253
DIAGN MICROBIOLINFECT DIS 1992;15:253-258
Diffusion of Teicoplanin and Vancomycin in Agar Luigi A. Cavenaghi, Elia Biganzoli, Antonio Danese, and Francesco Parenti
Teicoplanin, although more active than vancomycin [by minimum inhibitory concentration (MIC)], produces smaller inhibition zones in sensitivity testing with 30-tzg disks. Our data support the hypothesis that this is due to lower diffusion of teicoplanin in agar media. After 6 hr of incubation, -70% of vancomycin, but only 20% of teicoplanin entered the agar from a paper disk charged with 30 p,g of antibiotic. This is due to a difference between the diffusion coefficients: 0.47 mm2/hr for teicoplanin and 0.72 mm2/hr for vancomycin. With the
methodology used in this work, it is possible to calculate the range of concentrations of the antibiotic occurring at times likely to include the critical time--the time when the inhibition zone is formed---of most strains at any given distance from the reservoir. One can thus estimate the breakpoint diameter for a given MIC breakpoint; for example, an MIC breakpoint of -15-mm breakpoint diameter for vancomycin (30-1~g disk) and a >-13-mm breakpoint diameter for teicoplanin (30-tzg disk).
INTRODUCTION
tity of teicoplanin to be charged to the paper disk.
Teicoplanin is a novel glycopeptide of the dalbaheptide type similar to vancomycin, from which it differs in net charge (negative while vancomycin's is positive), lipophilicity, and molecular size (Parenti, 1988). These factors are likely to affect the diffusibility in agar and, consequently, the size of the inhibition zone. Barry et al. (1986 and 1987) have f o u n d that disks charged with 30 ~g teicoplanin s h o w e d smaller inhibition zones t h a n 30-~g teicoplanin s h o w e d smaller inhibition zones t h a n 30-~g disks of vancomycin, even t h o u g h the strains were more sensitive to teicoplanin; this suggested a lower diffusibility of teicoplanin. The present study compares the diffusibility of teicoplanin and vancomycin in an agar m e d i u m with the aim of suggesting the appropriate quanFrom the Marion Merrell Dow Research Institute--Lepetit Center, Gerenzano (VA), Italy. Address reprint requests to Dr. L.A. Cavenaghi, Marion Merrell Dow Research Institute--Lepetit Center, via R. Lepetit 34, 1-21040 Gerenzano (VA), Italy. Received 28 March 1991; revised and accepted 1 July 1991. © 1992 Elsevier Science Publishing Co., Inc. 655 Avenue of the Americas, New York, NY 10010 0732-8893/92/$5.00
MATERIALS A N D M E T H O D S Media All studies were carried out in M u e l l e r - H i n t o n agar (Difco Laboratories, Detroit, MI); pH after autoclaving 7.4.
Antimicrobial Agents We used a commercial vial of Vancocin HC1 and a Lepetit analytic standard batch of teicoplanin.
Sensidisk Preparation The antibiotics were dissolved in water, then diluted with reagent grade methanol to appropriate concentrations. Solution (0.010 ml) was applied to paper disks (Schleicher and Schull, paper type 2668/2, 6m m diameter).
Diffusibility of the Antibiotics The melted m e d i u m , cooled to -50°C, was poured into o p e n - e n d e d polycarbonate cylinders (7-mm in-
254
ternal diameter), sealed at one end with Saran Wrap and allowed to solidify in the upright position. Sensidisks of the appropriate potency were applied to the tops of the cylinders, which were turned on their sides and incubated overnight at 35°C. After incubation the sensidisks were discarded, and the agar cylinder was extracted from the polycarbonate tube by applying pressure with a piston. As the cylinder came out of the tube, slices were cut with a sharp razor blade. Each slice was weighed, transferred to a small beaker into which 1 ml of pH 7.4 phosphate buffer (0.1 M) was added, and homogenized with an Ultra Turrax homogenizer. An aliquot was then assayed for antibiotic concentration with a microbiologic method using Bacillus subtilis ATCC 6633 as test organism (Cavenaghi et al., 1986). The assay was based on a parallel line design with detection limits of 0.02 and 0.35 mg/L for teicoplanin and vancomycin, respectively, and confidence limits of +-10%. The diffusibility was assessed after incubation times of 6 and 18 hr; vancomycin was tested at 30 txg, whereas teicoplanin was tested at 30, 60, and 90 lag.
M i n i m u m Inhibitory Concentration (MIC) A Staphylococcus aureus strain (ATCC 10832) was chosen, and its MIC was determined in duplicate according to the NCCLS guidelines (1988). For sensitivity testing, we used Petri dishes of 93-mm internal diameter with 25 ml of medium containing scalar dilutions of the antibiotics and 2 ill of a 1 0 7 colony forming units (CFU)/ml suspension applied to the surface. The MIC thus determined was assumed to be equal to the "critical concentration," that is, the minimal concentration inhibiting growth under the conditions of radial diffusion (Humphrey and Lightbown, 1952).
