did not receive CRRT, which justifies our use of MIPD to ensure adequate dosing. The recommendations and concentration results were uploaded to the patient’s chart after a phone discussion.

Figure 1.

Predicted concentration-time profiles of meropenem with current and suggested dosing regimens. Closed circles represent measured concentrations. Solid red line indicates predicted concentration-time profile. Left side of blue vertical dotted line shows profile predicted using current meropenem dosing regimen, whereas right side displays profile predicted using newly suggested dosing regimen. Fluctuations in concentrations observed around CRRT initiation (Day 7) reflect patient’s intermittent use of CRRT during initial days. Black dashed line represents target concentration of 1 mg/L (4 × MIC of 0.25 mg/L).

CLINICIAN

The dosing interval was adjusted every 12 h based on PK analysis. The patient continued the regimen without further infectious complications and underwent a successful liver transplantation. Meropenem-related adverse reactions, including neurotoxicity, were not observed.

CONCLUSION

In this case, meropenem was selected to treat extended-spectrum beta-lactamase-producing E. coli in a critically ill neonate with liver failure and kidney injury on CARPEDIEM^® dialysis, a clinical scenario with unpredictable PK. We used an MIPD approach to individualize meropenem dosing to ensure optimal exposure.

Our individualized PK assessment, combined with Bayesian simulations for optimal dosing, enabled the clinical team to reduce the dosing frequency to every 12 h and avoid unnecessarily high concentrations. Notably, the relatively low MIC of the identified bacteria (<0.25 mg/L) facilitated this dosing strategy. If the MIC was higher, close to, or at the susceptibility breakpoint, more frequent dosing or extended infusions would have been necessary to achieve the target.

This case emphasizes the value of precise dosing in neonates, especially in the context of critical illness and extracorporeal support and provides an example of how model-informed dosing strategies can improve patient outcomes while minimizing the potential risks of overexposure, even in institutions without ready access to beta-lactam assays.