NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.
Lux LJ, Posey RE, Daniels LS, et al. Pharmacokinetic/Pharmacodynamic Measures for Guiding Antibiotic Treatment for Hospital-Acquired Pneumonia [Internet]. Rockville (MD): Agency for Healthcare Research and Quality (US); 2014 Nov. (Comparative Effectiveness Reviews, No. 136.)
Pharmacokinetic/Pharmacodynamic Measures for Guiding Antibiotic Treatment for Hospital-Acquired Pneumonia [Internet].
Show detailsThis chapter begins with the results of our literature search and a general description of the included studies of the effects of using pharmacokinetic/pharmacodynamics (PK/PD) measures for dosing and other decisions for hospital-acquired pneumonia (HAP). It is then organized by Key Question (KQ) and grouped by intervention. For each KQ, we give the key points, a more detailed synthesis of the literature, and the strength of evidence (SOE) grades. Additional details for included studies can be found in evidence tables (Appendix D).
Results of Literature Searches
Results of our searches appear in Figure 3. From an unduplicated pool of 2,134 possible articles, we excluded 1,894 at the title and abstract review stage and another 240 at the full-text review stage.
Figure 3
Disposition of articles about using PK/PD measures in hospital-acquired pneumonia. IPA = International Pharmaceutical Abstracts; KQ = Key Question; PK/PD = pharmacokinetic/pharmacodynamic; SIP = Scientific Information Packet
We included 10 studies reported in 11 published articles. Of these, one study pertained to KQ 1; nine pertained to KQ 2. We identified no studies addressing KQ 3 on subgroups.
Description of Included Studies
Table 4 describes the 10 included studies (listed in alphabetical order by first author). Seven studies were randomized controlled trials (RCTs).54-61Two were prospective cohort studies;62,63 one was a retrospective cohort study.64 All seven RCTs addressed KQ 2; three were conducted by the same group of investigators in the United States, and the other four were conducted in the United States, Thailand, Germany and China. One prospective cohort study for KQ 2 was conducted in Italy and the other in India. The retrospective cohort study for KQ 1 was performed in Spain. Five RCTs were funded by the pharmaceutical industry; one trial and two cohort studies were supported by government or an academic institution; and two studies, one trial and one cohort, reported no source of support. We rated five of the trials and one cohort study as medium risk of bias and two trials and two cohort studies as high risk of bias.
Table 4
Characteristics of included studies.
Key Question 1. PK/PD Measures for Dosing or Monitoring
Key Points
One prospective cohort study (high risk of bias) found significantly improved outcomes in terms of cure rates and mortality, although both measures were poorly constructed. Specifically, the study defined “cure” as no further specimens obtained for microbiologic testing, and the mortality outcome included both death and patients who left the hospital against medical advice.62 Evidence is insufficient to draw conclusions about the effect of using PK/PD measures for dosing or monitoring on intermediate and health outcomes.
Detailed Synthesis
Scaglione et al. studied a sample of patients receiving mechanical ventilation and who were treated in a special PK/PD program in Italy.62 The study excluded immunocompromised patients such as those with HIV, cystic fibrosis, active tuberculosis, lung cancer or another malignancy metastatic to the lungs, sepsis, or severe renal failure. The authors noted that they did not present their data on the three-way comparison of the impact of measuring and adjusting (versus not measuring and adjusting versus not measuring and not adjusting); however, they concluded that their analyses demonstrated that patients with PK/PD measures and subsequent dose adjustments had the best outcomes. We assessed this study as high risk of bias because of multiple reasons: unclear methods, outcomes inconsistent with definitions, and potential confounding.
Intermediate and Health Outcomes
The investigators defined clinical success as the absence or improvement of clinically significant symptoms and signs requiring no additional therapy. Those patients who had both PK/PD measures (serum concentration and minimum inhibitory concentration [MIC] monitoring) had a higher percentage classified as a success than those who had only one or no test (82 percent versus 68 percent, p=not reported) (Table 5). Clinical failure was defined as persistence or progression of symptoms and signs, or death. Failure was statistically significantly lower in patients who had both PK/PD measures than in those who did not (18 percent versus 32 percent, p<0.001) (Table 5). Patients who received both the serum concentration and MIC monitoring had a nonsignificantly lower duration of mechanical ventilation days than patients who received only one test or none (Table 5). Of the 205 patients in the group with both PK/PD measures, 81 had antibiotic dose adjustments based on the PK/PD information; however, the authors did not present their analyses based on those who received dose changes or not.