L.A. Cavenaghi et al.
(/~g/rnt) 4OO
300 0
o
rU
200 O C 0 0
*4
1O 0
o
+
-.,. ÷
0
'
0
' ....
1
2
3
z~ 5
distance
6 7 8 9 10 11 12 12 (ram) from reservoir
FIGURE 1 Diffusion profiles in agar tubes at 18 hr of teicoplanin (+, 30 txg; A, 60 txg; and O, 90 ixg) and vancomycin (A, 30 Ixg).
RESULTS The average concentrations of antibiotic in the sections of the agar cylinders after diffusion times of 18 and 6 hr are reported in Figures 1 and 2, respectively; also shown are the profiles obtained by plotting concentration versus distance from the reservoir. The total amount of antibiotic that diffused into the agar was calculated as the sum of the amounts (that is, concentration x volume of section) found in each section. Not all of the antibiotic present in the paper disk diffused into the agar: -72%-74% for vancomycin and only 20%-28.5% for teicoplanin in 6 hr. Prolonging the incubation time to 18 hr increased the amount of teicoplanin but not that of vancomycin, which diffused into the agar. These data clearly
(/~g/ml) 4OO
Critical Time The critical time (time of incubation after which no inhibition zone can be formed on a inoculated plate) was determined in duplicate according to Humphrey and Lightbown (1952). We used Petri dishes of 93-mm internal diameter with 25 ml of medium inoculated with 0.2 ml of a 1 0 7 CFU/ml suspension evenly distributed on the surface. The sensidisks prepared as described with 7.5, 15, 30, 60, 120, and 240 lag of the antibiotics were applied immediately after inoculation (0 hr) or after 100 min, 3, and 5 hr of preincubation at 35°C. The plates were then inverted and incubated for 18 hr. The inhibition zone diameters were measured, and the intercepts of the regression lines determined from plots of the square of the diameters versus preincubation time.
c 0
300
C
200
d) o c o o
A"..
•
1O0 it..
0 0
1
2
distance
3
4 5 6 7 8 (mm) from reservoir
FIGURE 2 Diffusion profiles in agar tube at 6 hr of teicoplanin (+, 30 p,g; A, 60 ~,g; and O, 90 Ixg) and vancomycin (A, 30 ~xg).
Diffusion of Teicoplanin and Vancomycin
255
indicate a lower agar diffusibility of teicoplanin in comparison with vancomycin. To calculate the diffusion coefficients, linearity of the gradient within each slice was assumed: that is, the concentration measured occurred at the midpoint of the slice. In each experiment, the slice closest to the reservoir was not included in the analysis as, owing to the steepness of the gradient near the reservoir, the concentration at midpoint would be much lower than the average concentration of the slice. In a linear diffusion gradient, the concentration can be described (Cooper, 1963; Finn, 1959) by the equation 1 C = Co .e 4D/;y2
(2)
If Co is assumed to remain constant, then one can plot the logarithm of C against the square of the distance, draw the best-fit regression line, and calculate the diffusion coefficient from D =
1
4kT
(3)
where k is the slope and T the time. Figure 3 shows the diffusion gradients at 6 hr for teicoplanin (30, 60, and 90 I~g) and vancomycin (30 t~g). The points are experimental values and the lines are the calculated best fit. The diffusion coefficients calculated from the curves of Figure 3 were 0.49, 0.46, and 0.47 mm2/hr (average 0.47) for 30, 60, and 90 t~g teicoplanin, respectively, and 0.69 for vancomycin. Diffusion coefficients were also calculated at 18 hr (curves not reported) and were found to be 0.40, 0.48, 0.53 mm2/hr (average 0.47) for teicoplanin (30, 60, and 90 t~g, respectively) and 0.74 for vancomycin. It is possible to calculate the expected diameter of the inhibition zone formed by an antibiotic diffusing radially using the equation
r2 = 4DTo [ I n ( M )
o
5 c 0
.~ .b-,
4
c o
3
c
0
o
2
c
1 0 0
(1)
where y is the distance from the reservoir in millimeters, C is the concentration in micrograms per milliliter at distance y, Co is the concentration in micrograms per milliliter in the reservoir, D is the diffusion coefficient in square millimeters per hour, and T is the elapsed time in hours. Equation 1 can also be expressed as Co _ 1 y2 In C 4DT
6
-ln(C~)-ln(4~DTo)]
10
20
squared
30
40
distance
50
60
( m m 2)
FIGURE 3 Diffusion gradient at 6 hr for teicoplanin (+, 30 lag; A, 60 t~g; and O, 90 lag) and vancomycin (A, 30 I~g). D is the diffusion coefficient (mm2/hr), M is the amount charged to the reservoir (t~g), Cc is the critical concentration (t~g/mm3), To is the critical time (hr), and H is the thickness of the agar (mm). In our experiments the thickness of the agar was 3.68 mm; the MICs were 0.5 mg/L (teicoplanin) and 1 mg/L (vancomycin) for the S. aureus strain used. The determination of critical times for teicoplanin and vancomycin are reported in Figure 4. The estimates were 5.9 hr for teicoplanin and 6.1 hr for vancomycin. In Table 1, the diameters of the inhibition zones calculated using equation 4 (assuming a To of 6 hr for both antibiotics, and a Cc equal to the MIC of each antibiotic for the strain of S. aureus used) are compared with those determined experimentally. In Table 2, we report our calculations of antibiotic concentrations at distances from the reservoir corresponding to zone diameters of 12, 13, 14, and 15 mm; concentrations were calculated for 30-1~g vancomycin disks and 30-, 60-, 75-, and 90-t~g teicoplanin disks at diffusion times of 4, 6, and 8 hr. One can see that, for a diffusion time of 6 hr (which is the time when the inhibition zone is formed) teicoplanin 75 gives a concentration profile similar to vancomycin 30. The concentrations would be lower for teicoplanin 75 than for vancomycin 30 at diameters ->15 ram. The concentrations generated by teicoplanin 75 are smaller than those of vancomycin 30 at a diffusion time of 4 hr. DISCUSSION
(4) where r is the radius of the inhibition zone (mm),
Barry and coworkers (1986 and 1987) reported that teicoplanin, although more active than vancomycin
256
L.A. Cavenaghi et al.
