1998 Vancomycin HD

8/22/2019 1998 Vancomycin HD

1/9

Determinants of Vancomycin Clearance by Continuous VenovenousHemofiltration and Continuous Venovenous Hemodialysis

Melanie S. Joy, PharmD, Gary R. Matzke, PharmD, FCCP, FCP, Reginald F. Frye, PharmD, PhD,and Paul M. Palevsky, MD

The clearance of vancomycin is significantly reduced in patients with acute, as well as, chronic renal failure.

Although multiple-dosage regimen adjustment techniques have been proposed for these patients, there is little

quantitative data to guide the individualization of vancomycin therapy in acute renal failure patients who are

receiving continuous renal replacement therapy (CRRT). To determine appropriate vancomycin dosing strategies

for patients receiving continuous venovenous hemofiltration (CVVH) and continuous venovenous hemodialysis

(CVVHD), we performed controlled clearance studies in five stable hemodialysis patients with three hemofilters: an

acrylonitrile copolymer 0.6 m2 (AN69), polymethylmethacrylate 2.1 m2 (PMMA), and polysulfone 0.65 m2 (PS).

Patients received 500 mg of vancomycin intravenously at least 12 hours before the start of the clearance study. The

concentration of vancomycin in multiple plasma and dialysate/ultrafiltrate samples was determined by EMIT (Syva,

Palo Alto, CA). The diffusional clearance and sieving coefficient (SC) of vancomycin were compared by a

mixed-model repeated-measures analysis of variance (ANOVA) with filter and blood (QB), dialysate inflow (QDI), or

ultrafiltration rate (QUF) as the main effects and patient as a random effect. Vancomycin was moderately protein

bound in these patients; free fraction ranged from 49% to 83%. The SCs of the three filters were similar and

significantly correlated with the free fraction of vancomycin (P 0.01; r2 0.465). Significant linear relationships

were observed between the diffusional clearance of vancomycin and QDI for all three filters: AN69 (slope 0.482;

r2 0.880); PMMA (slope 0.853; r2 0.966); and PS (slope 0.658; r2 0.887). Theslope of this relationship forthe

PMMA filter was significantly greater than that of the AN69 and PS filters. The clearance of vancomycin, urea, and

creatinine, however, was essentially constant at all QBs for all three filters. Thus, the clearance of vancomycin was

not membrane dependent during CVVH. However, during CVVHD, membrane dependence of vancomycin clearance

was noted at a QDI greater than 16.7 mL/min; vancomycin clearance with PMMA at a QDI of 25 mL/min was 66% and

43% greater than that with the AN69 and PS filters, respectively. CVVH (62% to 262%) and CVVHD (90% to 540%) can

significantly augment the clearance of vancomycin in acute renal failure patients. Dosing strategies for individual-

ization of vancomycin therapy in patients receiving CVVH and CVVHD are proposed.

1998 by the National Kidney Foundation, Inc.

INDEX WORDS: Vancomycin; continuous renal replacement therapy; pharmacokinetics.

CONTINUOUS RENALreplacement therapy(CRRT) is frequently used in the intensivecare setting to manage hemodynamically un-stable patients who are fluid overloaded or have

acute renal failure.1-3 Continuous venovenous

hemofiltration (CVVH) and continuous venove-

nous hemodialysis (CVVHD) are two of the

most commonly used CRRT methods. They pri-

marily use ultrafiltration/convection or diffusion,

respectively, for solute removal.4 Continuous ve-

novenous hemodiafiltration (CVVHDF), which

uses diffusion as well as convection, may also be

used, particularly for hypercatabolic patients.

Drug clearance by CVVH is dependent on the

ultrafiltration rate (QUF

) and the sieving coeffi-

cient (SC) for the particular solute or drug of

interest.5 The clearance of medications by

CVVHD is dependent on the dialysate flow rate

because solute and/or drug removal is primarily

diffusive.6 CVVHDF combines diffusion with

convection and, thus, the degree of solute/drug

clearance may be greater than can be attained

with either of the other two CRRT therapies. In

addition to the QB and dialysate flow rate, the

removal of solutes/drugs by CRRT may be depen-

dent on patient factors, eg, degree of binding toplasma proteins or the type of hemofilter used.7

The clearance of small and large molecules has

been reported to vary markedly between hemofil-

ters even when all other procedural variables are

From the Division of Nephrology and Hypertension, Uni-

versity of North Carolina, School of Medicine, Chapel Hill,NC; Schools of Medicine and Pharmacy, Center for Clinical

Pharmacology, University of Pittsburgh; and the RenalSection, Veterans Administration Pittsburgh Health Care

System, Pittsburgh, PA.Received September 25, 1997; accepted in revised form

January 16, 1998.Supported in part by grant no. 5M01 RR00056 from the

National Institutes of Health, National Center for Research

Resources/General Clinical Research Center, Bethesda, MD.Address reprint requests to Gary R. Matzke, PharmD,

FCCP, FCP, University of Pittsburgh, School of Pharmacy,724 Salk Hall, Pittsburgh, PA 15261. E-mail: matzke@

pitt.edu

1998 by the National Kidney Foundation, Inc.0272-6386/98/3106-0017$3.00/0

American Journal of Kidney Diseases, Vol 31, No 6 (June), 1998: pp 1019-1027 1019


2/9

held constant.8,9 Finally, the use of pump-driven

systems, ie, CVVH and CVVHD, may enhance

drug clearance because of the consistency of

blood flow compared with continuous arteriove-

nous hemofiltration and continuous arteriove-nous hemodialysis.