Table 5
Clinical response, days of mechanical ventilation, and mortality or other health outcome.
Of those patients who died or left the hospital against medical advice, patients who had both serum concentration and MIC monitoring had significantly lower mortality (10 percent versus 24 percent, p<0.001) than those who had one test or none (Table 5). Mortality was, however, a composite measure comprising undefined mortality (did not specify time interval or whether death occurred in the hospital or after discharge) and leaving hospital against medical advice; it is not a validated measure. The authors did not present any other evidence on relapse, reinfection, superinfection, mortality due to pneumonia, mortality in-hospital, or mortality within 30 days of discharge.
Antibiotic-Related Adverse Events
This prospective cohort study did not address organ toxicity, hematological effects, Clostridium difficile (C. difficile) infection, or antibiotic resistance. The investigators stated that all treatments were well tolerated and that study groups did not differ on these outcomes.
Strength of Evidence
For KQ 1, evidence was insufficient for the four outcomes addressed: clinical response, mechanical ventilation, treatment failure, and mortality. The evidence base was a single study with a high risk of bias (Table 6).62
Table 6
Strength of evidence for using PK/PD measures to influence dosing or monitoring.
Key Question 2. Prolonged or Continuous Infusions
Key Points
Evidence is insufficient to draw conclusions about the effect of continuous infusions compared with the effect of intermittent infusions on outcomes related to clinical response, mechanical ventilation, morbidity, or mortality. The evidence consisted of two small trials.55,57,59 and one prospective cohort.63
Evidence is insufficient to draw conclusions about the effect of continuous infusions versus intermittent infusions on the rates of antibiotic-related adverse events.54-58,60,64
Detailed Synthesis
KQ 2 addresses the issue of whether using prolonged or continuous infusions as compared with using bolus infusions for beta-lactams affects (a) clinical response or mechanical ventilation, (b) morbidity or mortality, or (c) rates of antibiotic-related adverse events. Our synthesis included nine studies (10 articles).54-61,63,64 All nine studies included patients with HAP in the intensive care unit (ICU) setting. Seven were RCTs;54-61 one was an historical cohort study,64 and one a prospective cohort.63 Characteristics of the patients in these studies are shown in Table 7.
Table 7
Severity of illness and other population characteristics.
Of the nine studies in our KQ 2 analysis, four medium risk of bias studies (three trials, one prospective cohort) evaluated the effect of continuous versus intermittent administration of beta-lactam antibiotics on intermediate clinical outcomes, duration of mechanical ventilation, and superinfection.55,57,59-61,63
Four RCTs (two medium risk of bias and two high risk of bias) reported rates of antibiotic-related adverse events.54,56,58,60
We excluded one study (high risk of bias) from the analysis of intermediate outcomes and morbidity or mortality because it was retrospective.64 We included it for the analysis of rates of adverse events.
Of the three studies rated high risk of bias, one study received this rating because of high risk of selection bias and confounding.64 The second study received this rating because of high risk of selection bias, measurement bias, and confounding.56 The third study had a very small number of patients, a high risk of selection bias, and confounding.60 Appendix B presents detailed information on risk-of-bias ratings.
Intermediate and Health Outcomes
Three RCTs and one prospective cohort study met our criteria for assessment of intermediate and health outcomes (Table 8). One open-label RCT reported clinical response, length of mechanical ventilation, and superinfection.55,57 The investigators excluded immunocompromised patients such as those with AIDS and neutropenia. Clinical cure was defined as complete resolution of all signs and symptoms of pneumonia and improvement or lack of progression of all abnormalities on the chest radiograph; improvement was defined as improvement of signs and symptoms of pneumonia with evidence of infection remaining. Failure was defined as persistence or progression of signs and symptoms of pneumonia, development of new pulmonary or extra-pulmonary clinical findings consistent with active infection, progression of radiographic abnormalities, or death from infection. Clinical cure, improvement, or failure did not differ significantly between the two groups.55,57
Table 8
Intermediate and health outcomes for studies addressing Key Question 2.
Another RCT, also using ceftazidime, defined success as complete resolution of all signs and symptoms of pneumonia and improvement or in lack of progression of all abnormalities on the chest radiograph.59 Patients with creatinine clearances of <30 mL/min or bacterial pathogens resistant to ceftazidime were excluded. The percentage of patients achieving success was higher in the intermittent infusion group than in the continuous infusion group, but the difference was not statistically significant (56% versus 71%, p = 0.63).
