Cystatin C Outperforms Creatinine in Predicting Cefepime Clearance in Pediatric Stem Cell Transplant Recipients.

ABSTRACT

Pediatric hematopoietic stem cell transplant (HSCT) patients are at risk of developing both sepsis and altered kidney function. Cefepime is used for empiric coverage post-HSCT and requires dose adjustment based on kidney function. Since cefepime's antimicrobial efficacy is determined by the time free concentrations exceed bacterial minimum inhibitory concentration (MIC), it is important to assess kidney function accurately to ensure adequate concentrations. Serum creatinine (SCr) is routinely used to estimate glomerular filtration rate (eGFR) but varies with muscle mass, which can be significantly lower in HSCT patients, making SCr an inaccurate kidney function biomarker. Cystatin C (CysC) eGFR is independent of muscle mass, though steroid use increases CysC. Objectives of this study were to describe how eGFR impacts cefepime pharmacokinetic/pharmacodynamic (PK/PD) target attainment in pediatric HSCT patients, to investigate which method of estimating GFR (SCr, CysC, combined) best predicts cefepime clearance, and to explore additional predictors of cefepime clearance. Patients admitted to the pediatric HSCT unit who received ≥2 cefepime doses were prospectively enrolled. We measured total cefepime peak/trough concentrations between the second and fourth cefepime doses and measured SCr and CysC if not already obtained clinically within 24h of cefepime samples. eGFRs were calculated with Chronic Kidney Disease in Children U25 equations. Bayesian estimates of cefepime clearance were determined with a pediatric cefepime PK model and PK software MwPharm++. Simple linear regression was used to compare cefepime clearance normalized to body surface area (BSA) to BSA-normalized SCr-, CysC-, and SCr-/CysC-eGFRs, while multiple linear regression was used to account for additional predictors of cefepime clearance. For target attainment, we assessed the percentage of time free cefepime concentrations exceeded 1x MIC (%fT>1x MIC) and 4x MIC (%fT>4x MIC) using a susceptibility breakpoint of 8 mg/L for Pseudomonas aeruginosa. We enrolled 53 patients (ages 1 to 30 years, median 8.9 years). SCr- and CysC-eGFRs were lower in patients who attained 100% fT>1xMIC compared to those who did not attain this target 115 versus 156 mL/min/1.73m2 (p = .01) for SCr-eGFR and 73.5 versus 107 mL/min/1.73m2 (p < .001) for CysC-eGFR. SCr-eGFR was weakly positively correlated with cefepime clearance (adjusted [a]r2= 0.14), while CysC-eGFR and SCr-/CysC-eGFR had stronger positive correlations (ar2 = 0.30 CysC, ar2 = 0.28 combo. There was a weak, significant linear association between increasing CysC-eGFR and decreased %fT>1xMIC (ar2 = 0.32) and %fT>4xMIC (ar2 = 0.14). No patients with a CysC-eGFR >120 mL/min/1.73 m2 achieved 100% fT>1xMIC or 50% fT>4x MIC. In multiple regression models, underlying diagnosis of hemoglobinopathy (in all models) and being pretransplant (in SCr and combined models) were associated with increased cefepime clearance, while concomitant use of calcineurin inhibitors was associated with decreased cefepime clearance in all models. Overall, the combo-eGFR model with timing pretransplant, hemoglobinopathy, and use of calcineurin inhibitors had the best performance (ar2 = 0.63). CysC-based eGFRs (CysC alone and combined) predicted cefepime clearance better than SCr-eGFR, even after considering steroid use. Increasing CysC eGFR correlated with decreased probability of PD target attainment, raising concerns for underdosing at high eGFRs. CysC should be included when estimating kidney function to provide adequate dosing of cefepime in pediatric HSCT patients.