Although middle-molecular-weight (1,000 to

5,000 d) agents are poorly removed by conven-

tional hemodialysis, effective removal by CRRT

methods has been documented. Vancomycin, a

prototypical middle-molecular-weight drug (1,448

d), is approximately 55% protein bound (range,

15% to 75%) and has a volume of distribution of

about 0.7 L/kg (range, 0.3 to 0.9 L/kg) in patients

with normal renal function.10 Renal elimination

of unchanged drug comprises 90% of the total-

body clearance, and the half-life of vancomycin

increases significantly in patients with renal insuf-ficiency. Thus, the primary patient factor that

could contribute to the variability in clearance by

CRRT is the degree of protein binding. The

reported clearance of vancomycin during CVVH,

CVVHD, CVVHDF, or their arterial derivatives

ranges from 4.2 to 24.0 mL/min.5,11-19 Unfortu-

nately, many of these studies had limited statisti-

cal power because of small sample size (n 1 to

3); poorly defined CRRT conditions, ie, dialy-

sate, QUF, and QB rates, hemofilter type, length of

therapy; or lack of documentation of adequacy of

removal of a reference solute, ie, urea or creati-

nine. Furthermore, the calculative methods used

to determine the clearance of vancomycin were

often not provided or consisted of multiple as-

sumptions (eg, normal degree of protein binding,

consistency of QB, dialysate, and QUF rates). This

study was, therefore, designed to rigorously

evaluate the clearance of the prototype middle

molecule, vancomycin, by CVVH and CVVHD

in stable end-stage renal disease (ESRD) patients

to assess the impact of the critical procedural and

patient variables on drug clearance.

METHODS

Eight patients with ESRD who were receiving conven-

tional maintenance hemodialysis participated in this study

after giving written informed consent. The study was ap-

proved by the Biomedical Institutional Review Board and

the General Clinical Research Center Committee of the

University of Pittsburgh. The clearance of vancomycin by

CVVH and CVVHD was determined during a 12-hour pro-

cedure (see details below) for each of the three hollow-fiber

hemofilters that were evaluated. These included a 0.6-m2

acrylonitrile and sodium methalyl sulfonate copolymer (AN69)

hemofilter (Hospal Multiflow 60; CGH Medical, Lyon,

France); a 2.1-m2 polymethylmethacrylate (PMMA) hemo-

filter (Filtryzer B1-2:1U; Toray Industries, Tokyo, Japan)

and a 0.65-m2

polysulfone (PS) hemofilter (Fresenius F40;Fresenius AG, Bad Homburg, Germany). These filters, al-

though varying in surface area, were selected because they

provide similar ultrafiltration coefficient (KUF) of 15 to 20

mL/mm Hg per hour. Five clearance studies were performed

with each hemofilter; two patients completed one procedure,

five patients completed two procedures, and one patient

completed three clearance studies (one with AN69 and two

with PMMA filters). Each 12-hour CRRT procedure was

performed in addition to the patients regularly scheduled

hemodialysis treatments.

CRRT Procedure

Venous access was obtained by cannulation of the pa-

tients hemodialysis arteriovenous fistula or graft. The inlet

and outlet ports of the filter were connected to the patientthrough CVVH tubing. QB was regulated by the use of a

roller pump (Sarns, Ann Arbor, MI). An air detector with an

automatic pump shut-off was located distal to the drip

chamber on the venous return. Dialysate was pumped coun-

tercurrent to blood by using linear peristaltic pumps that

controlled both the inflow and outflow rates (Flowgard

6300; Baxter Healthcare Corp, Deerfield, IL). These pumps

allowed a maximum delivery rate of 1,999 mL/min. Hemo-

diafiltration fluid (Baxter Healthcare Corp) was used as

dialysate. Heparin was infused through a prehemofilter port

with initial dosages corresponding to the rate that was

prescribed during the patients conventional hemodialysis

session. The heparin infusion rate was monitored during the

procedure and titrated to achievean activated clottingtime of 120

to 180seconds(AC Tester; Quest Medical Inc,Allen, TX).

