Pharmacokinetics of intravenously administered ciprofloxacin in ...

1 downloads 0 Views 797KB Size Report
Patients with Various Degrees of Renal Function. GEORGE L. DRUSANO,l 2* MATTHEW WEIR,3 ALAN FORREST,1'2 KAREN PLAISANCE,2 THOMAS EMM,2.
Vol. 31, No. 6

ANTIMICROBIAL AGENTS AND CHEMOTHERAPY, June 1987, p. 860-864

0066-4804/87/060860-05$02.00/0 Copyright X) 1987, American Society for Microbiology

Pharmacokinetics of Intravenously Administered Ciprofloxacin in Patients with Various Degrees of Renal Function GEORGE L. DRUSANO,l 2* MATTHEW WEIR,3 ALAN FORREST,1'2 KAREN PLAISANCE,2 THOMAS EMM,2 AND HAROLD C. STANDIFORD,"2'4 Divisions of Infectious Diseases1 and Nephrology,3 Department of Medicine, University of Maryland School of Medicine, and University of Maryland School of Pharmacy,2 Baltimore, Maryland 21201, and Loch Raven Veterans Administration Medical Center, Baltimore, Maryland 212184

Received 18 November 1986/Accepted 3 March 1987 We examined the pharmacokinetic behavior of 200 mg of ciprofloxacin administered intravenously to 32 volunteers whose renal function as measured by creatinine clearance ranged from 0 to 8.99 liters/h per 1.73 m2. Serum clearances (mean + standard deviation) were 26.8 + 5.7 and 15.4 ± 4.3 liters/h per 1.73 i2 in normal and anephric volunteers, respectively. The half-life (mean ± standard deviation) increased from 4.3 ± 0.8 h in normal volunteers to 8.6 ± 3.3 h in anephric volunteers. There was good correlation between normalized creatinine clearance and both normalized serum and renal clearance. The regression equation for serum clearance (CLs) versus creatinine clearance (CLCR) was CLs = 1.97 x CLCR + 13.23, where r = 0.697; for renal clearance versus creatinine clearance, the equation was CLR = 2.26 x CLCR, where r = 0.845. On the basis of these data, we recommend a maximum 50% reduction in dose when ciprofloxacin is instituted at a renal function of 1.2 to 1.8 liters/h per 1.73 M2 (20 to 30 ml/min per 1.73 m2). Because of the observed variation in ciprofloxacin half-life in our anephric volunteers, we also recommend that a schedule of administration every 12 h be maintained, even for patients without urine output.

constant infusion. For those in groups 2 to 4, 200 mg of ciprofloxacin was administered intravenously as a 30-min constant infusion. Patients drank 180 ml of water before the infusion was begun. Normal patients and patients in group 2 were hydrated throughout the period of study. Patients in groups 3 and 4 received fluid in volumes consistent with their clinical status. The patients fasted for at least 8 h prior to drug administration and for 3 h thereafter. Specimen collection and handling. Blood was obtained for ciprofloxacin assay prior to drug administration. Samples were also obtained (in all instances from the arm contralateral to drug administration) at the end of infusion for all groups. For group 1 subjects, samples were also taken at 0.25, 0.33, 0.42, 0.67, 0.92, 1.17, 1.67, 2.17, 3.17, 4.17, 6.17, 8.17, 12.17, and 24.17 h after the infusion was begun. For patients in groups 2 through 4, samples were also obtained at 0.75, 1, 1.25, 1.5, 2, 2.5, 3.5, 4.5, 6.5, 8.5, 12.5, 16.5, 24.5, 36.5, and 48.5 h after the infusion was begun. Blood was allowed to clot, and serum was separated and promptly frozen at -20°C or below until the time of assay. Urine was obtained for ciprofloxacin assay prior to drug administration and quantitatively at 0 to 2, 2 to 4, 4 to 8, 8 to 12, 12 to 24, and 24 to 48 h after the initiation of drug administration for all volunteers. Those in groups 2 through 4 also had urine collectd at 48 to 72 h. Urine volumes were measured, and samples were frozen at -20°C or below until the time of assay. Drug assay. Ciprofloxacin in serum and urine was measured by high-pressure liquid chromatography by a variant of the assay of Gau et al. (6). The serum assay involved the use of protein precipitation, reverse-phase high-pressure liquid chromatography and fluorometric detection. The urine assay involved the use of sample dilution followed by direct injection. Quinine was used as the internal standard. A Waters Associates C-18 micro-Bondapak reverse-phase column was used. The mobile phase consisted of 27% methanol, 0.8% tetrahydrofuran, and 0.67 M phosphate buffer (pH

