Health Condition 1: A419- Sepsis, unspecified organism Health Condition 2: A490- Staphylococcal infection, unspecified site
Conditions
Interventions
Intervention1: Nil: Nil
Intervention2: Vancomycin: Intravenous Vancomycin administered as per standard ICU protocol.
Loading Dose: 25 to 30 mg per kg Total Body Weight (maximum 3000 mg)
infused over
Sponsors
AIIMS Bhubaneswar
Eligibility
Inclusion criteria
Inclusion criteria: 1. Adult patients aged 18 years or above admitted to medical or central ICU receiving intravenous vancomycin for suspected or documented Gram-positive infection. 2. Minimum anticipated duration of vancomycin therapy of 48 hours or more. 3. Written informed consent obtained from patient or legally authorized representative.
Exclusion criteria
Exclusion criteria: 1. Pregnant patients. 2. Patients with burns involving more than 20 percent of Body Surface Area (BSA). 3. Patients on Extracorporeal Membrane Oxygenation (ECMO). 4. Patients with known anaphylaxis to vancomycin. 5. Patients requiring renal replacement therapy or dialysis.
Design outcomes
Primary
| Measure | Time frame |
|---|---|
| Development and validation of a two-compartment population pharmacokinetic (PopPK) model for vancomycin in adult ICU patients incorporating inflammatory markers as covariates. Primary pharmacokinetic parameters to be estimated: 1. Vancomycin Clearance (CL) in L/h 2. Central Volume of Distribution (Vc) in L 3. Peripheral Volume of Distribution (Vp) in L 4. Intercompartmental Clearance (Q) in L/h 5. Area Under the Concentration-Time Curve over 24 hours (AUC24) in mg.h/L Inflammatory covariates to be evaluated: C-reactive protein (CRP), Interleukin-6 (IL-6), Procalcitonin, and Fibrinogen.Timepoint: Pharmacokinetic sampling performed over the first 48 hours of vancomycin therapy. Day 1: 0.5 hours (mid-infusion), 1 hour (end of infusion), 2 hours (post-infusion), 6 hours (late distribution phase), 12 hours (pre-maintenance trough). Day 2: 24 hours (steady-state trough), 25 hours (steady-state peak), 30 hours (mid-interval). Final outcome assessment at end of vancomycin therapy (up to 7 days). | — |
Secondary
| Measure | Time frame |
|---|---|
| Comparison of predictive performance of Physics-Informed Neural Networks (PINNs) and other neural network architectures against traditional population pharmacokinetic (PopPK) model for vancomycin trough concentration and AUC24 estimation in adult ICU patients. Performance metrics: Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), R-squared, and percentage of predictions within plus or minus 20 to 30 percent of observed values.Timepoint: Model training and validation performed on pharmacokinetic data collected over first 48 hours (Day 1 and Day 2 sampling). Final model comparison performed at end of study (Month 18 to 24).;Probability of target attainment (PTA) for vancomycin AUC24/MIC ratio of 400 to 600 mg.h/L assessed using Monte Carlo simulations (n=5000). Development of ICU-specific dosing nomograms tailored to renal function (eGFR) and inflammatory status (CRP, IL-6).Timepoint: Monte Carlo simulations performed after final PopPK model validation. Nomograms developed during analysis phase (Month 13 to 18). Target attainment evaluated based on PK data from Days 1 to 7 of vancomycin therapy.;Assessment of vancomycin-associated nephrotoxicity as a safety outcome, defined as acute kidney injury (AKI) of KDIGO stage 1 or above, occurring within 7 days of vancomycin initiation. Monitored by serial serum creatinine and eGFR measurements at baseline (Day 0), Day 4, Day 5, and Day 7.Timepoint: Baseline (Day 0): Pre-vancomycin serum creatinine and eGFR. Day 4 and Day 5: Serial renal function monitoring. Day 7: Final renal function assessment (if therapy extended). End of therapy: Final safety outcome recorded.;Comparative evaluation of predictive performance of one-compartment versus two-compartment pharmacokinetic models for vancomycin using objective function value (OFV), goodness-of-fit diagnostics, visual predictive checks (VPC), normalised prediction distribution errors (NPDE), and bootstrap confidence intervals.Ti | — |
Countries
India
Contacts
Public ContactAnand Srinivasan
AIIMS Bhubaneswar
Outcome results
None listed