Clearance Studies

All patients received a 500-mg intravenous dose of vanco-

mycin administered as a 1-hour infusion during the last hour

of the hemodialysis session on the day before the study. The

time between the end of the infusion and the commencement

of the clearance study (minimum of 12 hours) assured that

the clearance evaluations were performed during the postdis-

tributive phase. Study participants were admitted to the

Clinical Research Center outpatient facility the morning of

the clearance study. All clearance studies were performed

under controlled dialysate, blood, and ultrafiltrate conditions

as described below.

The effect of dialysate inflow rate (QDI) on clearance was

determined by increasing the dialysate flow rate incremen-tally at hourly intervals from 8.3 to 16.7, 25, and 33.3

mL/min while QB and QUF were held constant at 100

mL/min and 0 mL/min, respectively. The effect of QB on

clearance was determined by increasing the QB hourly from

75 to 100, 125, and 150 mL/min while the dialysate flow rate

and QUF were held constant at 33.3 and 0 mL/min, respec-

tively. The SC and CVVH clearance were assessed at nomi-

nal QUFs of 500 and 1,000 mL/hr while maintaining QB and

dialysate flow rates of 100 mL/min and 0 mL/min, respec-

1020 JOY ET AL


3/9

tively. CVVH clearance of vancomycin, urea, and creatinine

at the two QUFs was determined during two 15-minute

periods, after an initial 15 minutes for equilibration. Each

CVVHD clearance study period consisted of an initial 20-

minute equilibration period and two 20-minute clearance

determinations. Blood samples were collected at the mid-point of each dialysate/ultrafiltrate collection period.

Analytic

The measurement of urea and creatinine concentrations in

the plasma and dialysate/ultrafiltrate specimens was per-

formed by using an Ektachem 700 XRC autoanalyzer (East-

man Kodak, Rochester, NY). The concentrationof vancomy-

cin in the plasma and dialysate/ultrafiltrate samples was

determined in duplicate by enzyme-multiplied immunoassay

(EMIT; Syva, Palo Alto, CA). The lower limit of quantifica-

tion of this procedure was 1.0 mg/L. The interassay coeffi-

cient of variation at 10 and 29.0 mg/L in plasma and/or

ultrafiltrate/dialysate was less than 9.0% and 6.2%, respec-

tively. Plasma protein binding of vancomycin was deter-

mined in triplicate from three samples for each subject afterCentrifree filtration (Amicon, Inc, Beverly, MA).

Pharmacokinetic Analysis

The clearance of urea, creatinine, and vancomycin (total

and unbound) was calculated during each CVVHD period as

CL [QDO x CDO] / C Pmid, where CL solute clearance dur-

ing CVVHD, QDO hemofilter outflow rate, CDO concen-

tration of solute in the hemofilter outflow, and C Pmid

concentration of solute in the plasma at midpoint of the

collection period.

The SC of vancomycin was calculated during each CVVH

period as SC CUF/CP, where ultrafiltrate concentration

(CUF) and plasma concentration (CP) were determined from

simultaneously collected specimens. The clearance of urea,

creatinine, and vancomycin (total and unbound) was calcu-lated during the four observation periods of CVVH as

CLCVVH (CUF * QUF)/Cpmid, where CUF concentration of

solute in the ultrafiltrate and QUF ultrafiltrate outflow rate.

Dosing regimen guidelines were derived from the ob-

served CVVH and CVVHD clearance data assuming a

once-a-day dosing strategy (dose [mg/24 hr] Css x Total

Clearance [(CLNR CLR) x 24 hr]. The average desired

serum vancomycin concentration (Css) was 20 mg/L, the

nonrenal clearance (CLNR) was assumed to be 16 mL/min as

reported by Macias et al,11 and residual renal vancomycin

clearances (CLR) associated with creatinine clearances of 0

to 20 mL/min were calculated as CLR 0.693 * creatinine

clearance (CLcr).10

StatisticsThe demographic characteristics of the three groups were

compared by analysis of variance (ANOVA). The clearance

of vancomycin, urea, and creatinine by the three filters dur-

ing CVVH and CVVHD was compared by a mixed-model

repeated-measures ANOVA with filter and flow rate as main

effects and patient as a random effect. Linear regression

analysis was performed to determine the relationship be-

tween dialysate,QB, or QUF rates, and CVVHD and CVVH

clearance of urea, creatinine, and vancomycin, respectively.

Regression lines were compared using t-tests for common

slopes. Results were calculated as mean standard devia-

tion (SD). Computations were performed with version 6.10

of Statistical Analysis Software (SAS Institute, Cary, NC)

and P less than 0.05 was considered statistically significant.

RESULTS

The patients in each of the three filter groups

were similar with regard to age, sex, race, weight,

and pertinent laboratory measurements (Table

1). None of the patients experienced any adverse

events while participating in this study.

CVVH Clearance

Vancomycin was moderately protein bound in

the ESRD patients (free fraction range, 0.49 to

0.83). No significant differences in fraction un-

bound to plasma proteins (fup) were noted be-

tween the three groups of patients: AN69, fup 0.67 0.07; PS, fup 0.65 0.15; PMMA,

fup 0.68 0.12. The SC of vancomycin

approximated the fup for the AN69 (0.70 0.15)

Table 1. Clinical Characteristics of the Study Patients

Patient

No.

