Workpackage 1-Single Drug

Published

June 6, 2024

WORKPACKAGE 1A: QUANTITATIVE RELATIONSHIP OF TPOS WITH BACTERIAL INOCULUM

METHODS

In this package we will establish the quantitative relationship of Tpos and KPC-carbapenemase producing Klebsiella pneumoniae inoculum in blood culture bottles.

The methods for the assay are a modification of those originally proposed by Kaltsas et al.1 Detailed procedures can be found at this link. Methodology for preparing the test inoculum was adapted from CLSI M21A and M26A guidelines.

  • Briefly, tubes containing 1.8 mL of pooled healthy human serum are inoculated with 0.2 mL of a series of ten-fold dilutions (5x101 to 5x107 CFU/mL) of the standardised inoculum of each test indicator strains.
  • The sera were are then transferred into BacT⁄ALERT bottles without antibiotic inactivating matrix (Biomérieux Inc) for aerobic incubation for 24 hours and monitored for time to positivity.
  • Tpos results were used to establish preliminary assay quality control ranges by testing in triplicate for five KPC-carbapenemase producing K. pneumoniae strains (KPC A, B and C strains; NDM, and VIM producing strains) and a reference K. pneumoniae ATCC strain producing ESBL enzyme only. More detailed information on the isolates can be found on the Protocols section.
  • We also compared how Tpos results change if the organism suspension prepared in phosphate buffered saline (PBS-0.9%) versus pooled human serum (serum).

RESULTS

The relationship between Tpos and the K. pneumoniae inoculum is shown below. A linear relationship was observed from approximately 101-108 K. pneumoniae CFU/mL and a Tpos measured from 10.5 hours-4.5 hours over the tested inoculum range, with R2 of 0.92-0.94. The linear relationship of Tpos versus inoculum was consistent if the inoculum was prepared in PBS or pooled human serum as shown in Table 1.

Note

In the paper by Kaltsas et al.1 the reported inoculum was the inoculum introduced in to the bloodculture bottle, 30-40 mL of growth media (depending on the manufacturer).

Therefore we have followed this precedent of reporting the inoculum introduced into each bottle. Therefore, this does not represent the final test inoculum that need to account for the total 42 mL volume in each bloodculture bottle.

All experiments were performed using bloodculture bottles without antibiotic inactivating matrix

KLEBSIELLA PNEUMONIAE ISOLATES

Download data

Code 
library (ggplot2)
library(scales)
theme_set(theme_bw())
## import raw data from .csv file
wp1 <- read.csv("~/Desktop/ACUTEWEBSITE/wp1a.csv")
## plot raw data as x-y graph Tpos graph vs. drug concentrations stratified by dilution matrix, method="lm" is the method for linear regression
fig1 <-ggplot(wp1, aes(x=inoculum, y=tpos, color=isolates, shape=diluent, fill=isolates)) + geom_point(size=4, alpha = 0.5) + 
scale_y_continuous(name="Tpos (hr)", limits=c(4,12)) +
theme(legend.text=element_text(size=12)) +
geom_smooth(aes(linetype=diluent), method=lm , color="black", fill="#69b3a2", se=TRUE, inherit.aes = TRUE)
fig1
fig1 + theme_bw(base_size = 14)+ scale_x_log10(name="Inoculum CFU/mL", breaks = trans_breaks("log10",n=7, function(x) 10^x),labels = trans_format("log10", math_format(10^.x)))
Figure 1: Relationship of time-to-positivity (Tpos) versus test inoculum
Figure 2: Relationship of time-to-positivity (Tpos) versus test inoculum
Code 
library(forestmangr)
library(kableExtra)
wp1 <- read.csv("~/Desktop/ACUTEWEBSITE/wp1a.csv")
df <- data.frame (wp1)
table1<-lm_table(df, log(inoculum) ~ tpos, "diluent")
kbl(table1)%>%
kable_paper("hover", full_width = F, position="left")
Table 1: Effect of inoculum diluent on Tpos
diluent b0 b1 Rsqr Rsqr_adj Std.Error
pbs 32.05074 -2.564975 0.9343070 0.9336227 1.110872
serum 33.02361 -2.840223 0.9231703 0.9216023 1.210130

ESCHERICHIA COLI ATCC 25922

Download data

Code 
## import raw data from .xlsx file
library (readxl)
ecoliatcc_inoc <- read_excel("datasets_single/ecoliatcc_inoculum.xlsx")
library (ggplot2)
library(scales)
theme_set(theme_bw())
## plot raw data as x-y graph Tpos graph vs. drug concentrations stratified by dilution matrix, method="lm" is the method for linear regression
figecoli <-ggplot(ecoliatcc_inoc, aes(x=inoculum, y=tpos)) + geom_point(size=4, alpha = 0.5) + 
scale_y_continuous(name="Tpos (hr)", limits=c(4,12)) +
theme(legend.text=element_text(size=12)) +
geom_smooth(aes(linetype=diluent), method=lm , color="black", fill="#69b3a2", se=TRUE, inherit.aes = TRUE )
figecoli + theme_bw(base_size = 14)+ scale_x_log10(name="Inoculum CFU/mL", breaks = trans_breaks("log10",n=7, function(x) 10^x),labels = trans_format("log10", math_format(10^.x)))
Figure 3: Relationship of time-to-positivity (Tpos) versus test inoculum
Code 
library(forestmangr)
library(kableExtra)
library (readxl)
ecoliatcc_inoc <- read_excel("datasets_single/ecoliatcc_inoculum.xlsx")
dfecoli <- data.frame (ecoliatcc_inoc)
table1<-lm_table(dfecoli, log(inoculum) ~ tpos, "diluent")
kbl(table1)%>%
kable_paper("hover", full_width = F, position="left")
Table 2: Effect of inoculum diluent on Tpos
diluent b0 b1 Rsqr Rsqr_adj Std.Error
pbs 29.52142 -2.295468 0.9933486 0.9929985 0.3948541

ACINETOBACTER BAUMANII ATCC XXXX

Download data

Code 
library (ggplot2)
library(scales)
theme_set(theme_bw())
## import raw data from .csv file
library (readxl)
abaumani <- read_excel("abaum_pbs.xlsx")
## plot raw data as x-y graph Tpos graph vs. drug concentrations stratified by dilution matrix, method="lm" is the method for linear regression
fig1ab <-ggplot(abaumani, aes(x=inoculum, y=tpos)) + geom_point(size=4, alpha = 0.5) + 
scale_y_continuous(name="Tpos (hr)", limits=c(4,12)) +
theme(legend.text=element_text(size=12)) +
geom_smooth(aes(linetype=diluent), method=lm , color="black", fill="#69b3a2", se=TRUE, inherit.aes = TRUE )
fig1ab + theme_bw(base_size = 14)+ scale_x_log10(name="Inoculum CFU/mL", breaks = trans_breaks("log10",n=7, function(x) 10^x),labels = trans_format("log10", math_format(10^.x)))
Figure 4: Relationship of time-to-positivity (Tpos) versus test inoculum for Acinetobacter baumanii
Code 
library(forestmangr)
library(kableExtra)
library (readxl)
abaumani <- read_excel("abaum_pbs.xlsx")
df <- data.frame (abaumani)
table1<-lm_table(df, log(inoculum) ~ tpos, "diluent")
kbl(table1)%>%
kable_paper("hover", full_width = F, position="left")
Table 3: Effect of inoculum diluent on Tpos for Acinetobacter baumanii
diluent b0 b1 Rsqr Rsqr_adj Std.Error
pbs 28.35235 -2.158901 0.9924317 0.9921406 0.4157561

PSEUDOMONAS AERUGINOSA ATCC 27853

Download data

Code 
library (ggplot2)
library(scales)
theme_set(theme_bw())
## import raw data from .csv file
library (readxl)
pseudo <- read_excel("pseudo_pbs.xlsx")
## plot raw data as x-y graph Tpos graph vs. drug concentrations stratified by dilution matrix, method="lm" is the method for linear regression
fig2f <-ggplot(pseudo, aes(x=inoculum, y=tpos, color=isolates, fill=isolates)) + geom_point(size=4, alpha = 0.5) + 
scale_y_continuous(name="Tpos (hr)", limits=c(4,12)) +
theme(legend.text=element_text(size=12)) +
geom_smooth(aes(linetype=diluent), method=lm , color="black", fill="#69b3a2", se=TRUE, inherit.aes = TRUE )
fig2f + theme_bw(base_size = 14)+ scale_x_log10(name="Inoculum CFU/mL", breaks = trans_breaks("log10",n=7, function(x) 10^x),labels = trans_format("log10", math_format(10^.x)))
Figure 5: Relationship of time-to-positivity (Tpos) versus test inoculum for Pseudomonas aeruginosa
Code 
library(forestmangr)
library(kableExtra)
library (readxl)
pseudo <- read_excel("pseudo_pbs.xlsx")
df_pseudo <- data.frame (pseudo)
table2<-lm_table(df_pseudo, log(inoculum) ~ tpos, "diluent")
kbl(table2)%>%
kable_paper("hover", full_width = F, position="left")
Table 4: Effect of inoculum diluent on Tpos for pseudomonas aeruginosa
diluent b0 b1 Rsqr Rsqr_adj Std.Error
pbs 26.16852 -1.465734 0.9842056 0.9838108 0.59305

CONCLUSIONS

These data confirm that Tpos exhibits a wide dynamic range and reproducible relationship as a surrogate indicator for viable CFU/mL.

