Download document

History of PK-PD:
1970s- 2008: From the Formative
Years to Mainstream Use
Michael N. Dudley, Pharm. D, FIDSA
Senior Vice President, Research and
Development and Chief Scientific Officer
MN Dudley
Michael N. Dudley, Pharm.D., FCCP
Vice President, Pharmacology and Microbiology
Essential Therapeutics, Mountain View, CA & Waltham, MA
San Diego, CA
MN Dudley, Craig PK-PD Symposium 2008
Efflux Pump
Inhibitors: A Viable
Option?
1
Four Decades of PK-PD: 1970-2008

What were the key tools?




Applications



Analytical
Models
Drugs
How evolved did this evolve?
Who cares? Now vs. Then
The Future
MN Dudley, Craig PK-PD Symposium 2008
2
Four Decades of Antibiotic PK-PD
" Those who cannot remember the past are condemned to
repeat it”.
George Santanaya in The Life of Reason, Volume 1, 1905;
US (Spanish-born) philosopher (1863 - 1952)
MN Dudley, Craig PK-PD Symposium 2008
3
Billy Craig:
Guitar Player or PK-PD Expert?
MN Dudley, Craig PK-PD Symposium 2008
4
1970’s: Understanding Antibiotic PK



Pharmacokinetic analysis tools
Aminoglycoside antibiotics
Protein binding of antimicrobials
MN Dudley, Craig PK-PD Symposium 2008
5
1970s: Importance of Individualizing Dose
Kinetic model for gentamicin dosing with the use of individual patient parameters.
Sawchuk RJ, Zaske DE, Cipolle RJ, Wargin WA, Strate RG. Clin Pharmacol Ther.
1977 Mar;21(3):362-9.
Multiple-infusion dosing regimens for gentamicin were established for 84 patients with the use of
individually calculated values of elimination kinetic parameters. Serum level-time data obtained
after a single infusion were used to determine the patient's gentamicin half-life (t 1/2) and
distribution volume. Patients with serum creatinine (Cr) less than 1.2 mg per 100 ml had t 1/2
(mean, 2.25 hr) and total body clearances (mean, 0.082 L/hr/kg) significantly different from
those with Cr greater than or equal to 1.2 mg/100 ml (means, 5.3 and 0.039, respectively).
Distribution volumes were not significantly different (means, 0.22 and 0.21 L/kg, respectively).
Calculations of dosing intervals and infusion rates, based on each patient's kinetic parameters
and desired steady-state peaks and nadirs, assumed a one-compartment model with first-order
elimination and 1-hr constant-rate input at fixed intervals. Follow-up steady-state peak and nadir
levels were measured in 63 of the regimens. Differences between predicted and measured
peak levels averaged --0.05 mug/ml with 60% of the measured values falling within 1 mug/ml of
that predicted. Predicted-measured nadir differences averaged --0.62 mug/ml (significantly
different from zero) indicating slight bias in the model. Fifty-six percent of these nadirs were
within 1 mug/ml of that predicted.
MN Dudley, Craig PK-PD Symposium 2008
6
1970s: A Mechanistic Understanding
Gentamicin tissue accumulation and nephrotoxic reactions.
Schentag JJ, Cubmo, TJ, Jusko, WJ, Plaut ME. JAMA. 1978 Nov
3;240(19):2067-9.
In 64 adults treated with gentamicin sulfate, peak and trough serum
concentrations rose gradually and declined in two phases after the final
dose. Seventeen patients experienced renal damage. The 17 patients had
greater amounts of gentamicin in tissues even after the first dose and
before any renal effects were noted. This pharmacokinetics analysis
provided evidence that patients who experience gentamicin-related
nephrotoxic effects while receiving recommended doses of gentamicin
could be distinguished from patients with no toxic effects because they
experienced abnormal tissue accumulation before detectable changes in
renal function occurred.
MN Dudley, Craig PK-PD Symposium 2008
7
MN Dudley, Craig PK-PD Symposium 2008
8
1970s: Considering the PD Opportunities
Persistent effect of antibiotics on Staphylococcus
aureus after exposure for limited periods of time. J
Infect Dis. 1977 Feb;135(2):217-23.
McDonald PJ, Craig WA, Kunin CM.
Persistent suppression of bacterial growth by certain antibiotics was tested by periodic counts of
viable organisms in a culture of Staphylococcus aureus that had been incubated in media
containing drugs for limited periods of time and then removed by centrifugation. During short (2
hr) periods of exposure of test cultures to penicillin G, cephalothin, erythromycin, clindamycin,
vancomycin, and tetracyline, effects on the growth of S. aureus were produced that persisted
after removal of the drug for periods of 1.7-4.1 hr. A persistent antibiotic effect was not
observed with gentamicin. The persistent effects of penicillin G and erythromycin were directly
related to duration of exposure and concentration of drug, up to a point of maximal response.
The maximal durations of bacterial suppression after exposure to penicillin G and erythromycin
were approximately 2 and 5 hr, respectively. These effects were observed over a wide range of
inocula.
MN Dudley, Craig PK-PD Symposium 2008
9
1980’s: Identifying the Paradigm