Age

(yr)

Weight

(kg ) Ra ce S ex

Hema-

tocrit

(%)

Albumin

(gm/dL)

Total

Bilirubin

(mg/dL)

Group 1AN69 filter

1 44 76 B F 36.4 3.8 0.4

2 46 80 B F 30.5 3.5 0.53 62 70 B M 38.7 4.2 0.8

4 48 76 B M 31.4 3.8 0.3

5 29 104 B F 32.2 4.6 0.6

Mean 45.8 81.2 33.8 4.0 0.5

SD 11.8 13.2 3.5 0.4 0.2

Group IIPMMA filter

1 41 64 B F 31.7 4.1 0.5

2 29 104 B F 31.6 4.5 0.5

3 64 52 W M 27.3 4.1 1.8

4 62 70 B M 38.1 3.9 0.6

5 29 104 B F 31.6 4.5 0.5

Mean 45.0 78.8 32.1 4.2 0.8

SD 17.2 23.9 3.9 0.3 0.6

Group IIIPS filter1 49 63 B F 34.5 3.9 0.6

2 44 76 B F 38.3 3.9 0.9

3 64 52 W M 36.1 3.9 1.4

4 48 76 B M 35.5 4.1 0.4

5 46 80 B F 25.8 3.6 0.6

Mean 50.2 69.4 34.0 3.9 0.8

SD 8.0 11.7 4.8 0.2 0.4

Abbreviations: B, black; W, white; F, female; M, male.

VANCOMYCIN CLEARANCE BY CVVH AND CVVHD 1021


4/9

and PS (0.68 0.19) groups, whereas it was

higher in the PMMA group (0.86 0.16), but

this did not achieve statistical significance (P 0.279). Although the fup of vancomycin was

significantly correlated with the SC (P

0.01), itonly accounted for 46.5% of the variability in the

measured SC.

The clearance of urea and creatinine at a QUFof 602 160 mL/hr for the AN69, PS, and

PMMA filters are shown in Fig 1. The clearance

of these two solutes was significantly increased

at the higher QUF. However, there was no differ-

ence between the three filter groups at either Q UF.

CVVHD Clearance Versus Dialysate Flow

Urea clearance increased linearly with QDI for

all three filters (Fig 2). The regression lines for

urea clearance when plotted against QDI haveslopes of 0.77 (r2 0.951; P 0.0001), 0.80(r2 0.926; P 0.0001), and 0.91 (r2 0.962;P 0.0001) for the AN69, PS, and PMMAfilters, respectively. Creatinine clearance also in-

creased linearly (Table 2).

The clearance of total vancomycin by the

PMMA filter exceeded the values observed with

the PS and AN69 filters at QDI greater than 16.7

mL/min (Table 2). Regression analysis showed

significant linear relationships between vancomy-

cin clearance and QDI for all three filters: AN69

(slope 0.482; r2 0.88; P 0.0001), PS

(slope 0.658; r2 0.887; P 0.0001), andPMMA (slope 0.853; r2 0.966; P 0.0001)(Fig 3). The slope of the PMMA regression line

was significantly greater than the values ob-

served with the AN69 and PS filters (P 0.05).The relationship between clearance of unbound

vancomycin and dialysate flow rate was similar.

Vancomycin total clearance was significantly

Table 2. Vancomycin, Creatinine, and Urea Clearance

(mL/min) in Relation to Dialysate Inflow for AN69,

PS, and PMMA Filters During Constant Blood Flow

Dialysate Inflow Rate (mL/min)

8.3 16.7 25.0* 33.3

AN69 filter

Vancomycin 5.81.5 10.01.5 13.73.5 13.46.7

Urea 7.31.2 12.13.8 18.45.6 26.75.3

Creatinine 6.50.9 10.43.0 15.34.7 22.04.2

PS filter

Vancomycin 5.21.6 11.44.8 16.04.8 22.19.3

Urea 5.91.7 12.84.0 20.66.0 26.98.8

Creatinine 5.31.7 11.63.3 19.25.9 23.37.8

PMMA filter

Vancomycin 7.50.9 14.73.3 22.83.7 27.05.6

Urea 7.00.7 17.55.0 21.55.8 30.03.8

Creatinine 6.20.8 15.84.6 19.86.4 26.43.7

*Vancomycin clearance by PMMA AN69 (P 0.006)

and PS (P 0.037).

Vancomycin clearanceby PMMAAN69 (P 0.0001).

Vancomycin clearance at QDI 8.3 25 and 33.3

mL/min (P 0.010).

Vancomycin clearance at QDI 8.3 25 and 33.3

mL/min; 16.7 33.3 mL/min (P 0.0002).