WORKPACKAGE 1B: EFFECT OF ANTIBIOTIC EXPOSURE ON TPOS

METHODS

This workpackage explores how antibiotic concentrations alter Tpos for a standard (1x104 CFU/mL) inoculum of Enterobacterales.

RESULTS

CEFTAZIDIME-AVIBACTAM

Preliminary experiments KPC (MIC 1 mg/L) and NDM (MIC > 64 mg/L) isolates

Download data

As shown in Figure 6, a marked increased in the Tpos for KPC_B from < 10 hrs to > 24 hrs when CAZ/AVI concentrations surpassed 1 mg/L. In contrast, the negative control KPC_NDM strain exhibited consistent Tpos < 10hr at all test concentrations with lack of antimicrobial activity. Trends in Tpos relative to antibiotic exposure were fitted by Loess.

Code 
library (ggplot2)
theme_set(theme_bw())
ceftaz <- read.csv("~/Desktop/ACUTEWEBSITE/datasets_single/ceftazidime.csv")
ggplot(ceftaz, aes(x=conc_s, y=tpos, color=isolate, shape=isolate)) + 
  geom_point(size=4, alpha = 0.7) + 
  geom_smooth(aes(linetype=isolate), method= loess, color="black", fill="#69b3a2", se=TRUE, inherit.aes = TRUE ) +
scale_x_log10(name="CAZ/AVI conc, mg/L") +
scale_y_continuous(name="Tpos (hr)", limits=c(4,25))
Figure 6: Effect of ceftazidime/aviactam concentrations on Tpos of K. pneumoniae



K. pneumoniae ATCC 700603 (MIC 0.75 mg/L) powder 4:1

Download data

Code 
## a four-parameter logistic regression model is fit to ceftazidime concentrations to estimated PD parameters
library (readxl)
library(drda)
caz_avi_7a <- read_excel("datasets_single/kpatcc_caz_avi_powder4-1.xlsx")
fit_atcc <- drda(tpos ~ ctz_s, data=caz_avi_7a, mean_function = "loglogistic4", max_iter = 1000)
plot(fit_atcc, xlab = "CAZ/AVI serum conc. (mg/L)", ylab = "Tpos (hr)")
Figure 7: Pharmacodynamic relationship of Tpos to ceftazidime/avibactam concentrations
Code 
## Analysis is repeated to produce a table reporting estimated EC10-EC95 parameter estiamtes plus 95% CI 
library (readxl)
library(drda)
library(broom)
library(kableExtra)
caz_avi_7a <- read_excel("datasets_single/kpatcc_caz_avi_powder4-1.xlsx")
fit_atcc <- drda(tpos ~ ctz_s, data=caz_avi_7a, mean_function = "loglogistic4", max_iter = 1000)
ed<-effective_dose(fit_atcc, y = c(0.10,0.25,0.50,0.75,0.90,0.95))
kbl(ed)%>%
kable_paper("hover", full_width = F, position = "left")
Table 5: Pharmacodynamic estimates
Estimate Lower .95 Upper .95
0.1 23.87874 19.31716 28.44033
0.25 26.93812 24.06902 29.80723
0.5 30.38948 29.11007 31.66888
0.75 34.28302 31.38838 37.17767
0.9 38.67541 33.84814 43.50268
0.95 41.98020 33.98412 49.97627



K. pneumoniae ATCC 700603 (MIC 0.75 mg/L) fixed 4 mg/L

Download data

Code 
## a four-parameter logistic regression model is fit to ceftazidime concentrations to estimated PD parameters
library (readxl)
library(drda)
caz_avi_7b <- read_excel("datasets_single/kpatcc_caz_avi_powderfix4.xlsx")
fit_atcc2 <- drda(tpos ~ ctz_s, data=caz_avi_7b, mean_function = "loglogistic4", max_iter = 1000)
plot(fit_atcc2, xlab = "CAZ/AVI serum conc. (mg/L)", ylab = "Tpos (hr)")
Figure 8: Pharmacodynamic relationship of Tpos to ceftazidime/avibactam concentrations
Code 
## Analysis is repeated to produce a table reporting estimated EC10-EC95 parameter estiamtes plus 95% CI 
library (readxl)
library(drda)
library(broom)
library(kableExtra)
caz_avi_7a <- read_excel("datasets_single/kpatcc_caz_avi_powderfix4.xlsx")
fit7a <- drda(tpos ~ ctz_s, data=caz_avi_7a, mean_function = "loglogistic4", max_iter = 1000)
ed<-effective_dose(fit7a, y = c(0.10,0.25,0.50,0.75,0.90,0.95))
kbl(ed)%>%
kable_paper("hover", full_width = F, position = "left")
Table 6: Pharmacodynamic estimates
Estimate Lower .95 Upper .95
0.1 8.099087 7.038485 9.159689
0.25 9.235833 8.728989 9.742676
0.5 10.532127 10.149119 10.915134
0.75 12.010361 10.856802 13.163921
0.9 13.696074 11.685276 15.706872
0.95 14.975851 12.415855 17.535846



KPC B (MIC 1 mg/L) commercial

Download data

Code 
## a four-parameter logistic regression model is fit to ceftazidime concentrations to estimated PD parameters
library (readxl)
library(drda)
caz_avi_4 <- read_excel("datasets_single/kpcb_caz_avi_com.xlsx")
fit4 <- drda(tpos ~ ctz_s, data=caz_avi_4, mean_function = "loglogistic4", max_iter = 1000)
plot(fit4, xlab = "CAZ/AVI serum conc. (mg/L)", ylab = "Tpos (hr)")
Figure 9: Pharmacodynamic relationship of Tpos to ceftazidime/avibactam concentrations
Code 
## Analysis is repeated to produce a table reporting estimated EC10-EC95 parameter estimates plus 95% CI 
library (readxl)
library(drda)
library(broom)
library(kableExtra)
caz_avi_4 <- read_excel("datasets_single/kpcb_caz_avi_com.xlsx")
fit4 <- drda(tpos ~ ctz_s, data=caz_avi_4, mean_function = "loglogistic4", max_iter = 1000)
ed<-effective_dose(fit4, y = c(0.10,0.25,0.50,0.75,0.90,0.95))
kbl(ed)%>%
kable_paper("hover", full_width = F, position = "left")
Table 7: Pharmacodynamic estimates
Estimate Lower .95 Upper .95
0.1 73.66731 71.92413 75.41048
0.25 78.61720 77.51043 79.72397
0.5 83.89969 83.05534 84.74404
0.75 89.53712 88.10788 90.96636
0.9 95.55335 93.12523 97.98147
0.95 99.87460 96.42020 103.32901



KPC B (MIC 1 mg/L) powder 4:1

Download data

Code 
## a four-parameter logistic regression model is fit to ceftazidime concentrations to estimated PD parameters
library (readxl)
library(drda)
caz_avi_6 <- read_excel("datasets_single/kpcb_caz_avi_powder_4-1.xlsx")
fit6 <- drda(tpos ~ ctz_s, data=caz_avi_6, mean_function = "loglogistic4", max_iter = 1000)
plot(fit6, xlab = "CAZ/AVI serum conc. (mg/L)", ylab = "Tpos (hr)")
Figure 10: Pharmacodynamic relationship of Tpos to ceftazidime/avibactam concentrations
Code 
## Analysis is repeated to produce a table reporting estimated EC10-EC95 parameter estimates plus 95% CI 
library (readxl)
library(drda)
library(broom)
library(kableExtra)
caz_avi_6 <- read_excel("datasets_single/kpcb_caz_avi_powder_4-1.xlsx")
fit6 <- drda(tpos ~ ctz_s, data=caz_avi_6, mean_function = "loglogistic4", max_iter = 1000)
ed<-effective_dose(fit6, y = c(0.10,0.25,0.50,0.75,0.90,0.95))
kbl(ed)%>%
kable_paper("hover", full_width = F, position = "left")
Table 8: Pharmacodynamic estimates
Estimate Lower .95 Upper .95
0.1 68.88793 61.14631 76.62954
0.25 75.52374 71.50425 79.54322
0.5 82.79876 79.85671 85.74080
0.75 90.77456 83.70890 97.84023
0.9 99.51866 86.97415 112.06317
0.95 105.94237 88.43755 123.44720