Model Development
Insights into Classes of Antibiotics




Beta-lactams
Fluoroquinolones
Aminoglycosides
Microcomputers
MN Dudley, Craig PK-PD Symposium 2008
10
MN Dudley, Craig PK-PD Symposium 2008
11
1980s: An Integration?
Integration of selected pharmacologic and microbiologic properties of
three new beta-lactam antibiotics: a hypothesis for rational
comparison. Rev Infect Dis. 1984 May-Jun;6(3):357-63
Drusano GL, Ryan PA, Standiford HC, Moody MR, Schimpff SC.
To enable the comparison of antibiotics in a standardized, meaningful manner, a system that integrates
the pharmacologic and microbiologic properties of beta-lactam antibiotics has been developed.
The system compares the duration of time that concentrations of both total and free drug exceed
the 90% minimal inhibitory concentration ( MIC90 ) of various pathogens as well as the area
under the concentration-time curve for both total and free drug, the latter measurement being an
indication of potential for antibiotic diffusion to the periphery. Of the three cephalosporin(-like)
compounds evaluated, moxalactam produced the longest duration of free drug above the MIC90
for the Enterobacteriaceae as well as the largest free-drug area under the concentration-time
curve following a standardized 2-g intravenous infusion. Both the duration of time cefotaxime was
above the MIC90 for the Enterobacteriaceae and the area under the concentration-time curve
were significantly less because of its short elimination half-life, results indicating the need for
more frequent dosing with cefotaxime than with moxalactam. Cefoperazone, because of its high
degree of protein binding and higher MIC90, develops the least duration of time above the MIC90
for most pathogens. None of these new agents provides free-drug concentrations above the
MIC90 for Pseudomonas aeruginosa for longer than 0.6 hr, which suggests that they may be
inadequate as single-agent therapy for many serious infections due to this pathogen. Because
promotion of comparisons by this method relies on in vitro data for analysis, the conclusions
relating to expected clinical efficacy are proffered and should be interpreted with
caution.(ABSTRACT TRUNCATED AT 250 WORDS)
SOME THINGS NEVER CHANGE!!
MN Dudley, Craig PK-PD Symposium 2008
12
or an Individualization?
Role for dual individualization with cefmenoxime.
Schentag JJ, Smith IL, Swanson DJ, DeAngelis C, Fracasso JE, Vari A,
Vance JW. Am J Med. 1984 Dec 21;77(6A):43-50.
Cefmenoxime concentration/effect relationships were retrospectively explored for gram-negative
bacteria isolated from 14 critical care patients treated for nosocomial pneumonia. The effects of
cefmenoxime concentrations on in vitro growth kinetics of 21 isolated pathogens were studied
using the Abbott MS-2 Research System, from which a dynamic response concentration was
derived. Serum pharmacokinetic profiles were obtained in each patient. These data were used
to calculate the in vivo total area under the curve over dynamic response concentration and the
time that cefmenoxime concentrations exceeded the dynamic response concentration for each
bacteria. The same determinations were made in 18 patients prospectively treated, except that
dosage was optimized on the basis of previous mathematical relations to achieve bacterial
