Context.— One purpose of early clinical trials is to establish the appropriate
dose of an antibiotic for phase 3 trials. Development of a relationship between
the ratio of drug exposure to organism minimum inhibitory concentration (MIC)
and therapeutic response early in the development process would allow an optimal
choice of dose to maximize response.
Objective.— To prospectively quantitate the relationship between plasma levels of
levofloxacin and successful clinical and/or microbiological outcomes and occurrence
of adverse events in infected patients.
Design.— Multicenter open-label trial.
Setting.— Twenty-two enrolling university-affiliated medical centers.
Patients.— A total of 313 patients with clinical signs and symptoms of bacterial
infections of the respiratory tract, skin, or urinary tract.
Main Outcome Measures.— Clinical response and microbiological eradication of pathogenic organisms.
Results.— Of 313 patients, 272 had plasma concentration-time data obtained. Of
these, 134 patients had a pathogen recovered from the primary infection site
and had an MIC of the pathogen to levofloxacin determined. These patients
constituted the primary analysis group for clinical outcome. Groups of 116
and 272 patients, respectively, were analyzed for microbiological outcome
and incidence of adverse events. In a logistic regression analysis, the clinical
outcome was predicted by the ratio of peak plasma concentration to MIC (Peak/MIC)
and site of infection (P<.001). Microbiological
eradication was predicted by the Peak/MIC ratio (P<.001).
Both clinical and microbiological outcomes were most likely to be favorable
if the Peak/MIC ratio was at least 12.2.
Conclusions.— Levofloxacin generated clinical and microbiological response rates of
95% and 96%, respectively. These response rates included fluoroquinolone "problem
pathogens," such as Streptococcus pneumoniae and Staphylococcus aureus. Exposure to levofloxacin was significantly
associated with successful clinical and microbiological outcomes. The principles
used in these analyses can be applied to other classes of drugs to develop
similar relationships between exposure and outcome. This pharmacokinetic modeling
could be used to determine optimal treatment dose in clinical trials in a
shorter time frame with fewer patients. This modeling also should be evaluated
for its potential to improve outcomes (maximizing therapeutic response, preventing
emergence of resistance, and minimizing adverse events) of patients treated
with this drug.