KPC A (MIC 2 mg/L) commercial

Download data

Code 
## a four-parameter logistic regression model is fit to ceftazidime concencentrations to estimated PD parameters
library (readxl)
library(drda)
caz_avi_7 <- read_excel("datasets_single/kpca_caz_avi_com.xlsx")
fit7 <- drda(tpos ~ ctz_s, data=caz_avi_7, mean_function = "loglogistic4", max_iter = 1000)
plot(fit7, xlab = "CAZ/AVI serum conc. (mg/L)", ylab = "Tpos (hr)")
Figure 11: Pharmacodynamic relationship of Tpos to ceftazidime/avibactam concentrations
Code 
## Analysis is repeated to produce a table reporting estimated EC10-EC95 parametersestiamtes plus 95% CI 
library (readxl)
library(drda)
library(broom)
library(kableExtra)
caz_avi_7 <- read_excel("datasets_single/kpca_caz_avi_com.xlsx")
fit7 <- drda(tpos ~ ctz_s, data=caz_avi_7, mean_function = "loglogistic4", max_iter = 1000)
ed<-effective_dose(fit7, y = c(0.10,0.25,0.50,0.75,0.90,0.95))
kbl(ed)%>%
kable_paper("hover", full_width = F, position = "left")
Table 9: Pharmacodynamic estimates
Estimate Lower .95 Upper .95
0.1 143.5857 138.9952 148.1763
0.25 154.0912 151.7888 156.3936
0.5 165.3653 163.4974 167.2333
0.75 177.4643 172.9862 181.9424
0.9 190.4485 183.1369 197.7601
0.95 199.8183 190.5343 209.1023



KPC A (MIC 2 mg/L) powder 4:1

Download data

Code 
library (readxl)
library(drda)
caz_avi_8 <- read_excel("datasets_single/kpca_caz_avi_powder_4-1.xlsx")
fit8 <- drda(tpos ~ ctz_s, data=caz_avi_8, mean_function = "loglogistic4", max_iter = 1000)
plot(fit8, xlab = "CAZ/AVI serum conc. (mg/L)", ylab = "Tpos (hr)")
Figure 12: Pharmacodynamic relationship of Tpos to ceftazidime/avibactam concentrations
Code 
## Analysis is repeated to produce a table reporting estimated EC10-EC95 parametersestiamtes plus 95% CI 
library (readxl)
library(drda)
library(broom)
library(kableExtra)
caz_avi_8 <- read_excel("datasets_single/kpca_caz_avi_powder_4-1.xlsx")
fit8 <- drda(tpos ~ ctz_s, data=caz_avi_8, mean_function = "loglogistic4", max_iter = 1000)
ed<-effective_dose(fit8, y = c(0.10,0.25,0.50,0.75,0.90,0.95))
kbl(ed)%>%
kable_paper("hover", full_width = F, position = "left")
Table 10: Pharmacodynamic relationship of Tpos to ceftazidime/avibactam concentrations
Estimate Lower .95 Upper .95
0.1 125.7716 114.3869 137.1562
0.25 143.8460 136.1835 151.5085
0.5 164.5179 158.8981 170.1377
0.75 188.1605 179.1859 197.1350
0.9 215.2007 197.7571 232.6444
0.95 235.7798 202.9734 268.5861



KPC A (MIC 2 mg/L) fixed 4 mg/L

Download data

Code 
## a four-parameter logistic regression model is fit to ceftazidime concentrations to estimated PD parameters
library (readxl)
library(drda)
caz_avi_9 <- read_excel("datasets_single/kpca_caz_avi_powder_fix4.xlsx")
fit9 <- drda(tpos ~ ctz_s, data=caz_avi_9, mean_function = "loglogistic4", max_iter = 1000)
plot(fit9, xlab = "CAZ/AVI serum conc. (mg/L)", ylab = "Tpos (hr)")
Figure 13: Pharmacodynamic relationship of Tpos to ceftazidime/avibactam concentrations
Code 
## Analysis is repeated to produce a table reporting estimated EC10-EC95 parameter estiamtes plus 95% CI 
library (readxl)
library(drda)
library(broom)
library(kableExtra)
caz_avi_9 <- read_excel("datasets_single/kpca_caz_avi_powder_fix4.xlsx")
fit9 <- drda(tpos ~ ctz_s, data=caz_avi_9, mean_function = "loglogistic4", max_iter = 1000)
ed<-effective_dose(fit9, y = c(0.10,0.25,0.50,0.75,0.90,0.95))
kbl(ed)%>%
kable_paper("hover", full_width = F, position = "left")
Table 11: Pharmacodynamic estimates
Estimate Lower .95 Upper .95
0.1 26.02302 22.05197 29.99406
0.25 37.21041 34.05205 40.36876
0.5 53.20729 50.04995 56.36463
0.75 76.08130 70.56354 81.59905
0.9 108.78892 81.37114 136.20669
0.95 138.74444 47.74622 229.74266



KPC B (MIC 1 mg/L) fixed 4 mg/L

Download data

Code 
## a four-parameter logistic regression model is fit to ceftazidime concentrations to estimated PD parameters
library (readxl)
library(drda)
caz_avi_10 <- read_excel("datasets_single/kpcb_caz_avi_powder_fix4.xlsx")
fit10<- drda(tpos ~ ctz_s, data=caz_avi_10, mean_function = "loglogistic4", max_iter = 1000)
plot(fit10, xlab = "CAZ/AVI serum conc. (mg/L)", ylab = "Tpos (hr)")
Figure 14: Pharmacodynamic relationship of Tpos to ceftazidime/avibactam concentrations
Code 
## Analysis is repeated to produce a table reporting estimated EC10-EC95 parameters estimates plus 95% CI 
library (readxl)
library(drda)
library(broom)
library(kableExtra)
caz_avi_10 <- read_excel("datasets_single/kpcb_caz_avi_powder_fix4.xlsx")
fit10 <- drda(tpos ~ ctz_s, data=caz_avi_10, mean_function = "loglogistic4", max_iter = 1000)
ed<-effective_dose(fit10, y = c(0.10,0.25,0.50,0.75,0.90,0.95))
kbl(ed)%>%
kable_paper("hover", full_width = F, position = "left")
Table 12: Pharmacodynamic estimates
Estimate Lower .95 Upper .95
0.1 21.16519 15.15120 27.17918
0.25 27.06283 22.30567 31.81998
0.5 34.60383 31.40317 37.80449
0.75 44.24612 39.07653 49.41571
0.9 56.57521 43.63228 69.51813
0.95 66.87013 38.60827 95.13199



KPC B (MIC 1 mg/L) high inoculum

Download data

We examined how the relationship of Tpos vs. ceftazidime/avibactam concentrations changes when tested against a higher inoculum (1x107 CFU/mL). The same methodology as described above was used for testing. As shown in Figure 15 and Table 13 the EC50 and EC90 were marginally higher when tested at the higher inoculum with a more shallow dose-response relationship.