eradication in four days. This method of dosage optimization is termed dual individualization.
Serial cultures of infected tissues were evaluated to determine the number of days to the
eradication of bacteria, and the pharmacokinetic and pharmacodynamic variables were used to
describe the bacteriologic response of the original pathogen isolated in pretreatment culture.
Bacterial eradication rates could be described from cefmenoxime pharmacokinetics in the
patient and from the relation between concentration and bacterial inhibition. Patients who were
prospectively treated using these retrospectively derived relationships had a predictable day of
bacterial eradication. This, in turn, was associated with a shorter duration of treatment (p less
than 0.05). The success of prospective dual individualization is encouraging and suggests that
more precise optimization of antibiotic dosage can yield a predictable rate of bacterial
eradication from the infection site.
MN Dudley, Craig PK-PD Symposium 2008
13
A PK-PD Target of Aminoglycosides
Clinical response to aminoglycoside therapy: importance of the ratio of peak
concentration to minimal inhibitory concentration. J Infect Dis. 1987
Jan;155(1):93-9.
Moore RD, Lietman PS, Smith CR.
In an examination of the relationships among plasma aminoglycoside concentrations,
the minimal inhibitory concentration (MIC) for the infecting organism, and
therapeutic outcome, data were analyzed from 236 patients with gram-negative
bacterial infections who were participants in four clinical trials of gentamicin,
tobramycin, and amikacin. Clinical response to therapy occurred in 188 (80%)
patients. Elevated maximal and mean peak aminoglycoside concentration/MIC
ratios were strongly associated with clinical response (P less than .00001 and P
less than .0001, respectively). A graded dose-response effect was found between
an increasing maximal peak concentration/MIC ratio and clinical response. By
logistic regression the peak concentration/MIC ratios were associated significantly
with clinical response after adjustment for underlying severity of illness and other
factors correlated with response. These results demonstrate that a high peak
concentration relative to the MIC for the infecting organism is a major determinant
of the clinical response to aminoglycoside therapy.
MN Dudley, Craig PK-PD Symposium 2008
14
1983
MN Dudley, Craig PK-PD Symposium 2008
15
1980s: An Important Tool is Introduced
Impact of dosing intervals on activity of gentamicin and ticarcillin against
Pseudomonas aeruginosa in granulocytopenic mice. J Infct Dis. 1983
May;147(5):910-7
Gerber AU, Craig WA, Brugger HP, Feller C, Vastola AP, Brandel J.
The influence of dosing intervals on the activity of gentamicin and ticarcillin against
Pseudomonas aeruginosa was studied in vivo. Granulocytopenic mice infected with
P. aeruginosa in the thigh muscle were treated with 1-hr or 3-hr injections of
gentamicin, ticarcillin, or gentamicin-ticarcillin. Plasma pharmacokinetics of the
drugs were correlated with antibacterial activity. Gentamicin injected every 1 hr
tended to be less active than gentamicin injected at longer intervals. In contrast,
ticarcillin given every 1 hr was significantly more efficacious than equivalent total
doses injected every 3 hr. The dosing schedule of gentamicin-ticarcillin was again