Vancomycin clearance at QDI 8.3 16.7, 25.0, and

33.3 mL/min; 16.7 25.0 and 33.3 mL/min (P 0.001).

Fig 1. The clearance of urea () and creatinine ()by CVVH at an ultrafiltration rate of 0.6 L/hr did notsignificantly differ within or between filters.

Fig 2. Urea clearance was highly correlated withdialysate inflow rate for the AN69 (; dotted line; r0.992), PS (, dashed line; r 0.999), and PMMAfilters(, solid line; r 0.988) at a constant Q B of 100

mL/min.

1022 JOY ET AL


5/9

correlated with the observed urea clearances for

all three filters (Fig 4). The slope of the relation-

ship with the AN69 and PS filters was similar,

whereas that of the PMMA filter was markedly

greater (P 0.05).

CVVHD Clearance Versus QB

Vancomycin, urea, and creatinine clearances were

also measured while QB was increased from 75 to

150 mL/min (Table 3). The clearance of all three

solutes was essentially constant at all QBs for all

three filters. A significant difference in vancomycinclearance, however, was noted between the three

filters at a QB of 75 and 100 mL/min (Table 3).

Comparison of Membranes

The CVVH clearance of urea and creatinine

was similar for all three membranes, and the SC

of vancomycin was not significantly different.

During CVVHD, the clearances of urea and

creatinine were not significantly different be-

tween filters at any level of QDI. However, at QDIof 25 and 33.3 mL/min, the clearance of total

vancomycin provided by the AN69 membranewas 60% and 50%, whereas the PS filter

clearances were 70% and 82% of the values

observed with the PMMA filter (P 0.03).

DISCUSSION

Several reports of the disposition of vancomy-

cin during CVVH have been published.5,11-14

Many of these were single case reports or a

Fig 4. Relationship between vancomycin clear-ance and urea clearance for (A) the AN69 filter (vanco-mycin clearance 0.650 [Clurea]; r2 0.884), (B) the PSfilter (vancomycin clearance 0.706 [Clurea]; r2 0.920),and(C) thePMMA filter (vancomycin clearance 0.892[Clurea]; r2 0.903).

Fig 3. Vancomycin clearance in relation to dialy-sate inflow rate for the AN69 (; dotted line), PS (;dashed line), and PMMA (, solid line) filters at con-

stant QB of 100 mL/min. Values are mean SD.



6/9

collection of a series of clinical cases. This

approach, unfortunately, does not allow one to

control the critical variables that may affect the

clearance of vancomycin. In this study, we mea-

sured the SC for three filters and the fup of

vancomycin in stable ESRD patients undergoing

controlled CVVH. The clearance of vancomycin

and two reference solutes (urea and creatinine)

determined at two QUFs confirmed the depen-

dence of CLCVVH on QUF and fup. No clinical

significant difference in large or small solute

clearance was attributed to the type of membrane

used for CVVH.

The mean fup of vancomycin in these patients

was similar to previous reports in healthy volun-

teers and infected patients.10, 20 The large degree

of interpatient variability, however, limits the

utility of a mean value as a predictor of an

individual patients fup. The fup of these patientswas significantly correlated with the observed

SC of vancomycin. This confirms the depen-

dence of SC and, thereby, vancomycin CVVH

clearance on the fup.

CVVHD allows one to independently regulate

blood, dialysate, and ultrafiltrate flow.4 Clear-

ance of solute during CVVHD is composed of a

diffusive and a convective or ultrafiltration com-

ponent. Because QUF was less than 10% of QDOduring the CVVHD segment of this study, the

changes in vancomycin clearance predominantly

reflect the effects of alterations in blood and

dialysate flow on diffusion across the membrane.Unfortunately, most of the previously published

studies of vancomycin clearance during CVVHD

did not control QUF to this degree, and thus, their

results cannot be directly compared with our

observations.6,15-19 Furthermore, the only filter

membrane for which vancomycin clearance has

been previously evaluated is the AN69. Thus,

this investigation is the first rigorous investiga-

tion of the determinants of vancomycin clear-

ance by CVVHD.

Increasing QDI from 8.3 to 33.3 mL/min pro-

duced a linear increase in the clearance of the

two reference solutes, urea and creatinine, witheach of the three hemofilters (Table 2 and Fig 2).

The slope of the urea and creatinine clearance to

QDI relationships with the AN69 filter were

slightly lower (0.78 and 0.64, respectively) but

similar to the values previously reported by If-

ediora et al21 (0.92 and 0.87) and Relton et al9

(0.88 and 0.74). Similar congruity in the urea

relationships was evident for the PS membrane.