Ciprofloxacin is a new quinoline carboxylic acid with a broad range of activity (5). Since this antibiotic may be used intravenously, it may have a place in the therapy of serious nosocomial infections. We previously studied the pharmacokinetic behavior of 200 mng of ciprofloxacin administered intravenously to normal volunteers (3). We found that renal clearance accounted for 67 ± 11% (mean ± standard deviation) of the total serum clearance. Consequently, one would expect that seriously ill patients, who may have renal dysfunction, would have an impaired clearance of ciprofloxacin compared with that of normal volunteers. We therefore undertook a study of the pharmacokinetic behavior of 200 mg of ciprofloxacin administered intravenously to patients with various degrees of renal functional impairment. MATERIALS AND METHODS Selection of patients. Thirty-two patients with various degrees of renal function were studied. All volunteers selected for study had given written informed consent according to institutional guidelines. They were divided into four groups on the basis of renal function. Patients in group 1 had measured creatinine clearances greater than 6 liters/h per 1.73 M2. Group 1 volunteers were a subset (8 of 12) of previously reported normal volunteers receiving 200 mg of ciprofloxacin intravenously (4). Group 2 subjects had creatinine clearances greater than or equal to 3.6 and less than 6.0 liters/h per 1.73 M2. Group 3 volunteers had measured creatinine clearances greater than or equal to 0.6 and less than 3.6 liters/h per 1.73 M2. Finally, those in group 4 had measured creatinine clearances less than 0.6 liters/h per 1.73 mi2. Hepatic function indices were known to be normal. Drug administration. For group 1 patients, 200 mg of ciprofloxacin was administered intravenously as a 10-min * Corresponding author. 860

VOL. 31, 1987

CIPROFLOXACIN PHARMACOKINETICS

TABLE 1. Demographic characteristics of volunteers with differing renal function receiving 200 mg of ciprofloxacin intravenously Group

Wt

CLCR (liter/h

Disease

88.2 74.5 81.9 60.5 74.1 89.4 68.5 81.6

6.2 7.5 9.0 6.9 6.1 6.1 7.9 7.0

NOR NOR NOR NOR NOR NOR NOR NOR

68.2 77.5 75.7 108.9 67.3

5.2 5.2 4.1 4.9 4.6

HTN CGN CGN CGN HTN

Male Male Male Female Male Male Male Male Male Male Male

88.2 80.4 68.2 58.4 51.3 88.9 78.6 80.0 97.7 70.9 72.3

3.4 1.8 0.8 1.2 3.2 3.3 1.2 1.5 3.4 3.2 3.4

CGN CGN HTN CGN HTN HTN CGN HTN HTN CPN HTN

Male Male Male Male Male Male Male Male

69.8 66.8 67.2 85.5 71.1 68.3 59.6 62.9

0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

HTN HTN HTN HTN CGN HTN HTN HTN

Age (yr)

Sex

(kg)

1 2 3 4 5 6 7 8

28 28 28 28 30 26 22 27

Male Male Male Male Male Male Male Male

1 2 3 4 5

42 27 31 32 48

Male Male Male Male Male

1 2 3 4 5 6 7 8 9 10 11

29 52 60 24 26 44 62 49 43 42 49

1 2 3 4 5 6 7 8

36 33 55 47 52 45

per 1.73

m2)a

statusb

1

2

3

4

52 44

CLCR, Creatinine clearance. NOR, Normal; HTN, hypertensive nephropathy; CGN, chronic glomerulonephritis; CPN, chronic pyelonephritis. a