important for ticarcillin but did not appreciably affect the antibacterial activity of
gentamicin. Thus, antimicrobial chemotherapy of P. aeruginosa infections in the
granulocytopenic host might be improved by administering ticarcillin rather than
gentamicin as a constant infusion.
“I have a very bad
feeling about this..its
going to be a very
long decade”
MN Dudley, Craig PK-PD Symposium 2008
16
Population Analysis in the
Mouse Thigh Model
Gerber, Vestola, Brandel, and Craig, JID
1982
MN Dudley, Craig PK-PD Symposium 2008
17
In Vitro PK-PD Models
An in vitro model for the study of antibacterial dosage regimen design.
Toothaker RD, Welling PG, Craig WA. J Pharm Sci. 1982 Aug;71(8):861-4.
A model was developed that is capable of simulating antibacterial agent
concentration versus time profiles commonly observed following
intravenous and intramuscular bolus injections, intravenous infusions, and
oral doses, administered as single or multiple doses. The model consisted
of two physical compartments separated by a membrane of a commercial
hemodialyzer. The 1.08 m2 membrane surface area allowed rapid
transmembrane passage of drugs and other small molecules, while
membrane pore size prevented bacterial passage. These characteristics
allowed bacteria in one of the two compartments of the model to be
exposed to time-variant drug concentrations without affecting the number
or concentration of bacteria. The model was used to study the effects of a
multiple intravenous bolus dosage regimen of ampicillin on Escherichia
coli ATCC 12407.
MN Dudley, Craig PK-PD Symposium 2008
18
In Vitro Hollow Fiber Pharmacodynamic Model
For Study of Antibacterial, Antiviral Pharmacodynamics
Zinner, Blaser, Dudley,
Am J Med, 1986
MN Dudley, Craig PK-PD Symposium 2008
19
Blaser et. al., AAC 1987
MN Dudley, Craig PK-PD Symposium 2008
20
Approach for Identifying Optimal Dosage
Regimens for Antiinfectives is
Correlation of antimicrobial pharmacokinetic parameters with therapeutic efficacy
in an animal model. J Infect Dis. 1988 Oct;158(4):831-47
Vogelman B, Gudmundsson S, Leggett J, Turnidge J, Ebert S, Craig WA.
Medical Service, William S. Middleton Memorial Veterans Hospital, Madison, Wisconsin
Current antimicrobial dosing regimens are designed to maintain active drug levels for
most of the dosing interval and are based on 40-y-old observations. With use of
numerous multiple-dosing regimens in an animal model, this study is the first to
successfully minimize the interdependence between pharmacokinetic parameters
and thereby determine, by stepwise multivariate regression analysis, that the time
that serum levels exceeded the minimum inhibitory concentration (MIC) was the
most significant parameter determining efficacy for beta-lactams and erythromycin
against various pathogens, whereas the log area under the curve was the major
parameter for aminoglycosides. Optimal dosing intervals were no greater than the
time that serum levels exceeded the MIC plus the duration of the postantibiotic
effect. Careful application of these concepts should allow other investigators to use
more optimally dosed regimens than those previously used in preclinical trials and
to design studies to improve on current dosing regimens for humans.
MN Dudley, Craig PK-PD Symposium 2008
21