Ifediora et al21 reported a slope of 0.85 for the

Renal Systems HF-500 filter and 0.91 for the

Fresenius F-8 filter, whereas we observed a slope

of 0.80 with the Fresenius F-40 filter. Our find-

ings concur with the previous investigations,

which indicate the choice of the filter membrane

is not a critical determinant of CVVHD perfor-

mance for control of azotemia.4,21

As QDI increased from 8.3 to 33.3 mL/min,

vancomycin clearance increased by 136% for the

AN69, 325% for the PS, and 260% for the

PMMA filters. These changes were similar in

magnitude to those observed with urea, and re-

gression analysis showed better correlation be-

tween vancomycin and urea clearance than QDIand vancomycin clearance (Fig 4). These data

strongly suggest that vancomycin clearance dur-ing CVVHD is mainly dependent on diffusion.

Vancomycin clearance by the PMMA filter

exceeded its clearance by the AN69 and PS

filters at all QDI. These differences were statisti-

cally significant at QDI of 25.0 and 33.3 mL/min

(Table 2). The observed clearances by the AN69

filter closely approximated previous reported val-

ues with the filter during in vitro19 and in vivo

Table 3. Vancomycin, Creatinine, and Urea Clearance

(mL/min) in Relation to Blood Flow for AN69, PS, and

PMMA Filters During Constant Dialysate Inflow

Blood Flow Rate (mL/min)

75* 100 125 150

AN69 filter

Vancomycin 12.83.0 13.46.7 17.32.9 15.15.3

Urea 20.110.3 26.75.3 25.27.6 25.62.7

Creatinine 18.13.6 22.04.2 20.66.3 21.21.6

PS filter

Vancomycin 17.17.1 22.19.3 19.57.3 21.49.5

Urea 20.45.6 26.98.8 21.67.5 25.55.8

Creatinine 16.53.6 23.37.8 19.54.4 22.05.1

PMMA filter

Vancomycin 26.38.1 27.05.6 23.17.5 25.37.7

Urea 32.616.3 30.03.8 27.06.7 29.07.2

Creatinine 29.415.6 26.43.7 23.66.3 22.05.1

*Vancomycin clearance by PMMAAN69 and PS, P

0.018.

Vancomycin clearance by PMMAAN69, P 0.038.

For all within-filter comparisons, P 0.05.

1024 JOY ET AL


7/9

CVVHD evaluations15-18 (Fig 5). Thus, the diffu-

sive clearance of vancomycin is dependent on

QDI and the type of hemofilter. However, we

cannot differentiate the degree of contribution to

the hemofilter differences that are a function of

the membrane composition per se or the higher

surface area of the PMMA hemofilter.

If diffusive clearance of vancomycin during

CVVHD were flow limited, as it is with intermit-

tent hemodialysis, then one would anticipate that

increasing blood flow could also result in an

increase in vancomycin clearance.22-24 Within-

filter comparisons showed that there was no

significant difference in vancomycin, urea, or

creatinine clearance as QB was increased from 75

to 150 mL/min. Thus, a QB of 75 mL/min is

sufficient to maximize the clearance of small and

middle molecules with each of the filters. Be-

tween-filter comparisons, however, showed that

at low QB, vancomycin clearance was signifi-

cantly greater with the PMMA filter than the

AN69 (QB 75 mL/min; P 0.018 and QB 100 mL/min; P 0.038) and PS filters (QB 75

mL/min; P

0.018). This may be a function ofthe larger surface area of the PMMA filter rela-

tive to the other two filters.

Our data suggest that the clearance of vanco-

mycin in patients with acute renal failure can be

significantly augmented by CVVH and CVVHD

therapy. The total-body clearance of vancomycin

in patients with acute renal failure is composed

of residual renal clearance and a nonrenal compo-

nent. In anuric patients, the nonrenal clearance

has ranged from 9 to 35 mL/min (weighted mean

of 16 mL/min).11,16-18 Furthermore, this compo-

nent of vancomycin disposition has been re-

ported to decline toward values observed inESRD patients (4 mL/min) and in those with a

prolonged course of acute renal failure (ie, 7 to

10 days).11

The total-body clearance of vancomycin could,

thus, be increased by 62% (early in the course of

acute renal failure) to 262% (late) by CVVH

with a QUF of 0.9 L/hr. Because the convective

clearance of vancomycin is linear, the percentage

increment in total-body clearance of vancomycin

would be even greater at higher ultrafiltration

rates. The contributions of CVVHD are even

more dramatic. Depending on the filter or Q DI

used, one could anticipate a maximum increasein the patients vancomycin clearance of 90% to

169% early, or 288% to 540% late in the course

of anuric acute renal failure. Although the contri-

butions of CRRT therapy to total vancomycin

clearance will be lower in those patients with

higher degrees of residual renal function, dosage

regimen adjustment will still be necessary.