Code 
## a four-parameter logistic regression model is fit to ceftazidime concentrations to estimated PD parameters
library (readxl)
library(drda)
caz_avi_5 <- read_excel("datasets_single/kpcb_caz_avi_com_hinoc.xlsx")
fithinc <- drda(tpos ~ ctz_s, data=caz_avi_5, mean_function = "loglogistic4", max_iter = 1000)
plot(fithinc, xlab = "CAZ/AVI serum conc. (mg/L)", ylab = "Tpos (hr)")
Figure 15: Pharmacodynamic relationship of Tpos to ceftazidime/avibactam concentrations against a high-inoculum
Code 
## Analysis is repeated to produce a table reporting estimated EC10-EC95 parameter estimates plus 95% CI 
library (readxl)
library(drda)
library(broom)
library(kableExtra)
caz_avi_5 <- read_excel("datasets_single/kpcb_caz_avi_com_hinoc.xlsx")
fithinc <- drda(tpos ~ ctz_s, data=caz_avi_5, mean_function = "loglogistic4", max_iter = 1000)
ed<-effective_dose(fithinc, y = c(0.10,0.25,0.50,0.75,0.90,0.95))
kbl(ed)%>%
kable_paper("hover", full_width = F, position = "left")
Table 13: Pharmacodynamic estimates for a high-inoculum
Estimate Lower .95 Upper .95
0.1 82.08148 71.35951 92.80344
0.25 117.28496 109.48524 125.08468
0.5 167.58667 159.06133 176.11201
0.75 239.46201 222.24533 256.67869
0.9 342.16357 290.42168 393.90546
0.95 436.16833 240.82684 631.50983



KPC Catania (MIC 16 mg/L) powder 4:1

Download data

Code 
## a four-parameter logistic regression model is fit to ceftazidime concentrations to estimated PD parameters
library (readxl)
library(drda)
caz_avi_catan <- read_excel("datasets_single/kpccatania_caz_avi_powder_4-1.xlsx")
fitcat <- drda(tpos ~ ctz_s, data=caz_avi_catan, mean_function = "loglogistic4", max_iter = 1000)
plot(fitcat, xlab = "CAZ/AVI serum conc. (mg/L)", ylab = "Tpos (hr)")
Figure 16: Pharmacodynamic relationship of Tpos to ceftazidime/avibactam concentrations against a resistant strain
Code 
## Analysis is repeated to produce a table reporting estimated EC10-EC95 parameter estimates plus 95% CI 
library (readxl)
library(drda)
library(broom)
library(kableExtra)
caz_avi_5 <- read_excel("datasets_single/kpccatania_caz_avi_powder_4-1.xlsx")
fitcat <- drda(tpos ~ ctz_s, data=caz_avi_catan, mean_function = "loglogistic4", max_iter = 1000)
ed<-effective_dose(fitcat, y = c(0.10,0.25,0.50,0.75,0.90,0.95))
kbl(ed)%>%
kable_paper("hover", full_width = F, position = "left")
Table 14: Pharmacodynamic estimates
Estimate Lower .95 Upper .95
0.1 357.0678 344.5988 369.5368
0.25 399.4416 391.4077 407.4755
0.5 446.8440 438.8360 454.8519
0.75 499.8716 486.9293 512.8140
0.9 559.1922 538.6586 579.7258
0.95 603.5119 566.2421 640.7818



KPC Catania (MIC 16 mg/L) fixed 4 mg/L

Download data

Code 
## a four-parameter logistic regression model is fit to ceftazidime concentrations to estimated PD parameters
library (readxl)
library(drda)
caz_avi_5 <- read_excel("datasets_single/kpccatania_caz_avi_powderfix.xlsx")
fitcat5 <- drda(tpos ~ ctz_s, data=caz_avi_5, mean_function = "loglogistic4", max_iter = 1000)
plot(fitcat5, xlab = "CAZ/AVI serum conc. (mg/L)", ylab = "Tpos (hr)")
Figure 17: Pharmacodynamic relationship of Tpos to ceftazidime/avibactam concentrations against a high-inoculum
Code 
## Analysis is repeated to produce a table reporting estimated EC10-EC95 parameter estimates plus 95% CI 
library (readxl)
library(drda)
library(broom)
library(kableExtra)
caz_avi_5 <- read_excel("datasets_single/kpccatania_caz_avi_powderfix.xlsx")
fit5 <- drda(tpos ~ ctz_s, data=caz_avi_5, mean_function = "loglogistic4", max_iter = 1000)
ed<-effective_dose(fit5, y = c(0.10,0.25,0.50,0.75,0.90,0.95))
kbl(ed)%>%
kable_paper("hover", full_width = F, position = "left")
Table 15: Pharmacodynamic estimates
Estimate Lower .95 Upper .95
0.1 227.1879 181.0095 273.3664
0.25 292.7908 255.0843 330.4973
0.5 377.3372 356.2659 398.4084
0.75 486.2972 454.9664 517.6279
0.9 626.7205 552.7132 700.7278
0.95 744.7437 610.2525 879.2349



KPC K. pneumoniae KPRAD (MIC 0.5 mg/L)

Download data

Code 
## a four-parameter logistic regression model is fit to ceftazidime concentrations to estimated PD parameters
library (readxl)
library(drda)
caz_avi_kprad <- read_excel("datasets_single/kp_PRAD_caz_avi_powder4-1.xlsx")
fitkprad <- drda(tpos ~ ctz_s, data=caz_avi_kprad, mean_function = "loglogistic4", max_iter = 1000)
plot(fitkprad, xlab = "CAZ/AVI serum conc. (mg/L)", ylab = "Tpos (hr)")
Figure 18: Pharmacodynamic relationship of Tpos to ceftazidime/avibactam concentrations against KPC- producing K. pneumoniae
Code 
## Analysis is repeated to produce a table reporting estimated EC10-EC95 parameter estimates plus 95% CI 
library (readxl)
library(drda)
library(broom)
library(kableExtra)
caz_avi_kprad <- read_excel("datasets_single/kp_PRAD_caz_avi_powder4-1.xlsx")
fitkprad <- drda(tpos ~ ctz_s, data=caz_avi_kprad, mean_function = "loglogistic4", max_iter = 1000)
ed<-effective_dose(fitkprad, y = c(0.10,0.25,0.50,0.75,0.90,0.95))
kbl(ed)%>%
kable_paper("hover", full_width = F, position = "left")
Table 16: Pharmacodynamic estimates
Estimate Lower .95 Upper .95
0.1 45.60828 44.00945 47.20710
0.25 48.96778 47.95219 49.98337
0.5 52.57475 51.68561 53.46388
0.75 56.44741 54.94883 57.94598
0.9 60.60532 58.19775 63.01290
0.95 63.60696 59.97950 67.23441

K. pneumoniae KFAB ESBL (MIC 0.38 mg/L)

Download data

Code 
## a four-parameter logistic regression model is fit to ceftazidime concentrations to estimated PD parameters
library (readxl)
library(drda)
caz_avi_kfab <- read_excel("datasets_single/kp_FAB_caz_avi_powder4-1.xlsx")
fitkfab <- drda(tpos ~ ctz_s, data=caz_avi_kfab, mean_function = "loglogistic4", max_iter = 1000)
plot(fitkfab, xlab = "CAZ/AVI serum conc. (mg/L)", ylab = "Tpos (hr)")
Figure 19: Pharmacodynamic relationship of Tpos to ceftazidime/avibactam concentrations against KPC- producing K. pneumoniae
Code 
## Analysis is repeated to produce a table reporting estimated EC10-EC95 parameter estimates plus 95% CI 
library (readxl)
library(drda)
library(broom)
library(kableExtra)
caz_avi_kfab <- read_excel("datasets_single/kp_FAB_caz_avi_powder4-1.xlsx")
fitkfab <- drda(tpos ~ ctz_s, data=caz_avi_kfab, mean_function = "loglogistic4", max_iter = 1000)
ed<-effective_dose(fitkfab, y = c(0.10,0.25,0.50,0.75,0.90,0.95))
kbl(ed)%>%
kable_paper("hover", full_width = F, position = "left")
Table 17: Pharmacodynamic estimates
Estimate Lower .95 Upper .95
0.1 18.54827 17.58236 19.51418
0.25 20.74190 20.01959 21.46420
0.5 23.19495 22.00443 24.38548
0.75 25.93813 23.96559 27.91067
0.9 29.00572 26.27565 31.73580
0.95 31.29690 27.87479 34.71901

K. pneumoniae WT (MIC 0.125 mg/L)

Download data

Code 
## a four-parameter logistic regression model is fit to ceftazidime concentrations to estimated PD parameters
library (readxl)
library(drda)
caz_avi_wt2 <- read_excel("datasets_single/kpwt_caz_avi_powder_4-1.xlsx")
fitwt <- drda(tpos ~ ctz_s, data=caz_avi_wt2, mean_function = "loglogistic4", max_iter = 1000)
plot(fitwt, xlab = "CAZ/AVI serum conc. (mg/L)", ylab = "Tpos (hr)")
Figure 20: Pharmacodynamic relationship of Tpos to ceftazidime/avibactam concentrations against a wt strain
Code 
library (readxl)
library(drda)
caz_avi_wt2 <- read_excel("datasets_single/kpwt_caz_avi_powder_4-1.xlsx")
fitwt <- drda(tpos ~ ctz_s, data=caz_avi_wt2, mean_function = "loglogistic4", max_iter = 1000)

