207 Pharmacokinetics of vancomycin in patients with continuous flow left-ventricular assist devices

Tuesday, October 23, 2012
Westin Diplomat Resort
Charles T. Makowski, PharmD1, Rachel M. Chambers, PharmD, BCPS, (AQ-ID)2 and Douglas L. Jennings, PharmD, BCPS, (AQ-CV)2
1Wayne State University Eugene Applebaum College of Pharmacy, Detroit, MI
2Henry Ford Hospital, Detroit, MI

Purpose:   To describe the pharmacokinetics of vancomycin in patients with continuous flow left-ventricular assist devices (CF-LVADs).

Methods:   This IRB-approved retrospective, descriptive analysis was conducted at an urban, tertiary hospital.  Eligible patients were ≥18yrs old, implanted with a HeartMateII CF-LVAD during January2008-April2012, and treated with vancomycin ≥48hrs for infection.  Key exclusion criteria include <2 vancomycin levels, ≥Stage1 AKIN acute kidney injury (change in Scr ≥0.3mg∙dL-1 or ≥1.5∙baseline or urine output <0.5mL∙kg-1∙hr-1 for ≥6hrs), acute heart failure exacerbation, hemodynamic instability, and surgery ≤5d before initiating vancomycin.  Methods for estimating first-order elimination rate constant (Ke)(Table1) and volume of distribution (Vd)[using ideal (IBW), adjusted (AdjBW), and actual (ABW) body weights]  were compared with the actual Ke and Vd.  Actual pharmacokinetic parameters were calculated from steady-state peak and trough vancomycin levels using one-compartment model equations.  Paired t-test or signed-rank test was used. 

Results:   Twelve patients were included (age 44.9±15 years, 91.7% male, 58.3% obese, CLcr 79.2±27 mL∙min-1).  Common reasons for exclusion were <2 vancomycin levels drawn (n=65) and unstable renal function or hemodynamics (n=4).  Common treatment indications (n≥2) were health-care associated pneumonia (41.7%), driveline infection (25%), and sepsis (16.7%). Methods CLcr1 and CLcr2 for estimating Ke were highly correlated with actual Ke (r=0.78 and 0.79, respectively; p<0.01).  CLcr3a tended to overestimate actual Ke for obese patients [0.0254hr-1 (-0.0001-0.051hr-1; p=0.051)].  Obese patients had a lower-than-expected actual Vd, which was numerically less compared with non-obese patients (0.44L∙kg-1ABW vs. 0.57L∙kg-1ABW; p=0.17).  Therefore, EstVd_ABW overestimated actual Vd for obese patients [23.2L (7.5-38.9L; p=0.028)].

Conclusion: General population methods may accurately estimate the pharmacokinetic parameters of vancomycin for stable non-obese patients with CF-LVADs.  Obese patients may require use of AdjBW or IBW for an accurate prediction of Vd.

Table1.  Methods for estimating Ke*
Method ABW∙IBW-1 Designated body weight for CLcr equation†,‡

CLcr1†

any IBW

CLcr2†

72kg

CLcr3a‡

≥1.3 (obese) ABW

CLcr3b†

AdjBW
*Ke=0.00117∙CLcr+0.03.  †Cockcroft-Gault‡Salazar-Corcoran.