Published Online, 2 June 2009, www.theannals.com, DOI 10.1345/aph.1L144b.
The Annals of Pharmacotherapy: Vol. 43, No. 6, pp. 1146. DOI 10.1345/aph.1L144b
© 2009 Harvey Whitney Books Company.
Authors' Reply
P Brandon Bookstaver, PharmD BCPS AAHIVE
Clinical Assistant Professor, SC College of Pharmacy, University of
South Carolina Campus, 715 Sumter St., Columbia, SC 29208, fax 803/777-2820,
bookstaver{at}sccp.sc.edu
James W Johnson, PharmD BCPS
Clinical Coordinator, Infectious Diseases, Department of Pharmacy,
Wake Forest University Baptist Medical Center, Winston-Salem, NC
David Stewart, PharmD BCPS
Clinical Assistant Professor, College of Pharmacy, East Tennessee
State University, Johnson City, TN
John C Williamson, PharmD BCPS
Clinical Coordinator, Infectious Diseases, Department of Pharmacy,
Wake Forest University Baptist Medical Center
Published Online, June 2, 2009. www.theannals.com, DOI 10.1345/aph.1L144b
We appreciate the comments submitted by Goutelle et al. concerning our
article and would like to respond to these observations. They have noted a few
specific concerns regarding the methods used in calculating both
patient-specific and population estimated values. They point out the
discrepancy in the formula we used (ke = 0.0024 x CrCl
+ 0.01) to estimate ke with those included in the article
cited by Sawchuk and
Zaske.1 It
was our intent to use this article to reference the overall concept and model
only. The method in this reference describes modeling in burn patients, which
would not be applicable to our medical surgery population. The method for
deriving a ke estimate from creatinine clearance by linear
regression was first proposed by Dettli et
al.2 The
Dettli method has been shown to be as good as other estimates for initial
gentamicin
dosing.3
Goutelle et al. state that "...such equations with nonvariable
coefficients are likely to provide poor parameter estimates because
aminoglycoside pharmacokinetic variability is inevitably
underestimated." We agree that nonvariable coefficients are not optimal;
however, we disagree that they are likely to provide poor estimates. The
method used in our study is an accepted routine clinical practice for
determining aminoglycoside doses during patient care in lieu of aminoglycoside
concentrations for individualized dosing. As such, we believe it represents an
appropriate method for this study.
It is true that both equations overestimate ke. However,
the MDRD formula offered a more precise estimate of patient-specific values as
represented by lower Akaike's Information Criterion (AIC) values for both
aminoglycoside ke (MDRD: AIC = 99.2; modified
Cockcroft-Gault [CGm]: AIC = 101.8) and clearance (MDRD: AIC
= 71.0; CGm: AIC = 82.3). In addition, Goutelle et al.
comment that the fact that the Vd was fixed at 0.3 L/kg was
an influence of clearance estimation. It is true that this was a fixed
variable in the a priori model; however, it was not fixed when the 2
predictive formulas were compared with patient-specific elimination and
clearance, as the median Vd was 24.98 ± 9.08 L or
0.32 ± 0.15 L/kg. Using a fixed Vd of 0.2 L/kg, as
suggested, would not have been appropriate in our population, as the majority
of our patients were housed in an intensive care unit. Furthermore, using a
Vd of 0.4 L/kg results in an additional 25% difference in
clearance estimations between the 2 formulas.
We appreciate the comments expressed by Goutelle et al. but maintain our
original conclusions regarding utilization of the MDRD formula as an estimate
of clearance for aminoglycoside dosing in routine clinical practice.
Individualized dosing through patient-specific aminoglycoside concentrations
is optimal for dose modification.
References
- Sawchuk RJ, Zaske DE. Pharmacokinetics of dosing regimens which
utilize multiple intravenous infusions: gentamicin in burn patients. J
Pharmacokinet Biopharm 1976;4:183-95.[CrossRef][Medline]
- Dettli RC. Drug dosage in patients with renal disease. Clin
Pharmacol Ther 1974;16:274-80.[Medline]
- Lesar TS, Rotschafer JC, Strand LM, Solem LD, Zaske DE. Gentamicin
dosing errors with four commonly used nomograms. JAMA 1982;10:1190-3.