## Analysis is repeated to produce a table reporting estimated EC10-EC95 parameter estimates plus 95% CI 
library (readxl)
library(drda)
library(broom)
library(kableExtra)
caz_avi_5 <- read_excel("datasets_single/kpwt_caz_avi_powder_4-1.xlsx")
fitwt2 <- drda(tpos ~ ctz_s, data=caz_avi_wt2, mean_function = "loglogistic4", max_iter = 1000)
ed<-effective_dose(fitwt, y = c(0.10,0.25,0.50,0.75,0.90,0.95))
kbl(ed)%>%
kable_paper("hover", full_width = F, position = "left")
Table 18: Pharmacodynamic estimates K. pneumoniae WT (MIC 0.125 mg/L)
Estimate Lower .95 Upper .95
0.1 4.104039 3.907545 4.300533
0.25 4.853535 4.727647 4.979422
0.5 5.739906 5.602824 5.876987
0.75 6.788149 6.522560 7.053738
0.9 8.027827 7.540413 8.515241
0.95 8.997963 8.099929 9.895996



E. coli ATCC 25922 (MIC 0.19 mg/L)

Download data

Code 
## a four-parameter logistic regression model is fit to ceftazidime concentrations to estimated PD parameters
library (readxl)
library(drda)
caz_avi_ecoli <- read_excel("datasets_single/ecoliatcc_caz_avi_powder4-1.xlsx")
fitecoli <- drda(tpos ~ ctz_s, data=caz_avi_ecoli, mean_function = "loglogistic4", max_iter = 1000)
plot(fitecoli, xlab = "CAZ/AVI serum conc. (mg/L)", ylab = "Tpos (hr)")
Figure 21: Pharmacodynamic relationship of Tpos to ceftazidime/avibactam concentrations against E. coli ATCC 25922 (MIC 0.19 mg/L)
Code 
## Analysis is repeated to produce a table reporting estimated EC10-EC95 parameter estimates plus 95% CI 
library (readxl)
library(drda)
library(broom)
library(kableExtra)
caz_avi_ecoli <- read_excel("datasets_single/ecoliatcc_caz_avi_powder4-1.xlsx")
fitecoli <- drda(tpos ~ ctz_s, data=caz_avi_ecoli, mean_function = "loglogistic4", max_iter = 1000)
ed<-effective_dose(fitecoli, y = c(0.10,0.25,0.50,0.75,0.90,0.95))
kbl(ed)%>%
kable_paper("hover", full_width = F, position = "left")
Table 19: Pharmacodynamic estimates for a high-inoculum
Estimate Lower .95 Upper .95
0.1 3.823371 3.241785 4.404957
0.25 4.312649 3.937461 4.687837
0.5 4.864540 4.644781 5.084298
0.75 5.487056 5.137118 5.836995
0.9 6.189236 5.480141 6.898332
0.95 6.717490 5.482243 7.952737

COMPARISON OF POSSIBLE INDICATOR ISOLATES

Download data:
kpca
kpcb
kpccatania
caz_avi_kfab
kpwt
caz_avi_kprad
kpatcc

Code 
## a four-parameter logistic regression model is fit to ceftazidime concentrations to estimated PD parameters
library (readxl)
library(drda)
kpca <- read_excel("datasets_single/kpca_caz_avi_powder_4-1.xlsx")
kpcb <- read_excel("datasets_single/kpcb_caz_avi_powder_4-1.xlsx")
kpccatania <- read_excel("datasets_single/kpccatania_caz_avi_powder_4-1.xlsx")
caz_avi_kfab <- read_excel("datasets_single/kp_FAB_caz_avi_powder4-1.xlsx")
kpwt <- read_excel("datasets_single/kpwt_caz_avi_powder_4-1.xlsx")
caz_avi_kprad <- read_excel("datasets_single/kp_PRAD_caz_avi_powder4-1.xlsx")
kpatcc <- read_excel("datasets_single/kpatcc_caz_avi_powder4-1.xlsx")

## fit models for each of the isolates
fitkpca <- drda(tpos ~ ctz_s, kpca, mean_function = "loglogistic4", max_iter = 1000)
fitkpcatcc <- drda(tpos ~ ctz_s, kpatcc, mean_function = "loglogistic4", max_iter = 1000)
fitkpcb <- drda(tpos ~ ctz_s, kpcb, mean_function = "loglogistic4", max_iter = 1000)
fitkpccatania <- drda(tpos ~ ctz_s, kpccatania, mean_function = "loglogistic4", max_iter = 1000)
fitkpwt <- drda(tpos ~ ctz_s, kpwt, mean_function = "loglogistic4", max_iter = 1000)
fitkfab <- drda(tpos ~ ctz_s, data=caz_avi_kfab, mean_function = "loglogistic4", max_iter = 1000)
fitkprad <- drda(tpos ~ ctz_s, data=caz_avi_kprad, mean_function = "loglogistic4", max_iter = 1000)
## plot all of the isolates together
p <-plot(fitkpwt, fitkfab, fitkprad,  fitkpcatcc, fitkpcb, fitkpca, fitkpccatania, base="10", xlab = "Ceftazidime-avibactam conc. (mg/L)", ylab = "Tpos (hr)",cex = 0.9, legend_location="topleft", legend = c("0.125 mg/L", "0.38 mg/L", "0.5 mg/L",  "0.75 mg/L", "1 mg/L", "2 mg/L", "16 mg/L"))
Figure 22: Pharmacodynamic relationship of Tpos to ceftazidime/avibactam concentrations

CONCLUSIONS

These data demonstrate that Tpos is a robust PD endpoint and the relationship of ceftazidime/avibactam concentrations versus Tpos follows classical sigmoidal dose response relationship with a steep transitional portion of the dose response curve that occurs near the MIC of the pathogen. Although ED50/90 estimates were broadly similar if the commercial (pharmaceutical) and analytical powder formulations were tested, used of high-fixed concentrations of avibactam or testing at very high K. pneumoniae inocula resulted in broader concentration-effect curves and higher EC50/90 estimates.

MEROPENEM

K. pneumoniae ATCC (MIC 0.03 mg/L)

  • The impact of increasing meropenem concentrations on Tpos observed with the ATCC ESBL producing Klebsiella pneumoniae was tested using similar methodology as previously described

  • Analytical powder was used to produce serum concentration

  • As shown in Figure 23, Tpos increased as meropenem concentrations surpassed the MIC, with an estimated pharmacodynamic parameters consistent with previous experiments that showed transition in the EC50/EC90 at simulated serum concentrations near the MIC as shown in Table 20.

Downlad data

Code 
## a four-parameter logistic regression model is fit to meropenem concentrations to estimated PD parameters
library (readxl)
library(drda)
mero <- read_excel("datasets_single/ATCC_meropenem.xlsx")
fit11 <- drda(tpos ~ mero_s, data=mero, mean_function = "loglogistic4", max_iter = 1000)
plot(fit11, xlab = "Meropenem serum conc. (mg/L)", ylab = "Tpos (hr)", xlim =c(0,100))
Figure 23: Pharmacodynamic relationship of Tpos to meropenem concentrations
Code 
## Analysis is repeated to produce a table reporting estimated EC10-EC95 parameter estimates plus 95% CI 
library (readxl)
library(drda)
library(broom)
library(kableExtra)
mero <- read_excel("datasets_single/ATCC_meropenem.xlsx")
fit11 <- drda(tpos ~ mero_s, data=mero, mean_function = "loglogistic4", max_iter = 1000)
ed<-effective_dose(fit11, y = c(0.10,0.25,0.50,0.75,0.90,0.95))
kbl(ed)%>%
kable_paper("hover", full_width = F, position = "left")
Table 20: Pharmacodynamic estimates
Estimate Lower .95 Upper .95
0.1 1.165027 1.026833 1.303220
0.25 1.460430 1.351831 1.569029
0.5 1.830736 1.764894 1.896577
0.75 2.294935 2.177266 2.412605
0.9 2.876837 2.586027 3.167648
0.95 3.354810 2.861582 3.848038
KPC B (MIC 32 mg/L)