(mm z) 360 300
~
240
f~
z3
180 120 60 0
a
0
1
2
3
4
5
6
7
(mm~ 360 300 240 '13 "o
\ \
180 120
860 0 b
0
1
2
3
time
4
,.5
6
7
(h)
FIGURE 4 Determination of the critical times for (a) teicoplanin and (b) vancomycin (O, 7.5 ~g; A, 15 ~g, V, 30 ~g;, [] 60 bLg; 0, 120 ~g; and B, 240 p,g). (by MIC), produced smaller inhibition zones; they suggested that this was due to lower diffusion of teicoplanin in agar media. Our data support this hypothesis. After 6 hr of incubation (the time when the inhibition zone is formed), -70% of vancomycin,
TABLE 1
but only 20% of teicoplanin, entered the agar from a paper disk charged with 30 btg of antibiotic. This is due to a difference between the diffusion coefficients: 0.47 mma/hr for teicoplanin and 0.72 mm2/hr for vancomycin. The equation used to calculate the diffusion coefficient is rigorously valid when the concentration of antibiotic in the reservoir is constant throughout the diffusion period. This was not the case under the experimental conditions utilized in this work, as the antibiotic content of the disk was not replenished during incubation; this is unlikely, however, to affect significantly the calculated values of the diffusion coefficients, because the concentration of the antibiotic in the disk (which has a very small volume) remains much higher than the concentration in the first section of the agar. In fact, the diffusion coefficients measured with different charges of teicoplanin and at different diffusion times (6 or 18 hr) were very similar. Because more vancomycin than teicoplanin diffused from the paper disk into the agar, the experimental conditions used would, if anything, tend to underestimate the difference between the diffusion coefficients of the two antibiotics. The size of the inhibition zone under standardized experimental conditions (medium, agar thickness, incubation temperature, and disk content) is a function of the diffusibility of the antibiotic, the susceptibility of the strain used, and its growth rate. During incubation, the antibiotic diffuses into the agar generating a concentration gradient, and the microorganism multiplies in those areas of the plate where the antibiotic concentration is lower than the MIC. For every strain there will be a time when, at a certain distance from the reservoir, the antibiotic concentration will be just below the inhibitory concentration. This is the site of the formation of the inhibition zone; the time when the inhibition zone formed is called the critical time (To) and the concentration at the zone edge is the critical concentration (Co). It is possible to determine T0 experimentally for a given microorganism, under standardized experimental conditions, by preincubating inoculated
Inhibition Zones Formed by Application of Sensidisks Diameter (mm) Found
Antibiotic
Amount
Predicted
Mean (n = 6)
SD
Teicoplanin Teicoplanin Teicoplanin
30 p,g 60 ~g 120 p.g
16.5 17.4 18.3
16.7 17.5 18.4
0.45 0.47 0.44
Vancomycin
30 b~g
18.4
18.1
0.42
Diffusion of Teicoplanin and Vancomycin
TABLE 2
Concentrations of Antibiotic at Different Distances from the Reservoir Calculated as a Function of Disk Charge and Diffusion Time 12
Time (hr) 4
6
8
257
Paper Disk Charge (~,g) Teico 30 60 75 90 Vanco 30 Teico 30 60 75 90 Vanco 30 Teico 30 60 75 90 Vanco 30
Radial Diffusion Diameter (ram) 13 14
15
Calculated Concentrations (~g/ml) 2.88 5.75 7.19 8.63
1.25 2.51 3.13 3.76
0.51 1.02 1.28 1.53
0.19 0.39 0.49 0.58
9.02
5.12
2.78
1.44
9.46 18.91 23.64 28.37
5.43 10.87 13.59 16.30
2.99 5.97 7.47 8.96
1.57 3.14 3.93 4.71
17.82
12.22
8.13
5.25
15.75 31.50 39.38 47.26
10.40 20.79 25.99 31.19
6.64 13.27 16.59 19.91
4.10 8.20 10.25 12.29
23.02
17.34
12.77
9.20
Teico, teicoplanin;and Vanco, vancomycin.