Maintenance dosage recommendations, after

the initiation of vancomycin therapy with a load-

ing dose of 15 to 20 mg/kg, for patients receiving

CVVH with any of the three filters evaluated in

this study, are listed in Table 4. Vancomycin

dosage may need to be reduced for those patientswith prolonged acute renal failure because the

assumed CLNR may decline with time. Further-

more, because marked interpatient and interfilter

variability in the SC of vancomycin was ob-

served, individualized pharmacokinetic evalua-

tions should be considered to guide subsequent

Table 4. Vancomycin Dosage Guidelines During

CVVH (mg/24 hr)

Residual

Renal

Function

(CLcr in

mL/min)

Ultrafiltration Flow Rate (mL/min)

2 5 10 15 20 30

0 500 550 650 750 850 1,050

5 600 650 750 850 950 1,150

10 700 750 850 1,000 1,050 1,250

15 800 850 1,000 1,100 1,150 1,350

20 900 1,000 1,100 1,200 1,250 1,500

Abbreviation: CLcr, creatinine clearance.

Fig 5. Vancomycin clearance during CVVHD of pa-tients with acute renal failure correlates well with pre-vious in vitro observations with the AN69 filter () andcurrent findings ().



8/9

dosing decisions. The dosage recommendations

in Table 4 are based on the desire to achieve an

average plasma concentration of 20 mg/L. The

peak and trough concentrations associated with

these regimens will range from 25 and 15 mg/Lto 34 and 7 mg/L, respectively, for patients

receiving the lowest (500 mg) and highest (1,500

mg) projected doses proposed in Table 4. If

lower or higher concentrations are desired, the

dosage per day can be increased or decreased

proportionally. For example, if the desired aver-

age concentration is 15 mg/L, the dose per day

for a patient with a residual renal function of 15

mL/min receiving CVVH at a QUF of 10 mL/min

would be 750 mg (15 mg/L/20 mg/L * 1,000

mg/day).

Projected vancomycin dosage requirements to

maintain plasma concentrations of 20 mg/L for

patients receiving CVVHD are similar to those

for CVVH-managed patients (Table 5). Because

significant differences in vancomycin clearance

between the three filters were observed, dosage

guidelines were developed on the basis of the

filter, QDI, QUF, and the patients residual renal

function. Combinations of convective and diffu-

sive transport may be beneficial in many clinical

settings. If greater convective clearance is de-

sired, the dosage of vancomycin should be in-

creased by about 60 mg daily for each 3-mL/min

increase in QUF. Alternatively, the urea clearanceduring CVVHD could be measured and the dos-

age adjustment of vancomycin could be individu-

alized on the basis of the estimated vancomycin

clearance from the relationships in Fig 4, plus the

patients residual renal and nonrenal clearance as

described previously.

Our data indicate that the removal of vancomy-

cin by CVVH is dependent on the fup of the

patient and the delivered QUF. No filter mem-

brane effect was observed to be statistically or

clinically significant. In contrast, the removal of

vancomycin by CVVHD was primarily depen-

dent on filter membrane and QDI. Vancomycindosage regimens can be initiated on the basis of

the recommendations in Tables 4 and 5. How-

ever, because of the marked interpatient variabil-

ity in fup, volume of distribution, and residual

renal and nonrenal clearance, prospective phar-

macokinetic dosage regimen individualization

with serial monitoring of vancomycin serum

concentrations may be indicated.

ACKNOWLEDGMENT

The authors greatly appreciate the technical assistance of

Cheryl Galloway and the nursing staff of the General Clini-cal Research Unit and the secretarial assistance of Diane

Kenna, and thank Karen Baker, RN, William Dawson, RN,

Colleen Fitzpatrick, RN, Michelle Gazella, RN, Mary Hrinya,

RN, Stephanie Hughes, RN, Elaine Lander, RN, Ruth Le-

hotsky RN, Chris Lion, RN, and Patricia Shortridge, RN for

their invaluable assistance in conducting these studies.

REFERENCES

1. Conger JD: The role of continuous renal replacement

therapy in the future treatment of acute renal failure. Am J

Kidney Dis 28:S108-S113, 1996 (suppl 3)

2. Joy MS, Matzke GR, Armstrong DA, Marx M, Zarow-

itz BJ: A primer on continuous renal replacement therapy for

critically ill patients. Ann Pharmacother 32:362-375, 19983. Ronco C: Continuous renal replacement therapies:

Evolution towards a new era. Semin Dial 9:215-221, 1996

4. Sigler MH: Transport characteristics of the slow thera-

pies: Implications for achieving adequacy of dialysis in

acute renal failure. Adv Renal Replace Ther 4:68-80, 1997

5. Bresolle F, Kinowski JM, de la Coussaye JE, Wynn N,

Eledjam JJ, Galtier M: Clinical pharmacokinetics during

continuous haemofiltration. Clin Pharmacokinet 26:457-

471, 1994

Table 5. Vancomycin Dosage Guidelines During CVVHD (mg/24 hr)

Residual

Renal

Function(CLcr in mL/min)

AN69 QDI PS QDI PMMA QDI

16.7

QUF (mL/min)

33.3

QUF (mL/min)

16.7

QUF (mL/min)

33.3

QUF (mL/min)

16.7

QUF (mL/min)