Key summary of the “state
of the art” and prospectus
for the future
MN Dudley, Craig PK-PD Symposium 2008
22
1990s: Wide Application of PK-PD





PK-PD studies become widely implemented in
preclinical models for new drugs
Early use acceptance by the FDA as part of
planning for clinical trials, analysis of data (Carl
Peck, MD)
Internet
Better microcomputers
Population pharmacokinetics
MN Dudley, Craig PK-PD Symposium 2008
23
Phase II Trial of IV Ciprofloxacin in Lower
Respiratory Infections





50 Patients with LRTI, 50% Failed Previous Rx
Ciprofloxacin 200-300 mg IV q 12 h.
All Isolates Intially Susc. to Ciprofloxacin (MIC < 1)
Pharmacokinetic Measurements
Results

Good Result for Pathogens w/ MICs < 0.25 mg/L

Poor Results with Infections due to Ps. aeruginosa
 69% of Isolates Persisting in Sputum were Resistant to
Ciprofloxacin.
Peloquin et. al. Arch Intern Med 1989;149:2269
MN Dudley, Craig PK-PD Symposium 2008
24
% of patients remaining culture-positive
Ciprofloxacin: Bacterial Eradication from
Sputum and AUC:MIC24 in Hospitalized Patients
100
Forrest A et al. Antimicrobial Agents Chemother 37:1073–1081, 1993.
75
AUC:MIC < 125
50
AUC:MIC 125-250
25
AUC:MIC > 250
0
0
2
4
6
8
10
12
14
Days of treatment
MN Dudley, Craig PK-PD Symposium 2008
25
MN Dudley, Craig PK-PD Symposium 2008
26
Modeling Bacterial Subpopulations for
Fluoroquinolones and Ps. aeruginosa
Observed Total CFU
Resistant Subpop
Sum of Susc. + Resist
Susc. Subpop
Dudley & Zinner, ICAAC 1987;
Jumbe
et.al,2008
2003
MN Dudley, Craig
PK-PD Symposium
27
Memorable and Historic Battles




George Drusano
Jerry Schentag
John Powers
Others?
MN Dudley, Craig PK-PD Symposium 2008
28
Comparison of Serum Levels of Amoxicillin
(7mg/kg Human Dose) in Humans and Mice
With Uranyl Nitrate-Induced Renal Failure
Andes & Craig, AAC 1997
MN Dudley, Craig PK-PD Symposium 2008
29
Effect of Amoxicillin 7mg/kg q8h on Str. pneumoniae With
Varying Levels of Penicillin Susceptibility in the
Neutropenic Mouse Thigh Model Using Human PK
Andes & Craig, AAC 1997
MN
Dudley
MN Dudley, Craig PK-PD Symposium 2008
30



Landmark paper showing full integration of information from animal
model studies, population PK and MC simulation to identify dosing
Breakpoints for MIC testing
Monte Carlo simulation becomes an important tool for drug
development as well as consensus work in setting breakpoints in
CLSI, EUCAST
MN Dudley, Craig PK-PD Symposium 2008
31
MN Dudley, Craig PK-PD Symposium 2008
32
William Craig, MD








Tireless researcher
Teacher and mentor
Enthusiastic investigator
Creative
Encourager to young investigators
Visionary
Organizer and “peacemaker”
Highly respected by peers, students
MN Dudley, Craig PK-PD Symposium 2008
33
The Duke or The Dean of PK-PD??
MN Dudley, Craig PK-PD Symposium 2008
34
MN Dudley, Craig PK-PD Symposium 2008
35
MouseThigh Infection Model
Strains with MexEF-OprN overexpression are more susceptible to levofloxacin
in vivo than strains without MexEF-OprN overexpression
3
nfxC strains (N = 8)
Emax model of nfxC strains
2
Non-nfxC strains (N = 14)
Emax model of non-nfxC strains
Change in Log CFU/thigh
1
0
-1
Emax = 4.8
EC50 = 29.6
MIC = 0.015 -2.2 ug/ml
2
-2
R = 0.87
Gamma = 3.8
-3
MN Dudley
-4
MIC = 0.125- 7.2 ug/ml
-5
10
Griffith et. al. AAC 2006
100
AUC:MIC
MN Dudley, Craig PK-PD Symposium 2008
Emax = 4.2
EC50 = 68.8
R2 = 0.86
Gamma = 3.2
Efflux Pump
Inhibitors: A Viable
Option?
1000
36
MN Dudley, Craig PK-PD Symposium 2008
37
Prospective Evaluation of a
Mathematical Model Predicted
Levofloxacin Regimen for
Suppression of Resistant
Subpopulations in Mouse Thigh
Model
AUC:MIC= 52
Resistance
Predicted
Jumbe et. al. J Clin Invest 2003;112:275.
Monte Carlo Simulation of Levofloxacin
750 mg QD for 10,000 Patients
Validation of Mathematical Model
Predicted AUC:MIC= 157 to
Prevent Selection of Resistance
MN Dudley, Craig PK-PD Symposium 2008
38
MN Dudley, Craig PK-PD Symposium 2008
39