plates for different times before applying the charged disk and determining the preincubation time for which no inhibition zone is formed. It is more difficult to determine the critical concentration. In this study we have assumed it to be equal to the agar dilution MIC. By inserting the values of the diffusion coefficient D, the critical time (T0), and the critical concentration Cc (MIC) in equation 4, one can calculate an expected r2 (and, therefore, diameter) of the inhibition zone for an antibiotic. We have shown (Table 1) that the predicted diameters are very close to those found experimentally. This agreement between predicted and observed diameters suggests that our assumptions and calculations (determination of D; equivalence of MIC and critical concentration) are reasonable. The most relevant part of the diffusion gradient is the segment around the breakpoint, that is, the zone diameter that discriminates between sensitive and resistant microorganisms. The breakpoint is usually determined experimentally by correlating inhibition zones with MIC for a large number of microorganisms (scattergrams). With the methodology used in this work, it is
possible to calculate the range of concentrations of the antibiotic occurring at times likely to include the To of most strains at any given distance from the reservoir; one can thus estimate the breakpoint diameter for a given MIC breakpoint. If the MIC breakpoint were ~4 ~g/ml for sensitivity and >4 for resistance, then the diameter breakpoint would be ->15 mm for vancomycin 30 (concentration, 5.25 ~,g/ml), ---13 mm (5.43), ->14 mm (5.97), or ---15 mm (4.71) for teicoplanin 30, 60, and 90, respectively, assuming a To of 6 hr. With a breakpoint diameter -->14 mm for teicoplanin sensitivity (Barry et al., 1986 and 1987), some isolates with an MIC of 4 ~g/ml (and thus "susceptible") would appear resistant using a teicoplanin 30-~,g disk but susceptible with 60- or 90-~g disks. Our data indicate that the vancomycin 30-~g disk should perform adequately as indeed reported (Barry et al., 1986). The most difficult parameter to standardize in susceptibility determination with the agar diffusion method is the inoculum size. Variations in the inoculum size are reflected in variation in To (the time at which the inhibition zone is formed); the higher the inoculum, the shorter the To. Variations in in-
258
oculum have a greater effect on susceptibility determinations for teicoplanin than for vancomycin (and, probably, most other antibiotics) because, at any To, the gradient of teicoplanin is steeper than that of vancomycin (Table 2); at a given distance, the teicoplanin concentration declines more rapidly as a function of decreasing To. There have also been reports of Staphylococcus haemolyticus strains with reduced susceptibility to teicoplanin and large inhibition zones. This may be due to slow growth under test conditions (that is, long To). The aim of this study was to see whether, by altering the disk content of teicoplanin, we could generate a concentration profile similar to that ob-
L.A. Cavenaghi et al.
tained with a 30-~g vancomycin disk. Because the diffusion of the two antibiotics is different, it is possible to generate similar gradients only over smalldistance intervals. Extrapolation from our data indicate, for example, that a charge of 75 ~g teicoplanin would give a concentration gradient in the interval corresponding to the 13- to 15-ram diameter similar to that obtained with the 30-~g vancomycin disk.
The authors thank B. P. Goldstein for critical review of the manuscript.
REFERENCES Barry AL, Thornsberry C, Jones RN (1986) Evaluation of teicoplanin and vancomycin disk susceptibility tests. J Clin Microbiol 23:100-103. Barry AL, Jones RN, Gavan TL, Thornsberry C, and the Collaborative Antimicrobial SusceptibilityTesting Group (1987) Quality control limits for teicoplanin susceptibility tests and confirmation of disk diffusion interpretative criteria. J Clin Microbiol 25:1812-1814. Cavenaghi L, Corti A, Cassani G (1986) Comparison of the solid phase enzyme receptor assay (SPERA) and the microbiological assay for teicoplanin. J Hosp Infect 7(Suppl A):85-89. Cooper KE (1963) The theory of antibiotic inhibition zones.
In Analytical Microbiology. Ed, FW Kavanagh. New York and London: Academic Press, pp 1-86. Finn RK (1959) Theory of agar diffusion methods for bioassay. Anal Chem 31:975-977. Humphrey JH, Lightbown JW (1952) A general theory for plate assay of antibiotics with some practical applications. J Gen Microbiol 7:129-143. Parenti F (1988) Glycopeptide antibiotics. J Clin Pharmacol 28:136-140. National Committee for Clinical Laboratory Standards (NCCLS) (1988) Document M7-T2: methods for dilution antimicrobial susceptibility tests for bacteria that grow aerobically, 2d ed. Villanova, PA: NCCLS.