b

3.0). A flow rate of 1.3 ml/min was used. The column was heated to 50°C. Ciprofloxacin peaks in urine chromatograms were checked for purity (to rule out assay interference from metabolites) by full-spectrum UV scanning with an HP1040 Diode Array high-pressure liquid chromatography detector (Hewlett-Packard Co.). The standard curves were linear between 0.01 and 1 ,ug/ml for serum and between 0.025 and 1.5 ,ug/ml for urine. Dilutions at these ranges were prepared by using blank serum for the serum assay and phosphate buffer for the urine assay. The sensitivity of the assays was at least 0.01 ,ug/ml. Percent coefficients of variation between days for serum for serum were 5.5% at 0.091 ,ug/ml and 4.3% at 0.859 ,g/ml. For the urine assay, percent coefficients of variation were 4.2% at 0.046 jig/ml and 5.3% at 0.455 ,ug/ml. Pharmacokinetic methodology. An open, linear two- or three-compartment model with zero-order input was fit to the data by an iterative, nonlinear, weighted least-squares regression technique (2). Weighting was by the inverse variance of the assay. The Nelder-Mead simplex search

861

algorithm was used. The microscopic parameters were identified. Hybrid pharmacokinetic parameters were calculated by standard methods (7). Choices between models (two

versus three compartments) were made by an F-ratio test.

Model-independent serum clearance, renal clearance, and volume of distribution in the post-distributional phase (Varea) were determined as follows: Serum clearance = dose/ AUCo_o; Renal clearance = UCiP(UHt2)IAUC(t1-t2), where UCiP(tOt2) is the amount of ciprofloxacin recovered in the urine unchanged over the collection interval from t, to t2. and Varea = dose/(AUC>. x ,B), where 1 is the hydrid rate constant for the terminal portion of the curve. The area under the concentration-time curve (AUC) was determined by the linear trapezoidal rule. Extrapolation to infinity was made by adding C,BI3, where C,, is the last observed concentration in serum to the area under the curve determined by the trapezoidal rule. The 1 for the model-independent analysis was determined by nonlinear, weighted least-squares regression of the terminal, log-linear portion of the concentration-time curve for serum clearance. Linear regression relationships were developed with normalized serum clearance, normalized renal clearance, normalized nonrenal clearance, and normalized creatinine clearance. In each instance, normalized creatinine clearance served as the independent variable. Because anephric patients have no renal clearance, the renal clearance-creatinine clearance regression was constrained to a zero intercept (i.e., the regression line was forced through zero). RESULTS Demographic information for the 32 patients studied is summarized in Table 1. The mean concentration-time profiles for serum clearance for the four groups of patients are displayed in Fig. 1. The concentrations in serum (in micrograms per milliliter) at the end of infusion for groups 1 through 4 (mean ± standard deviation) were 6.30 ± 1.77, 4.14 ± 1.05, 5.44 ± 0.82, and 5.39 ± 1.59, respectively. At 1 h after termination of infusion, concentrations in the four groups were 0.874 ± 0.084, 0.948 ± 0.345, 1.379 ± 0.304, and 1.473 ± 0.321. At 12 h after termination of infusion, concentrations in the four groups were 0.105 ± 0.026, 0.128 ± 0.064, 0.268 ± 0.110, and 0.367 ± 0.060. Model-independent pharmacokinetic parameters for the four groups are displayed in Table 2. As we have seen in previous studies, the apparent volume of distribution is large, ranging from 2.38 ± 0.62 liters/kg (group 3) to 3.19 ± 1.26 liters/kg (group 2). Serum clearance in the four groups falls off, as one would expect, with declining renal function, averaging 26.8 ± 5.7 liters/h per 1.73 m2 in group 1 and declining to 15.4 ± 4.3 liters/h per 1.73 m2 in group 4. The Half-life also increases with increasing renal dysfunction, from 4.27 ± 0.84 h in normal volunteers to 8.55 ± 3.27 h in anephric patients. Of note is the large range of observed values for terminal elimination half-life in anephric patients. The regression relationship between normalized serum clearance and normalized creatinine clearance is in Fig. 2. As can be seen, there is a rather good correlation between serum clearance and creatinine clearance (r = 0.697, P < 0.001). This is particularly impressive when one considers that only 50 to 70% of the drug is renally cleared, even in normal individuals, and that much of the renal clearance must be by active secretion (3). The regression relationship between normalized renal clearance and creatinine clearance is displayed in Fig. 3. As

ANTIMICROB. AGENTS CHEMOTHER.