Download data

Code 
## a four-parameter logistic regression model is fit to meropenem concentrations to estimated PD parameters
library (readxl)
library(drda)
mero2 <- read_excel("datasets_single/KPCB_meropenem.xlsx")
fit11 <- drda(tpos ~ mero_s, data=mero2, mean_function = "loglogistic4", max_iter = 1000)
plot(fit11, xlab = "Meropenem serum conc. (mg/L)", ylab = "Tpos (hr)", xlim =c(0,100))
Figure 24: Pharmacodynamic relationship of Tpos to meropenem concentrations
Code 
## Analysis is repeated to produce a table reporting estimated EC10-EC95 parameter estimates plus 95% CI 
library (readxl)
library(drda)
library(broom)
library(kableExtra)
mero2 <- read_excel("datasets_single/KPCB_meropenem.xlsx")
fit11 <- drda(tpos ~ mero_s, data=mero2, mean_function = "loglogistic4", max_iter = 1000)
ed<-effective_dose(fit11, y = c(0.10,0.25,0.50,0.75,0.90,0.95))
kbl(ed)%>%
kable_paper("hover", full_width = F, position = "left")
Table 21: Pharmacodynamic estimates
Estimate Lower .95 Upper .95
0.1 39.09324 38.35251 39.83397
0.25 40.62838 40.06818 41.18857
0.5 42.22380 41.61877 42.82882
0.75 43.88187 43.02567 44.73807
0.9 45.60505 44.37855 46.83155
0.95 46.81557 45.21113 48.42000
Comparison of ATCC and KPC B isolates


A comparison of meropenem activity against the ATCC ESBL-producing reference (MIC 0.03 mg/L) and KPC-carbapenemase producing K. pneumoniae isolates are shown in Figure 25. The meropenem EC50 measured by Tpos was 23-fold higher for the KPC-carbapenemase producing isolate versus the the ESBL-producing ATCC isolate. Despite variability in the response was noted for the KPC-B isolate at 20 mg/L and 80 mg/mL concentrations (experiments are currently being repeated), the EC50 were nearly identical (indicated by the dotted lines) when corrected for pathogen MIC. These data suggest that it may be possible to substitute sensitive indicator isolates with low MICs to predict pharmacodynamic responses of isolates with higher MICs.

Download dataset1
Download dataset2

Code 
## a four-parameter logistic regression model is fit to ceftazidime concentrations to estimated PD parameters
library (readxl)
library(drda)
mero2 <- read_excel("datasets_single/KPCB_meropenem.xlsx")
mero <- read_excel("datasets_single/ATCC_meropenem.xlsx")
fit4a <- drda(tpos ~ mero, mero, mean_function = "loglogistic4", max_iter = 1000)
fit4b <- drda(tpos ~ mero, mero2, mean_function = "loglogistic4", max_iter = 1000)
plot(fit4a, fit4b, xlab = "Meropenem conc. (mg/L)", ylab = "Tpos",cex = 0.9,legend = c("KP ATCC ESBL", "KP KPC_2"))
Figure 25: Pharmacodynamic relationship of Tpos to ceftazidime/avibactam concentrations

CONCLUSIONS

Similar to ceftazidime-avibactam, experiments with meropenem demonstrated that Tpos is a robust PD endpoint and the relationship of meropenem concentrations versus Tpos follows classical sigmoidal dose response relationship with a steep transitional portion of the dose response curve that occurs near the MIC of the pathogen.

A key observation is that the pharmacodynamics were similar when meropenem was tested against a highly-susceptible ATCC isolate (MIC 0.03 mg/L) and and the KPC-producing K.pneumoniae B isolate (MIC 32 mg/L) with at proportional difference in the EC50/90. Therefore, it may be possible to use highly-susceptible “indicator” strains for testing to predict activity against more resistant isolates. This is important because direct testing of 1 mL inoculum may routinely result in limited antimicrobial activity measured by Tpos (<10 hours) as dilutional effects when the serum samples is introduced into the bottle containing a total volume of 42 mL will reduce the actual testing concentrations of the antibiotics below the MIC for highly resistant pathogens.

This effect could be counteracted by testing with highly sensitive “indicator” isolates based on the expected serum concentrations of the antibiotic.

GENTAMICIN

KPC A (MIC 2 mg/L)

Download data

Code 
## a four-parameter logistic regression model is fit to gentamicin concentrations to estimated PD parameters
library (readxl)
library(drda)
gent1 <- read_excel("datasets_single/KPCA_gent.xlsx")
fit12 <- drda(tpos ~ gent_s, data=gent1, mean_function = "loglogistic4", max_iter = 1000)
plot(fit12, xlab = "Gentamicin serum conc. (mg/L)", ylab = "Tpos (hr)")
Figure 26: Pharmacodynamic relationship of Tpos to gentamicin concentrations
Code 
## Analysis is repeated to produce a table reporting estimated EC10-EC95 parameter estimates plus 95% CI 
library (readxl)
library(drda)
library(broom)
library(kableExtra)
gent1 <- read_excel("datasets_single/KPCA_gent.xlsx")
fit12 <- drda(tpos ~ gent_s, data=gent1, mean_function = "loglogistic4", max_iter = 1000)
ed<-effective_dose(fit12, y = c(0.10,0.25,0.50,0.75,0.90,0.95))
kbl(ed)%>%
kable_paper("hover", full_width = F, position = "left")
Table 22: Pharmacodynamic estimates
Estimate Lower .95 Upper .95
0.1 56.65250 53.43037 59.87462
0.25 62.53591 60.25523 64.81660
0.5 69.03033 67.01957 71.04109
0.75 76.19920 73.71019 78.68821
0.9 84.11256 79.42354 88.80158
0.95 89.95932 78.37830 101.54033
KPC B (MIC 0.5 mg/L)

Download data

Code 
## a four-parameter logistic regression model is fit to gentamicin concentrations to estimated PD parameters
library (readxl)
library(drda)
gent2 <- read_excel("datasets_single/KPCB_gent.xlsx")
fit13 <- drda(tpos ~ gent_s, data=gent2, mean_function = "loglogistic4", max_iter = 1000)
plot(fit13, xlab = "Gentamicin serum conc. (mg/L)", ylab = "Tpos (hr)")
Figure 27: Pharmacodynamic relationship of Tpos to gentamicin concentrations
Code 
## Analysis is repeated to produce a table reporting estimated EC10-EC95 parameter estimates plus 95% CI 
library (readxl)
library(drda)
library(broom)
library(kableExtra)
gent2 <- read_excel("datasets_single/KPCB_gent.xlsx")
fit13 <- drda(tpos ~ gent_s, data=gent2, mean_function = "loglogistic4", max_iter = 1000)
ed<-effective_dose(fit13, y = c(0.10,0.25,0.50,0.75,0.90,0.95))
kbl(ed)%>%
kable_paper("hover", full_width = F, position = "left")
Table 23: Pharmacodynamic estimates
Estimate Lower .95 Upper .95
0.1 18.56332 16.93133 20.19530
0.25 19.92766 19.06150 20.79382
0.5 21.39227 20.34817 22.43636
0.75 22.96452 20.81244 25.11660
0.9 24.65233 21.10862 28.19605
0.95 25.87062 20.96263 30.77861

COLISTIN

KPC B (MIC 0.125 mg/L)

Download data

Code 
## a four-parameter logistic regression model is fit to colistin concentrations to estimated PD parameters
library (readxl)
library(drda)
coli1 <- read_excel("datasets_single/KPCB_coli.xlsx")
fit14 <- drda(tpos ~ coli_s, data=coli1, mean_function = "loglogistic4", max_iter = 1000)
plot(fit14, xlab = "Colistin serum conc. (mg/L)", ylab = "Tpos (hr)")
Figure 28: Pharmacodynamic relationship of Tpos to colistin concentrations
Code 
## Analysis is repeated to produce a table reporting estimated EC10-EC95 parameter estimates plus 95% CI 
library (readxl)
library(drda)
library(broom)
library(kableExtra)
coli1 <- read_excel("datasets_single/KPCB_coli.xlsx")
fit14 <- drda(tpos ~ coli_s, data=coli1, mean_function = "loglogistic4", max_iter = 1000)
ed<-effective_dose(fit14, y = c(0.10,0.25,0.50,0.75,0.90,0.95))
kbl(ed)%>%
kable_paper("hover", full_width = F, position = "left")
Table 24: Pharmacodynamic estimates
Estimate Lower .95 Upper .95
0.1 3.814341 3.374386 4.254296
0.25 4.696574 4.395993 4.997155
0.5 5.782863 5.463324 6.102402
0.75 7.120404 6.500514 7.740293
0.9 8.767309 7.696486 9.838132
0.95 10.100091 8.130500 12.069682