DIAGN MICROBIOLINFECT DIS 1992;15:253-258
Diffusion of Teicoplanin and Vancomycin in Agar Luigi A. Cavenaghi, Elia Biganzoli, Antonio Danese, and Francesco Parenti
Teicoplanin, although more active than vancomycin [by minimum inhibitory concentration (MIC)], produces smaller inhibition zones in sensitivity testing with 30-tzg disks. Our data support the hypothesis that this is due to lower diffusion of teicoplanin in agar media. After 6 hr of incubation, -70% of vancomycin, but only 20% of teicoplanin entered the agar from a paper disk charged with 30 p,g of antibiotic. This is due to a difference between the diffusion coefficients: 0.47 mm2/hr for teicoplanin and 0.72 mm2/hr for vancomycin. With the
methodology used in this work, it is possible to calculate the range of concentrations of the antibiotic occurring at times likely to include the critical time--the time when the inhibition zone is formed---of most strains at any given distance from the reservoir. One can thus estimate the breakpoint diameter for a given MIC breakpoint; for example, an MIC breakpoint of -15-mm breakpoint diameter for vancomycin (30-1~g disk) and a >-13-mm breakpoint diameter for teicoplanin (30-tzg disk).
INTRODUCTION
tity of teicoplanin to be charged to the paper disk.
Teicoplanin is a novel glycopeptide of the dalbaheptide type similar to vancomycin, from which it differs in net charge (negative while vancomycin's is positive), lipophilicity, and molecular size (Parenti, 1988). These factors are likely to affect the diffusibility in agar and, consequently, the size of the inhibition zone. Barry et al. (1986 and 1987) have f o u n d that disks charged with 30 ~g teicoplanin s h o w e d smaller inhibition zones t h a n 30-~g teicoplanin s h o w e d smaller inhibition zones t h a n 30-~g disks of vancomycin, even t h o u g h the strains were more sensitive to teicoplanin; this suggested a lower diffusibility of teicoplanin. The present study compares the diffusibility of teicoplanin and vancomycin in an agar m e d i u m with the aim of suggesting the appropriate quanFrom the Marion Merrell Dow Research Institute--Lepetit Center, Gerenzano (VA), Italy. Address reprint requests to Dr. L.A. Cavenaghi, Marion Merrell Dow Research Institute--Lepetit Center, via R. Lepetit 34, 1-21040 Gerenzano (VA), Italy. Received 28 March 1991; revised and accepted 1 July 1991. © 1992 Elsevier Science Publishing Co., Inc. 655 Avenue of the Americas, New York, NY 10010 0732-8893/92/$5.00
MATERIALS A N D M E T H O D S Media All studies were carried out in M u e l l e r - H i n t o n agar (Difco Laboratories, Detroit, MI); pH after autoclaving 7.4.
Antimicrobial Agents We used a commercial vial of Vancocin HC1 and a Lepetit analytic standard batch of teicoplanin.
Sensidisk Preparation The antibiotics were dissolved in water, then diluted with reagent grade methanol to appropriate concentrations. Solution (0.010 ml) was applied to paper disks (Schleicher and Schull, paper type 2668/2, 6m m diameter).
Diffusibility of the Antibiotics The melted m e d i u m , cooled to -50°C, was poured into o p e n - e n d e d polycarbonate cylinders (7-mm in-
254
ternal diameter), sealed at one end with Saran Wrap and allowed to solidify in the upright position. Sensidisks of the appropriate potency were applied to the tops of the cylinders, which were turned on their sides and incubated overnight at 35°C. After incubation the sensidisks were discarded, and the agar cylinder was extracted from the polycarbonate tube by applying pressure with a piston. As the cylinder came out of the tube, slices were cut with a sharp razor blade. Each slice was weighed, transferred to a small beaker into which 1 ml of pH 7.4 phosphate buffer (0.1 M) was added, and homogenized with an Ultra Turrax homogenizer. An aliquot was then assayed for antibiotic concentration with a microbiologic method using Bacillus subtilis ATCC 6633 as test organism (Cavenaghi et al., 1986). The assay was based on a parallel line design with detection limits of 0.02 and 0.35 mg/L for teicoplanin and vancomycin, respectively, and confidence limits of +-10%. The diffusibility was assessed after incubation times of 6 and 18 hr; vancomycin was tested at 30 txg, whereas teicoplanin was tested at 30, 60, and 90 lag.
M i n i m u m Inhibitory Concentration (MIC) A Staphylococcus aureus strain (ATCC 10832) was chosen, and its MIC was determined in duplicate according to the NCCLS guidelines (1988). For sensitivity testing, we used Petri dishes of 93-mm internal diameter with 25 ml of medium containing scalar dilutions of the antibiotics and 2 ill of a 1 0 7 colony forming units (CFU)/ml suspension applied to the surface. The MIC thus determined was assumed to be equal to the "critical concentration," that is, the minimal concentration inhibiting growth under the conditions of radial diffusion (Humphrey and Lightbown, 1952).