DRUSANO ET AL.

862

ug/SI

Us/ml

.0-t

e.o

7.0+

7.0 S S

E

E

R U

R U

6.0

C

C 0

5.0-

N C E N

N C E 40

N T

6.0

5.0O 4.0'

R 3.0-

A 3.0

T I 0 20'

T

I N 20 0

N

1.0.

1.0

oIl I

i

2

!

!

3

4

5

7

I

9

W0

is

i

u

2

3

4

6

7

a

9

10t

1

12

TIE (hours after beginning infusion)

TINE (hours after beginning infusion) B

S S

E

RI U

N

TIME

0hors efter beinning infusion) 0

C

FIG. 1. Mean concentrations in serum (± standard deviation) after 200-mg intravenous infusions of ciprofloxacin. (A) Group 1 volunteers, 10-min infusion, n = 8; (B) group 2 volunteers, 30-min infusion, n = 5; (C) group 3 volunteers, 30-min infusion, n = 11; (D) group 4 volunteers, 30-min infusion, n = 8.

can be seen, the line has been forced through zero. The slope of the line is very similar to that seen for the serum clearance relationship. An even stronger correlation is noted between renal clearance and creatinine clearance (r = 0.845, P < 0.001). As the slope of this line attests, the renal clearance is approximately twice the creatinine clearance and is probably explained by substantial active tubular secretion. Much of the unexplained variance in these regressions is probably due to differences in active secretion into the urine.

A less strong correlation between 1B and normalized creatinine clearance was noted. The correlation coefficient for this relationship was r = 0.658 (P < 0.001). The greater inherent variability of the elimination rate constant is reflected in the scatter observed in the half-life (Fig. 4). The hyperbola in this figure is derived from the 1-creatinine clearance regression relationship. No correlation was detected between normalized nonrenal clearance and creatinine clearance. The best-fit line had a

TABLE 2. Model independent pharmacokinetic parameters for 200 mg of ciprofloxacin administered to patients with various degrees of renal dysfunction' Group

2.49 3.19 2.38 2.73

1 2 3 4 a

Varea (liters/kg) (liters/kg)

± 0.46 ± 1.26

± 0.62 ± 0.92

CLs CL per 1.73Mi2)

(liters/h

26.8 26.3 15.0 15.4

± 5.7

10.3 ± 3.8 ± 4.3 ±

CLR

(liters/h per

in2) ~~~~~~~~1.73

t1/2 (h)

% CLR/CLs

16.4 ± 3.5 11.9 ± 5.3 4.5 ± 2.6 0.0

4.27 ± 0.84 6.12 ± 1.61 7.70 ± 1.22 8.55 ± 3.27

61.6 ± 8.7 44.9 ± 11.9 29.2 ± 13.2 0.0

Varea, Volume of distribution in the post distributional phase; CLs, serum clearance; CLR, renal clearance; t112, terminal half-life. Values are mean + standard

deviation.

VOL. 31, 1987 L/hr/1.73 sq. U. 40 35 1

30-

25 e 20 0

** M

I

CIPROFLOXACIN PHARMACOKINETICS Iq.

863

*.

* *

*

ft

ft

i

ft

la

w

s5

o

S cL

-

1.97 *

#

0

1

CrcL

+ 133.23

r - 0.697

2

3

i

5

6

i

6

9

10

NORMALIZED CREATININE CLEARANCE L/hr/1.73 sq. U.