33.3

QUF (mL/min)

2 5 2 5 2 5 2 5 2 5 2 5

0 800 850 900 950 850 900 1,150 1,200 950 1,000 1,300 1,350

5 900 950 1,000 1,050 950 1,000 1,250 1,300 1,050 1,100 1,400 1,450

10 1,000 1,050 1,100 1,150 1,050 1,100 1,350 1,400 1,150 1,200 1,500 1,550

15 1,100 1,150 1,200 1,250 1,150 1,200 1,450 1,500 1,250 1,300 1,600 1,650

20 1,200 1,250 1,300 1,350 1,250 1,300 1,550 1,600 1,350 1,400 1,700 1,750

1026 JOY ET AL


9/9

6. Kroh UF, Hall TJ, Steinhauber W: Management of

drug dosing in continuous renal replacement therapy. Semin

Dial 9:161-165, 1996

7. Sigler MH, Manns M: Membranes and devices in

continuous renal replacement therapy. Semin Dial 9:98-106,

19968. Kronfol NO, Lau AH, Colon-Rivera J, Libertin CL:

Effect of CAVH membrane types on drug-sieving coeffi-

cients and clearances. ASAIO Trans 32:85-87, 1986

9. Relton S, Greenberg A, Palevsky PM: Dialysate and

blood-flow dependence of diffusive solute clearance during

CVVHD.ASAIO J 38:M691-M696, 1992

10. Matzke GR: Vancomycin, in Evans WE, Schentag JJ,

Jusko WJ (eds): Applied Pharmacokinetics: Principles of

Therapeutic Drug Monitoring (ed 3). Spokane, WA, Applied

Therapeutics Inc, 1992, pp 15/1-15/31

11. Macias WL, Mueller BA, Scarim SK: Vancomycin

pharmacokinetics in acute renal failure: Preservation of

nonrenal clearance. Clin Pharmacol Ther 50:688-694, 1991

12. Dupuis RE, Matzke GR, Maddux FW, ONeil MG:

Vancomycin disposition during continuous arteriovenoushemofiltration. Clin Pharm 8:371-374, 1989

13. Lau AH, Kronfol NO, John E: Increased vancomycin

elimination with continuous hemofiltration. ASAIO Trans

33:772-774, 1987

14. Thomson AH, Grant AC, Rodger SC, Hughes RL:

Gentamicin and vancomycin removal by continuous venove-

nous hemofiltration. Ann Pharmacother 25:127-129, 1991

15. Bellomo R, Ernest D, Parkin G, Boyce N: Clearance

of vancomycin during continuous arteriovenous hemodiafil-

tration. Crit Care Med 18:181-183, 1990

16. Reetze-Bonorden P, Bohler J, Kohler C, Schollmeyer

P, Keller E: Elimination of vancomycin in patients on

continuous arteriovenous hemodialysis. Contrib Nephrol.

93:135-139, 1991

17. Davies SP, Azadian BS,Kox WJ,Brown EA: Pharma-

cokinetics of ciprofloxacin and vancomycin in patients with

acute renal failure treated by continuous haemodialysis.

Nephrol Dial Transplant 7:848-854, 199218. Santre C, Leroy O, Simon M, Georges H, Guery B,

Beuscart C, Beaucaire G: Pharmacokinetics of vancomycin

during continuous hemodialfiltration. Intensive Care Med

19:347-350, 1993

19. Kronfol N, Adams L. Powell S, Lau A: Clearance of

drugs during continuous arteriovenous hemodialysis

(CAVHD) correlates with dialysate flow rate (Qdi). Clin Res

37:9A, 1989 (abstr)

20. Li L, Miles MV, Lakkis H, Zaritsdy AL: Vancomycin-

binding characteristics in patients with serious infections.

Pharmacotherapy 16:1024-1029, 1996

21. Ifediora OC, Teehan BP, Sigler MH: Solute clearance

in continuous venovenous hemodialysis: A comparison of

cuprophane, polyacrylonitrile, and polysulfone membranes.

ASAIO J 38:M697-M701, 1992

22. Touchette MA, Patel RV, Anandan JV, Dumler F,

Zarowitz BJ: Vancomycin removal by high-flux poly-

sulfone hemodialysis membranes in critically ill patients

with end-stage renal disease. Am J Kidney Dis 26:469-474,

1995

23. Welage LS, Mason NA, Hoffman EJ, Odeh RM,

Dombrowski J, Patel A, Swartz RD: Influence of cellulose

triacetate hemodialyzers on vancomycin pharmacokinetics.

J Am Soc Nephrol 6:1284-1290, 1995

24. Pollard TA, Lampasona V, Akkerman S, Tom K,

Hooks MA, Mullins RE, Maroni BJ: Vancomycin redistribu-

tion: Dosing recommendations following high-flux hemodi-

alysis. Kidney Int 45:232-237, 1994


Documents

1998 Vancomycin HD