MEROPENEM-VABORBACTAM

KPC B (MIC 0.06 mg/L)

Download data

Code 
## a four-parameter logistic regression model is fit to meropenem-vaborbactam concentrations to estimated PD parameters
library (readxl)
library(drda)
vabo1 <- read_excel("datasets_single/KPCB_vabo.xlsx")
fit14 <- drda(tpos ~ vabo_s, data=vabo1, mean_function = "loglogistic4", max_iter = 1000)
plot(fit14, xlab = "Meropenem-vabobactam serum conc. (mg/L)", ylab = "Tpos (hr)")
Figure 29: Pharmacodynamic relationship of Tpos to meropenem-vaborbactam concentrations
Code 
## Analysis is repeated to produce a table reporting estimated EC10-EC95 parameter estimates plus 95% CI 
library (readxl)
library(drda)
library(broom)
library(kableExtra)
vabo1 <- read_excel("datasets_single/KPCB_vabo.xlsx")
fit14 <- drda(tpos ~ vabo_s, data=vabo1, mean_function = "loglogistic4", max_iter = 1000)
ed<-effective_dose(fit14, y = c(0.10,0.25,0.50,0.75,0.90,0.95))
kbl(ed)%>%
kable_paper("hover", full_width = F, position = "left")
Table 25: Pharmacodynamic estimates
Estimate Lower .95 Upper .95
0.1 5.391292 4.092373 6.690211
0.25 7.797856 6.671079 8.924633
0.5 11.278661 10.098053 12.459269
0.75 16.313228 13.817427 18.809029
0.9 23.595123 18.006802 29.183445
0.95 30.327549 19.089870 41.565227

TIGECYCLINE

Important

Experiments performed using freshly collected blood samples that use EDTA as an anticoagulant result in enhanced tigecycline activity as previously reported by Deitchman et al.2 This phenomena was evident in preliminary experiments as shown in Figure 30.

Download data

Code 
library (readxl)
library (ggplot2)
theme_set(theme_bw())
edta <- read_excel("datasets_single/edtaeffect.xlsx")
ggplot(edta, aes(x=tig_s, y=tpos, color=tube, shape=tube)) + 
  geom_point(size=4, alpha = 0.7) + 
scale_x_continuous (name="Tigecycline conc, mg/L") +
scale_y_continuous(name="Tpos (hr)", limits=c(4,25))
Figure 30: Effect of tigeycvcline concentrations on Tpos of K. pneumoniae using serum collected with and withouth EDTA

KPC-A (MIC 0.75 mg/L)

Download data

Code 
## a four-parameter logistic regression model is fit to tigecycline concentrations to estimated PD parameters
library (readxl)
library(drda)
tig1 <- read_excel("datasets_single/kpca_tig.xlsx")
fit15 <- drda(tpos ~ tig_s, data=tig1, mean_function = "loglogistic4", max_iter = 1000)
plot(fit15, xlab = "Tigecycline serum conc. (mg/L)", ylab = "Tpos (hr)")
Figure 31: Pharmacodynamic relationship of Tpos to tigecycline concentrations
Code 
## Analysis is repeated to produce a table reporting estimated EC10-EC95 parameter estimates plus 95% CI 
library (readxl)
library(drda)
library(broom)
library(kableExtra)
tig1 <- read_excel("datasets_single/kpca_tig.xlsx")
fit15 <- drda(tpos ~ tig_s, data=tig1, mean_function = "loglogistic4", max_iter = 1000)
ed<-effective_dose(fit15, y = c(0.10,0.25,0.50,0.75,0.90,0.95))
kbl(ed)%>%
kable_paper("hover", full_width = F, position = "left")
Table 26: Pharmacodynamic estimates
Estimate Lower .95 Upper .95
0.1 26.21193 22.64088 29.78298
0.25 33.33312 30.73822 35.92801
0.5 42.38896 40.56986 44.20806
0.75 53.90507 50.72689 57.08325
0.9 68.54984 60.64442 76.45527
0.95 80.72308 56.72937 104.71680

KPC-B (MIC 0.38 mg/L)

Download data

Code 
## a four-parameter logistic regression model is fit to tigecycline concentrations to estimated PD parameters
library (readxl)
library(drda)
tig2 <- read_excel("datasets_single/kpcb_tig.xlsx")
fit16 <- drda(tpos ~ tig_s, data=tig2, mean_function = "loglogistic4", max_iter = 1000)
plot(fit16, xlab = "Tigecycline serum conc. (mg/L)", ylab = "Tpos (hr)")
Figure 32: Pharmacodynamic relationship of Tpos to tigecycline concentrations
Code 
## Analysis is repeated to produce a table reporting estimated EC10-EC95 parameter estimates plus 95% CI 
library (readxl)
library(drda)
library(broom)
library(kableExtra)
tig2 <- read_excel("datasets_single/kpca_tig.xlsx")
fit16 <- drda(tpos ~ tig_s, data=tig2, mean_function = "loglogistic4", max_iter = 1000)
ed<-effective_dose(fit16, y = c(0.10,0.25,0.50,0.75,0.90,0.95))
kbl(ed)%>%
kable_paper("hover", full_width = F, position = "left")
Table 27: Pharmacodynamic estimates
Estimate Lower .95 Upper .95
0.1 26.21193 22.64088 29.78298
0.25 33.33312 30.73822 35.92801
0.5 42.38896 40.56986 44.20806
0.75 53.90507 50.72689 57.08325
0.9 68.54984 60.64442 76.45527
0.95 80.72308 56.72937 104.71680

E. coli ATCC 25922 (MIC 0.125 mg/L)

Download data

Code 
## a four-parameter logistic regression model is fit to tigecycline concentrations to estimated PD parameters
library (readxl)
library(drda)
tigecoli <- read_excel("datasets_single/ecoliatcc_tig.xlsx")
fittigecoli <- drda(tpos ~ tig_s, data=tigecoli, mean_function = "loglogistic4", max_iter = 1000)
plot(fittigecoli, xlab = "Tigecycline serum conc. (mg/L)", ylab = "Tpos (hr)")
Figure 33: Pharmacodynamic relationship of Tpos to tigecycline concentrations
Code 
## Analysis is repeated to produce a table reporting estimated EC10-EC95 parameter estimates plus 95% CI 
library (readxl)
library(drda)
library(broom)
library(kableExtra)
tigecoli <- read_excel("datasets_single/ecoliatcc_tig.xlsx")
fittigecoli <- drda(tpos ~ tig_s, data=tigecoli, mean_function = "loglogistic4", max_iter = 1000)
ed<-effective_dose(fittigecoli, y = c(0.10,0.25,0.50,0.75,0.90,0.95))
kbl(ed)%>%
kable_paper("hover", full_width = F, position = "left")
Table 28: Pharmacodynamic estimates
Estimate Lower .95 Upper .95
0.1 4.099751 3.742536 4.456966
0.25 6.123028 5.816673 6.429382
0.5 9.144816 8.798910 9.490721
0.75 13.657892 12.976866 14.338918
0.9 20.398226 18.087056 22.709396
0.95 26.796582 18.484108 35.109056