FIG. 2. Regression relationship between normalized serum clearance and normalized creatinine clearance (n = 32) (P < 0.001).

correlation coefficient of only 0.201. The slope of this line (-0.32) was not significantly different from zero, and the intercept (13.97) was very similar to the intercept obtained in the serum clearance-creatinine clearance regression relationship (13.23). DISCUSSION Ciprofloxacin is a new quinoline carboxylic acid whose clearance pathways include renal clearance by glomerular filtration and active tubular secretion, as well as significant metabolic clearance with the production of at least four metabolites (a metabolite in which the piperazine ring on position 7 is opened with the loss of an ethyl function, a sulfated metabolite, an oxo-compound, and a formyl metabolite [A. Heyd, personal communication]). Hospitalized and elderly patients frequently have impaired renal function. Because of multiple clearance pathways for ciprofloxacin, the effect of declining renal function on overall serum clearance is difficult to predict. Because both drug efficacy and toxicity may be related to concentrations in serum, an accurate idea of the impact of altered renal function on drug clearance is important. Because ciprofloxacin possesses multiple pathways of clearance, anephric patients still had a total serum clearance averaging 15.4 liters/h per 1.73 m2 and a lowest observed value of 11 liters/h per 1.73 m2 in the eight patients we studied. This is in contrast to the mean value of 26.8 liters/h per 1.73 m2, with a minimum observed value of 21.4 liters/h per 1.73 m2 in the normal volunteers. Consequently, it is apparent that major dose alteration is unnecessary for ciprofloxacin and should be limited to a maximum 50% dose reduction, which should be initiated at creatinine clearances between 1.2 and 1.8 liters/h per 1.73 m2 (20 to 30 ml/min per 1.73 mi2). Table 3 gives the mean concentration in serum at steady state for the proposed dose and schedule alterations for different categories of patients with declining renal function. The average steady-state concentration in serum (Css) was calculated from the serum clearance (CLs) predicted from our regression equation according to the equation Cs3 = dose/(CLs X T), where T is the dosing interval. It is apparent that only minimal accumulation (less than twofold) will occur under the proposed alteration of dose and schedule. Furthermore, therapeutic efficacy can be expected to be

RCL

- 2.26 *

CrCL

0.845

r -

NORMALIZED CREATININE CLEARNC L/W/1.73 sq. *.

FIG. 3. Regression relationship between normalized renal clearance and normalized creatinine clearance (n = 24). Anephric patients were not included in the analysis. The regression line was constrained to pass through zero (P < 0.001).

maintained as the minimum average concentration in serum for the patients receiving half the normal daily dose (i.e., those with creatinine clearances less than 1.2 liters/h per 1.73 mi2) will be similar to the mean concentration in serum seen for normals receiving the full dose. It is also clear that this dose adjustment should be done by reducing the dose and maintaining the normal schedule (administration every 12 h). In our group of eight anephric patients, although there was relatively little variation in serum clearance, there was a very wide variation in observed terminal elimination half-life, with a minimum of 3.9 h and a maximum of 13.5 h. Clearly, if one is dealing with a patient for whom the half-life is 10 to 12 h, administration of a normal dose (200 mg) every 24 h may be appropriate. However, if the half-life is 4 to 6 h, administration of a normal dose every 24 h would result in long periods at potentially subtherapeutic concentrations. Consequently, because of the unpredictability of half-life even for anephric patients, the safety of the patients would best be served by altering the dose size and maintaining the dosing interval. The effect of renal dysfunction on oral ciprofloxacin 15T hr

10

U.I II XFso -i

I

1

5

L

tf

*

t~~~~~~f

i

*ft

4

5

6

7

i

9

to

NORMALIZED CREATININE CLEARANCE L/hr/1.73 sq. *.

FIG. 4. Relationship between terminal half-life and normalized creatinine clearance. The solid curve is a hyperbola, developed from the P-creatinine clearance regression relationship.

864

ANTIMICROB. AGENTS CHEMOTHER.

DRUSANO ET AL.

TABLE 3. Mean concentration of ciprofloxacin in serum of patients with various degrees of renal dysfunction Mean concn Renal function CLs in serum Dose (mg)c (liters/h per (CLCRa; liters/h per 1.73 m2)

1.73 m2)b

7.2 (>6.0) 4.8 (>3.6, s6.0) 2.4 (>0.6,