L.A. Cavenaghi et al.
(/~g/rnt) 4OO
300 0
o
rU
200 O C 0 0
*4
1O 0
o
+
-.,. ÷
0
'
0
' ....
1
2
3
z~ 5
distance
6 7 8 9 10 11 12 12 (ram) from reservoir
FIGURE 1 Diffusion profiles in agar tubes at 18 hr of teicoplanin (+, 30 txg; A, 60 txg; and O, 90 ixg) and vancomycin (A, 30 Ixg).
RESULTS The average concentrations of antibiotic in the sections of the agar cylinders after diffusion times of 18 and 6 hr are reported in Figures 1 and 2, respectively; also shown are the profiles obtained by plotting concentration versus distance from the reservoir. The total amount of antibiotic that diffused into the agar was calculated as the sum of the amounts (that is, concentration x volume of section) found in each section. Not all of the antibiotic present in the paper disk diffused into the agar: -72%-74% for vancomycin and only 20%-28.5% for teicoplanin in 6 hr. Prolonging the incubation time to 18 hr increased the amount of teicoplanin but not that of vancomycin, which diffused into the agar. These data clearly
(/~g/ml) 4OO
Critical Time The critical time (time of incubation after which no inhibition zone can be formed on a inoculated plate) was determined in duplicate according to Humphrey and Lightbown (1952). We used Petri dishes of 93-mm internal diameter with 25 ml of medium inoculated with 0.2 ml of a 1 0 7 CFU/ml suspension evenly distributed on the surface. The sensidisks prepared as described with 7.5, 15, 30, 60, 120, and 240 lag of the antibiotics were applied immediately after inoculation (0 hr) or after 100 min, 3, and 5 hr of preincubation at 35°C. The plates were then inverted and incubated for 18 hr. The inhibition zone diameters were measured, and the intercepts of the regression lines determined from plots of the square of the diameters versus preincubation time.
c 0
300
C
200
d) o c o o
A"..
•
1O0 it..
0 0
1
2
distance
3
4 5 6 7 8 (mm) from reservoir
FIGURE 2 Diffusion profiles in agar tube at 6 hr of teicoplanin (+, 30 p,g; A, 60 ~,g; and O, 90 Ixg) and vancomycin (A, 30 ~xg).
Diffusion of Teicoplanin and Vancomycin
255
indicate a lower agar diffusibility of teicoplanin in comparison with vancomycin. To calculate the diffusion coefficients, linearity of the gradient within each slice was assumed: that is, the concentration measured occurred at the midpoint of the slice. In each experiment, the slice closest to the reservoir was not included in the analysis as, owing to the steepness of the gradient near the reservoir, the concentration at midpoint would be much lower than the average concentration of the slice. In a linear diffusion gradient, the concentration can be described (Cooper, 1963; Finn, 1959) by the equation 1 C = Co .e 4D/;y2
(2)
If Co is assumed to remain constant, then one can plot the logarithm of C against the square of the distance, draw the best-fit regression line, and calculate the diffusion coefficient from D =
1
4kT
(3)
where k is the slope and T the time. Figure 3 shows the diffusion gradients at 6 hr for teicoplanin (30, 60, and 90 I~g) and vancomycin (30 t~g). The points are experimental values and the lines are the calculated best fit. The diffusion coefficients calculated from the curves of Figure 3 were 0.49, 0.46, and 0.47 mm2/hr (average 0.47) for 30, 60, and 90 t~g teicoplanin, respectively, and 0.69 for vancomycin. Diffusion coefficients were also calculated at 18 hr (curves not reported) and were found to be 0.40, 0.48, 0.53 mm2/hr (average 0.47) for teicoplanin (30, 60, and 90 t~g, respectively) and 0.74 for vancomycin. It is possible to calculate the expected diameter of the inhibition zone formed by an antibiotic diffusing radially using the equation
r2 = 4DTo [ I n ( M )
o
5 c 0
.~ .b-,
4
c o
3
c
0
o
2
c
1 0 0
(1)
where y is the distance from the reservoir in millimeters, C is the concentration in micrograms per milliliter at distance y, Co is the concentration in micrograms per milliliter in the reservoir, D is the diffusion coefficient in square millimeters per hour, and T is the elapsed time in hours. Equation 1 can also be expressed as Co _ 1 y2 In C 4DT
6
-ln(C~)-ln(4~DTo)]
10
20
squared
30
40
distance
50
60
( m m 2)
FIGURE 3 Diffusion gradient at 6 hr for teicoplanin (+, 30 lag; A, 60 t~g; and O, 90 lag) and vancomycin (A, 30 I~g). D is the diffusion coefficient (mm2/hr), M is the amount charged to the reservoir (t~g), Cc is the critical concentration (t~g/mm3), To is the critical time (hr), and H is the thickness of the agar (mm). In our experiments the thickness of the agar was 3.68 mm; the MICs were 0.5 mg/L (teicoplanin) and 1 mg/L (vancomycin) for the S. aureus strain used. The determination of critical times for teicoplanin and vancomycin are reported in Figure 4. The estimates were 5.9 hr for teicoplanin and 6.1 hr for vancomycin. In Table 1, the diameters of the inhibition zones calculated using equation 4 (assuming a To of 6 hr for both antibiotics, and a Cc equal to the MIC of each antibiotic for the strain of S. aureus used) are compared with those determined experimentally. In Table 2, we report our calculations of antibiotic concentrations at distances from the reservoir corresponding to zone diameters of 12, 13, 14, and 15 mm; concentrations were calculated for 30-1~g vancomycin disks and 30-, 60-, 75-, and 90-t~g teicoplanin disks at diffusion times of 4, 6, and 8 hr. One can see that, for a diffusion time of 6 hr (which is the time when the inhibition zone is formed) teicoplanin 75 gives a concentration profile similar to vancomycin 30. The concentrations would be lower for teicoplanin 75 than for vancomycin 30 at diameters ->15 ram. The concentrations generated by teicoplanin 75 are smaller than those of vancomycin 30 at a diffusion time of 4 hr. DISCUSSION
(4) where r is the radius of the inhibition zone (mm),
Barry and coworkers (1986 and 1987) reported that teicoplanin, although more active than vancomycin
256
L.A. Cavenaghi et al.