E. coli ATCC 25922 (MIC 0.125 mg/L) protein binding

Download data

Code 
## a four-parameter logistic regression model is fit to ceftazidime concentrations to estimated PD parameters
library (readxl)
library(drda)
serum <- read_excel("datasets_single/ecoliatcc_tig_serum.xlsx")
pbs <- read_excel("datasets_single/ecoliatcc_tig_pbs.xlsx")
fitserum <- drda(tpos ~ tig_s, serum, mean_function = "loglogistic4", max_iter = 1000)
fitpbs <- drda(tpos ~ tig_s, pbs, mean_function = "loglogistic4", max_iter = 1000)
plot(fitserum, fitpbs, xlab = "Tigecycline conc. (mg/L)", ylab = "Tpos",cex = 0.9,legend = c("serum", "PBS"))
Figure 34: Pharmacodynamic relationship of Tpos to ceftazidime/avibactam concentrations
Code 
## a four-parameter logistic regression model is fit to ceftazidime concentrations to estimated PD parameters
library (readxl)
library(drda)
library(broom)
library(kableExtra)
serum <- read_excel("datasets_single/ecoliatcc_tig_serum.xlsx")
pbs <- read_excel("datasets_single/ecoliatcc_tig_pbs.xlsx")
fitserum <- drda(tpos ~ tig_s, serum, mean_function = "loglogistic4", max_iter = 1000)
fitpbs <- drda(tpos ~ tig_s, pbs, mean_function = "loglogistic4", max_iter = 1000)
edserum<-effective_dose(fitserum, y = c(0.10,0.25,0.50,0.75,0.90,0.95))
edpbs<-effective_dose(fitpbs, y = c(0.10,0.25,0.50,0.75,0.90,0.95))
kbl(edserum)%>%
kable_paper("hover", full_width = F, position = "left")
Estimate Lower .95 Upper .95
0.1 2.123423 2.085530 2.161317
0.25 3.333204 3.299344 3.367064
0.5 5.232235 5.196848 5.267621
0.75 8.213202 8.144849 8.281556
0.9 12.892521 12.733520 13.051522
0.95 17.519723 17.041417 17.998028
Figure 35: Pharmacodynamic relationship of Tpos to ceftazidime/avibactam concentrations
Code 
kbl(edpbs)%>%
kable_paper("hover", full_width = F, position = "left")
Estimate Lower .95 Upper .95
0.1 4.283286 3.530962 5.035611
0.25 6.954032 6.402163 7.505901
0.5 11.290059 10.592118 11.988000
0.75 18.329717 17.061603 19.597832
0.9 29.758793 17.553542 41.964045
0.95 41.376913 0.000000 90.646797
Figure 36: Pharmacodynamic relationship of Tpos to ceftazidime/avibactam concentrations

CEFTAZIDIME

K. pneumoniae ATCC 700603 (MIC 12 mg/L)

Download data

Code 
## a four-parameter logistic regression model is fit to ceftazidime concentrations to estimated PD parameters
library (readxl)
library(drda)
ceftaz_atcc <- read_excel("datasets_single/kpatcc_ceftazidime.xlsx")
fit_ceftazatcc <- drda(tpos ~ ctz_s, data=ceftaz_atcc, mean_function = "loglogistic4", max_iter = 1000)
plot(fit_ceftazatcc, xlab = "Ceftazidime serum conc. (mg/L)", ylab = "Tpos (hr)")
Figure 37: Pharmacodynamic relationship of Tpos to ceftazidime concentrations against an ESBL producing ATCC strain
Code 
## Analysis is repeated to produce a table reporting estimated EC10-EC95 parameter estimates plus 95% CI 
library (readxl)
library(drda)
library(broom)
library(kableExtra)
ceftaz_atcc <- read_excel("datasets_single/kpatcc_ceftazidime.xlsx")
fit_ceftazatcc <- drda(tpos ~ ctz_s, data=ceftaz_atcc, mean_function = "loglogistic4", max_iter = 1000)
ed<-effective_dose(fit_ceftazatcc, y = c(0.10,0.25,0.50,0.75,0.90,0.95))
kbl(ed)%>%
kable_paper("hover", full_width = F, position = "left")
Table 29: Pharmacodynamic estimates
Estimate Lower .95 Upper .95
0.1 245.8990 206.8423 284.9557
0.25 295.0812 265.6254 324.5370
0.5 354.1003 335.4615 372.7390
0.75 424.9237 398.3337 451.5137
0.9 509.9125 446.8074 573.0176
0.95 577.2350 434.9467 719.5233

K. pneumoniae KP-DAM (MIC 4 mg/L)

Download data

Code 
## a four-parameter logistic regression model is fit to ceftazidime concentrations to estimated PD parameters
library (readxl)
library(drda)
ceftaz1 <- read_excel("datasets_single/kp_dam_ceftaz.xlsx")
fit17 <- drda(tpos ~ ctz_s, data=ceftaz1, mean_function = "loglogistic4", max_iter = 1000)
plot(fit17, xlab = "Ceftazidime serum conc. (mg/L)", ylab = "Tpos (hr)")
Figure 38: Pharmacodynamic relationship of Tpos to ceftazidime concentrations
Code 
## Analysis is repeated to produce a table reporting estimated EC10-EC95 parameter estimates plus 95% CI 
library (readxl)
library(drda)
library(broom)
library(kableExtra)
ceftaz1 <- read_excel("datasets_single/kp_dam_ceftaz.xlsx")
fit17 <- drda(tpos ~ ctz_s, data=ceftaz1, mean_function = "loglogistic4", max_iter = 1000)
ed<-effective_dose(fit17, y = c(0.10,0.25,0.50,0.75,0.90,0.95))
kbl(ed)%>%
kable_paper("hover", full_width = F, position = "left")
Table 30: Pharmacodynamic estimates
Estimate Lower .95 Upper .95
0.1 107.0276 78.27442 135.7809
0.25 136.4083 116.29285 156.5238
0.5 173.8544 165.49765 182.2112
0.75 221.5801 196.00814 247.1520
0.9 282.4071 224.19394 340.6202
0.95 333.0621 230.10196 436.0223

CIPROFLOXACIN

K. pneumoniae WT (MIC <= 0.06)

Download data

Code 
## a four-parameter logistic regression model is fit to ceftazidime concentrations to estimated PD parameters
library (readxl)
library(drda)
cipro1 <- read_excel("datasets_single/kpwt_cipro_powder.xlsx")
fitcipro1 <- drda(tpos ~ cipro_s, data=cipro1, mean_function = "loglogistic4", max_iter = 1000)
plot(fitcipro1, xlab = "Ciprofloxacin serum conc. (mg/L)", ylab = "Tpos (hr)")
Figure 39: Pharmacodynamic relationship of Tpos to ceftazidime concentrations
Code 
## Analysis is repeated to produce a table reporting estimated EC10-EC95 parameter estimates plus 95% CI 
library (readxl)
library(drda)
library(broom)
library(kableExtra)
cipro1 <- read_excel("datasets_single/kpwt_cipro_powder.xlsx")
fitcipro1 <- drda(tpos ~ cipro_s, data=cipro1, mean_function = "loglogistic4", max_iter = 1000)
ed<-effective_dose(fitcipro1, y = c(0.10,0.25,0.50,0.75,0.90,0.95))
kbl(ed)%>%
kable_paper("hover", full_width = F, position = "left")
Table 31: Pharmacodynamic estimates
Estimate Lower .95 Upper .95
0.1 0.2429975 0.1876993 0.2982956
0.25 0.3259756 0.2760740 0.3758772
0.5 0.4372888 0.3916951 0.4828826
0.75 0.5866130 0.5247666 0.6484594
0.9 0.7869281 0.6566179 0.9172382
0.95 0.9609705 0.7089475 1.2129934

KPC B (MIC <= 0.38)

Download data

Code 
## a four-parameter logistic regression model is fit to ceftazidime concentrations to estimated PD parameters
library (readxl)
library(drda)
cipro2 <- read_excel("datasets_single/kpcb_cipro_powder.xlsx")
fitcipro2 <- drda(tpos ~ cipro_s, data=cipro2, mean_function = "loglogistic4", max_iter = 1000)
plot(fitcipro2, xlab = "Ciprofloxacin serum conc. (mg/L)", ylab = "Tpos (hr)")
Figure 40: Pharmacodynamic relationship of Tpos to ceftazidime concentrations
Code 
## Analysis is repeated to produce a table reporting estimated EC10-EC95 parameter estimates plus 95% CI 
library (readxl)
library(drda)
library(broom)
library(kableExtra)
cipro2 <- read_excel("datasets_single/kpcb_cipro_powder.xlsx")
fitcipro2 <- drda(tpos ~ cipro_s, data=cipro2, mean_function = "loglogistic4", max_iter = 1000)
ed<-effective_dose(fitcipro2, y = c(0.10,0.25,0.50,0.75,0.90,0.95))
kbl(ed)%>%
kable_paper("hover", full_width = F, position = "left")
Table 32: Pharmacodynamic estimates
Estimate Lower .95 Upper .95
0.1 7.164090 6.590145 7.738036
0.25 8.982345 8.565516 9.399174
0.5 11.262075 10.893533 11.630618
0.75 14.120404 13.417518 14.823290
0.9 17.704179 16.211318 19.197041
0.95 20.648354 17.657499 23.639209

REFERENCES

1.
Kaltsas, P., Want, S. & Cohen, J. Development of a time-to-positivity assay as a tool in the antibiotic management of septic patients. Clinical microbiology and infection: the official publication of the European Society of Clinical Microbiology and Infectious Diseases 11, 109–114 (2005).
2.
Deitchman, A. N., Singh, R. S. P., Rand, K. H. & Derendorf, H. Enhanced in vitro tigecycline activity in the presence of chelating agents. International journal of antimicrobial agents 51, 799–802 (2018).