Dosing of Beta-Lactam Antibiotics in Critically Ill Patients Is Often Inadequate
By Stan Deresinski, MD, FACP, FIDSA
Clinical Professor of Medicine, Stanford University, Editor of Infectious Disease Alert
Dosing of beta-lactam antibiotics in critically ill patients is often inadequate and results in poor clinical outcomes.
Roberts JA, Paul SK, Akova M, et al; DALI Study. DALI: defining antibiotic levels in intensive care unit patients: are current beta-lactam antibiotic doses sufficient for critically ill patients? Clin Infect Dis. 2014; 58:1072-83.
The in vivo antibacterial effect of beta-lactam antibiotics is time-dependent. That is, it depends on the proportion of the inter-dosing interval that the serum concentration of free (non-protein bound) drug remains above the minimum inhibitory concentration (MIC) of the infecting pathogen. A number of studies have suggested that usual antibiotic dosing practices often produce inadequate drug exposure in critically ill patients. This is frequently the consequence of the fact that these patients often exhibit pharmacokinetics (PK) that differ from those seen in the healthy volunteers who participate in Phase 1 studies that evaluate dosages.
In addressing this issue, Roberts and colleagues evaluated the adequacy of clinician-chosen dosing of beta-lactam antibiotics in critically ill patients by performing a point-prevalence study of pharmacodyamics (PD) target achievement. They studied 361 evaluable adult patients in 68 ICUs in 10 countries. All had a creatinine clearance >30 mL/min; their mean Apache score was 18. The antibiotics were administered for prophylaxis in approximately one-third and for therapy of infection in the remainder.
Among those with infection, a bacterial pathogen was isolated in 72.9% and an MIC was available in 34.2% of these. Of the pathogens identified, 18% were Pseudomonas aeruginosa and 16% were Escherichia coli. A positive clinical outcome was achieved in 144 of the 248 (58.1%) who received the antibiotic because of existing infection. One or more antibiotics in addition to a beta-lactam was administered to 77 (62%) of patients and a favorable outcome was achieved in 63% compared to 50% in the beta-lactam monotherapy cohort.
The beta-lactam was administered by prolonged infusion (>2 hours) to one-third of patients and by "bolus" in the remainder. While prolonged infusion failed to achieve the minimal target of 50% fT>MIC (free drug concentration greater than the MIC for greater than 50% of the dosing interval the least aggressive target) in only 7%, bolus administration missed this pharmacodynamics target in 20%. Of the total treated for infection,16% failed to achieve this pharmacodynamic target and, when compared to those who reached it, these were 32% less likely to have a favorable clinical outcome (OR, 0.68; 95% CI 0.52 to 0.91; P=0.009). Multivariate regression analysis identified the following as independent factors associated with clinical outcome: APACHE II score, SOFA score, and PK/PD targets both 50% fT>MIC and 100% fT>MIC. Of note is that achievement of the 100% fT>MIC was associated with a somewhat higher likelihood of a positive clinical response than was 50% fT>MIC. Among the 24 patients with bacteremia, reaching 50% fT>MIC was associated with a significantly improved likelihood of a favorable clinical outcome when compared to those who did not achieve this threshold, but this was not demonstrated with either pulmonary or intra-abdominal infection.
COMMENTARY
This study, while ambitious and yielding useful results, has a number of shortcomings that have been acknowledged by the authors.
These include lack of identification of a pathogen in many patients as well as a lack of accurate (or, in some cases, any) MIC data for many isolates a problem that was dealt with by using the high end of published data.
The fact that only 38% of patients received monotherapy with a beta-lactam is another obvious confounder. Actual antibiotic doses administered are also not reported.
Nonetheless, the results are consistent with an expanding literature regarding the inadequacy of antibiotic dosing in many critically ill patients together with retrospective studies reporting improved outcomes in patients with Gram-negative bacteremia receiving prolonged, as opposed to brief, beta-lactam infusions.1
This is, in part, likely a consequence of the fact that pharmacokinetics are frequently altered in critically ill patients who may have increased volume of distribution and, in some, increased drug clearance.
Overall, the data are as clear as they are likely to get that many critically ill patients receiving beta-lactam therapy are significantly under-dosed, either in terms of actual dose or of the duration of infusion, and that such under-dosing is associated with poor clinical outcomes.
All patients in this study had creatinine clearances >30 mL/min. One also wonders what proportion of patients receiving renal replacement therapy are receiving beta-lactam antibiotics in doses according to published guidelines are achieving even minimal pharmacodynamic targets.
An immediate solution to the problem of inadequate beta-lactam exposure in critically ill patients is intervention by knowledgeable Infectious Diseases consultants and Antimicrobial Stewardship Programs. The ultimate answer is therapeutic drug monitoring, something that is not currently generally available.
Reference
- Deresinski S. Extended infusion cefepime. In the Literature. Clin Infect Dis 2013:57: iii=iv.