(mm z) 360 300
~
240
f~
z3
180 120 60 0
a
0
1
2
3
4
5
6
7
(mm~ 360 300 240 '13 "o
\ \
180 120
860 0 b
0
1
2
3
time
4
,.5
6
7
(h)
FIGURE 4 Determination of the critical times for (a) teicoplanin and (b) vancomycin (O, 7.5 ~g; A, 15 ~g, V, 30 ~g;, [] 60 bLg; 0, 120 ~g; and B, 240 p,g). (by MIC), produced smaller inhibition zones; they suggested that this was due to lower diffusion of teicoplanin in agar media. Our data support this hypothesis. After 6 hr of incubation (the time when the inhibition zone is formed), -70% of vancomycin,
TABLE 1
but only 20% of teicoplanin, entered the agar from a paper disk charged with 30 btg of antibiotic. This is due to a difference between the diffusion coefficients: 0.47 mma/hr for teicoplanin and 0.72 mm2/hr for vancomycin. The equation used to calculate the diffusion coefficient is rigorously valid when the concentration of antibiotic in the reservoir is constant throughout the diffusion period. This was not the case under the experimental conditions utilized in this work, as the antibiotic content of the disk was not replenished during incubation; this is unlikely, however, to affect significantly the calculated values of the diffusion coefficients, because the concentration of the antibiotic in the disk (which has a very small volume) remains much higher than the concentration in the first section of the agar. In fact, the diffusion coefficients measured with different charges of teicoplanin and at different diffusion times (6 or 18 hr) were very similar. Because more vancomycin than teicoplanin diffused from the paper disk into the agar, the experimental conditions used would, if anything, tend to underestimate the difference between the diffusion coefficients of the two antibiotics. The size of the inhibition zone under standardized experimental conditions (medium, agar thickness, incubation temperature, and disk content) is a function of the diffusibility of the antibiotic, the susceptibility of the strain used, and its growth rate. During incubation, the antibiotic diffuses into the agar generating a concentration gradient, and the microorganism multiplies in those areas of the plate where the antibiotic concentration is lower than the MIC. For every strain there will be a time when, at a certain distance from the reservoir, the antibiotic concentration will be just below the inhibitory concentration. This is the site of the formation of the inhibition zone; the time when the inhibition zone formed is called the critical time (To) and the concentration at the zone edge is the critical concentration (Co). It is possible to determine T0 experimentally for a given microorganism, under standardized experimental conditions, by preincubating inoculated
Inhibition Zones Formed by Application of Sensidisks Diameter (mm) Found
Antibiotic
Amount
Predicted
Mean (n = 6)
SD
Teicoplanin Teicoplanin Teicoplanin
30 p,g 60 ~g 120 p.g
16.5 17.4 18.3
16.7 17.5 18.4
0.45 0.47 0.44
Vancomycin
30 b~g
18.4
18.1
0.42
Diffusion of Teicoplanin and Vancomycin
TABLE 2
Concentrations of Antibiotic at Different Distances from the Reservoir Calculated as a Function of Disk Charge and Diffusion Time 12
Time (hr) 4
6
8
257
Paper Disk Charge (~,g) Teico 30 60 75 90 Vanco 30 Teico 30 60 75 90 Vanco 30 Teico 30 60 75 90 Vanco 30
Radial Diffusion Diameter (ram) 13 14
15
Calculated Concentrations (~g/ml) 2.88 5.75 7.19 8.63
1.25 2.51 3.13 3.76
0.51 1.02 1.28 1.53
0.19 0.39 0.49 0.58
9.02
5.12
2.78
1.44
9.46 18.91 23.64 28.37
5.43 10.87 13.59 16.30
2.99 5.97 7.47 8.96
1.57 3.14 3.93 4.71
17.82
12.22
8.13
5.25
15.75 31.50 39.38 47.26
10.40 20.79 25.99 31.19
6.64 13.27 16.59 19.91
4.10 8.20 10.25 12.29
23.02
17.34
12.77
9.20
Teico, teicoplanin;and Vanco, vancomycin.
