Renal Function Estimation Using Serum Creatinine and Cystatin C in Critically Ill Pediatric Patients

First Author(s)

• 

  Gideon Stitt, BCPPS, BCCCP, PharmD,

  T32 Post-Doctoral Research Fellow in Clinical Pharmacology
  The Children's Hospital of Philadelphia
  Philadelphia, Pennsylvania

  Disclosure information not submitted.

  Slides and Audio

Co-Author(s)

• 

  Julie Fitzgerald, MD, PhD, FCCM

  Children's Hospital of Philadelphia
  Philadelphia, Pennsylvania

  Disclosure information not submitted.
• ML

  Marissa Leff

  Student Clinical Research Assistant
  Children's Hospital of Philadelphia, United States

  Disclosure information not submitted.
• 

  Kevin Downes

  Children's Hospital of Philadelphia
  Philadelphia, Pennsylvania

  Disclosure information not submitted.

Title: Renal Function Estimation Using Serum Creatinine and Cystatin C in Critically Ill Pediatric Patients

Introduction: Serum creatinine (SCr) has known confounders as a marker of renal function that may lead to drug over- or under-dosing. Cystatin C (CysC) is less influenced by body composition and gaining traction in practice as an alternative to SCr. We hypothesized that renal function estimates would be significantly lower using CysC compared to SCr.

Methods: We performed a retrospective study of PICU patients who had SCr and CysC measured simultaneously from 8/2020-6/2021. Estimated glomerular filtration rate (eGFR) was calculated using the bedside Schwartz (eGFR_SCr = 0.413*ht(cm)/SCr) and CKiD cystatin C (eGFR_CysC = 70.69*CysC^-0.931) equations. Correlation and agreement between GFR estimates were assessed using Spearman correlation and Bland-Altman tests, respectively. Factors associated with eGFR differences >30% were examined using a Fisher’s exact test and impacts of differences on drug dosing with descriptive statistics.

Results: 57 patients (60% male) had simultaneous SCr and CysC measurements during our study period. Median (IQR) age was 9.1 yrs (4.3-16.6), weight 24.4 kg (16.5-40), and BMI 18.5 (16.5-21.7). Most common PICU admission diagnoses included respiratory failure (56%), sepsis (19%), and need for neurologic monitoring (19%). Median eGFR_SCr and eGFR_CysC were 128 (85-183) and 71 (53-118) mL/min/1.73m² (p< 0.0001). Correlation between eGFR_SCr and eGFR_CysC was 0.575. Bland-Altman analysis showed a mean difference between eGFR_SCr and eGFR_CysC of -52 mL/min/1.73m² (95%CI -69 to -35). 35 (61%) patients had eGFR_SCr >30% higher than eGFR_CysC; no patient factors (age, weight, BMI, diagnoses) were associated with this difference. Assessing potential dosing impacts, 16 (28%) patients had concurrent eGFR_SCr >130 with eGFR_CysC < 130, 19 (33%) had eGFR_SCr >60 and eGFR_CysC < 60, and 6 (10.5%) had eGFR_SCr >130 with eGFR_CysC < 60 mL/min/1.73m².

Conclusions: eGFR_CysC was significantly lower than eGFR_SCr, and agreement between the two GFR estimates was poor. Use of SCr versus CysC may lead to substantial differences in GFR estimates in critically ill children, which can have a large impact on drug dosing.