plates for different times before applying the charged disk and determining the preincubation time for which no inhibition zone is formed. It is more difficult to determine the critical concentration. In this study we have assumed it to be equal to the agar dilution MIC. By inserting the values of the diffusion coefficient D, the critical time (T0), and the critical concentration Cc (MIC) in equation 4, one can calculate an expected r2 (and, therefore, diameter) of the inhibition zone for an antibiotic. We have shown (Table 1) that the predicted diameters are very close to those found experimentally. This agreement between predicted and observed diameters suggests that our assumptions and calculations (determination of D; equivalence of MIC and critical concentration) are reasonable. The most relevant part of the diffusion gradient is the segment around the breakpoint, that is, the zone diameter that discriminates between sensitive and resistant microorganisms. The breakpoint is usually determined experimentally by correlating inhibition zones with MIC for a large number of microorganisms (scattergrams). With the methodology used in this work, it is
possible to calculate the range of concentrations of the antibiotic occurring at times likely to include the To of most strains at any given distance from the reservoir; one can thus estimate the breakpoint diameter for a given MIC breakpoint. If the MIC breakpoint were ~4 ~g/ml for sensitivity and >4 for resistance, then the diameter breakpoint would be ->15 mm for vancomycin 30 (concentration, 5.25 ~,g/ml), ---13 mm (5.43), ->14 mm (5.97), or ---15 mm (4.71) for teicoplanin 30, 60, and 90, respectively, assuming a To of 6 hr. With a breakpoint diameter -->14 mm for teicoplanin sensitivity (Barry et al., 1986 and 1987), some isolates with an MIC of 4 ~g/ml (and thus "susceptible") would appear resistant using a teicoplanin 30-~,g disk but susceptible with 60- or 90-~g disks. Our data indicate that the vancomycin 30-~g disk should perform adequately as indeed reported (Barry et al., 1986). The most difficult parameter to standardize in susceptibility determination with the agar diffusion method is the inoculum size. Variations in the inoculum size are reflected in variation in To (the time at which the inhibition zone is formed); the higher the inoculum, the shorter the To. Variations in in-
258
oculum have a greater effect on susceptibility determinations for teicoplanin than for vancomycin (and, probably, most other antibiotics) because, at any To, the gradient of teicoplanin is steeper than that of vancomycin (Table 2); at a given distance, the teicoplanin concentration declines more rapidly as a function of decreasing To. There have also been reports of Staphylococcus haemolyticus strains with reduced susceptibility to teicoplanin and large inhibition zones. This may be due to slow growth under test conditions (that is, long To). The aim of this study was to see whether, by altering the disk content of teicoplanin, we could generate a concentration profile similar to that ob-
L.A. Cavenaghi et al.
tained with a 30-~g vancomycin disk. Because the diffusion of the two antibiotics is different, it is possible to generate similar gradients only over smalldistance intervals. Extrapolation from our data indicate, for example, that a charge of 75 ~g teicoplanin would give a concentration gradient in the interval corresponding to the 13- to 15-ram diameter similar to that obtained with the 30-~g vancomycin disk.
The authors thank B. P. Goldstein for critical review of the manuscript.
REFERENCES Barry AL, Thornsberry C, Jones RN (1986) Evaluation of teicoplanin and vancomycin disk susceptibility tests. J Clin Microbiol 23:100-103. Barry AL, Jones RN, Gavan TL, Thornsberry C, and the Collaborative Antimicrobial SusceptibilityTesting Group (1987) Quality control limits for teicoplanin susceptibility tests and confirmation of disk diffusion interpretative criteria. J Clin Microbiol 25:1812-1814. Cavenaghi L, Corti A, Cassani G (1986) Comparison of the solid phase enzyme receptor assay (SPERA) and the microbiological assay for teicoplanin. J Hosp Infect 7(Suppl A):85-89. Cooper KE (1963) The theory of antibiotic inhibition zones.
In Analytical Microbiology. Ed, FW Kavanagh. New York and London: Academic Press, pp 1-86. Finn RK (1959) Theory of agar diffusion methods for bioassay. Anal Chem 31:975-977. Humphrey JH, Lightbown JW (1952) A general theory for plate assay of antibiotics with some practical applications. J Gen Microbiol 7:129-143. Parenti F (1988) Glycopeptide antibiotics. J Clin Pharmacol 28:136-140. National Committee for Clinical Laboratory Standards (NCCLS) (1988) Document M7-T2: methods for dilution antimicrobial susceptibility tests for bacteria that grow aerobically, 2d ed. Villanova, PA: NCCLS.