Two trials, one randomized61 and the other nonrandomized,63 used the Clinical Pulmonary Infection Score (CPIS) as their marker of success. In the randomized trial, all patients were infected with A. baumannii and treated with meropenem. Success was a CPIS of <6; the authors presented mean scores, with no statistical testing for differences, for days 3, 5, and 7. All patients achieved a CPIS of <6 by day 7.61 In the nonrandomized trial, investigators excluded immunocompromised patients (i.e., AIDS, neutropenia) and those with early-onset HAP or HCAP without any risk factors for multidrug-resistant pathogens.63 The investigators evaluated the use of piperacillin/tazobactam by either intermittent infusion or prolonged infusion for the treatment of VAP. The two groups did not differ significantly in CPIS scores on days 1, 3 or 8; mean CPIS scores in each group rose at each measurement day, ending with 8.51 (intermittent) versus 8.60 (prolonged) on day 8.
We excluded one study from the analysis of intermediate and health outcomes because of its retrospective design.64
Duration of mechanical ventilation also did not differ significantly between the groups in the three trials.55,57,59,63 One trial presented data on relapse and mortality; on both measures, differences between groups were not statistically significant.61
No investigators reported on rates of reinfection; two trials reported on rates of superinfection.55,57,59 In one trial, methicillin-resistant Staphylococcus aureus (MRSA) occurred in one patient in the intermittent infusion group and in no patient in the continuous infusion group.55,57 The other trial reported high rates of superinfection, most commonly with A. calcoaceticus. Pneumonia caused by Acinetobacter calcoaceticus occurred in 44 percent of their continuous infusion group and 22 percent of the intermittent infusion group.59 For patients with treatment failures, 71 percent of the continuous infusion group and 75 percent of the intermittent infusion group developed superinfections.59 Neither study presented any results for tests of statistical significance for these data.
Antibiotic-Related Adverse Events
Six studies (four RCTs; one retrospective and one prospective cohort study) reported information on rates of antibiotic-related adverse events (Table 9). Four studies reported no adverse events attributed to the treatment regimens.54,56,58,60 One RCT (n=41) reported nephrotoxicity in three patients—two patients in the continuous infusion group and one patient in the intermittent infusion group; all patients had received concomitant IV tobramycin therapy.55,57 This trial also reported Clostridium difficile infection in three patients—two patients in the intermittent infusion group and one patient in the continuous infusion group.55,57 No study reported on hematological adverse effects.
Table 9
Antibiotic-related adverse event outcomes for studies addressing Key Question 2.
One RCT and the retrospective cohort study reported rates of resistance or development of resistance during the study periods.55,57,64 The trial prospectively evaluated data (333 serial MICs) for the identified isolates, but the investigators reported that they did not observe any development of resistance during the study period in either group.55,57 The cohort study researchers reported that they observed no antibiotic resistance during the treatment course in either group.64
Table 10 presents the characteristics of the organisms identified for the studies included for KQ 2. The majority of the organisms identified were Gram-negative. Four studies reported on susceptibility data for the organisms isolated.56,57,61,63 Two studies used these MIC data to evaluate pharmacodynamic profiles of the regimens given.56,57
Table 10
Organism characteristics for studies addressing Key Question 2.
In one Nicolau et al. study, the continuous infusion regimen of ceftazidime produced drug serum concentrations that exceeded the MIC breakpoint of 8 mg/L for Pseudomonas aeruginosa for 100 percent of the dosing interval for all patients in the continuous infusion group.58 This means that serum antibiotic concentrations were sufficient to inhibit the growth of ceftazidime-susceptible Pseudomonas aeruginosa for 100 percent of the dosing interval. For patients in the intermittent infusion group, the MIC was exceeded for 100 percent of the dosing interval for organisms with an MIC ≤2 mg/L, an average of 92 percent of the dosing interval for organisms with an MIC ≤4 mg/L, and an average of 82 percent of the dosing interval for organisms with an MIC of 8 mg/L. So, for more susceptible organisms (those with lower MICs), intermittent infusion of ceftazidime provided antibiotic concentrations sufficient to inhibit bacterial growth for more time during the dosing interval than for less susceptible organisms (those with higher MICs).
The Sakka et al. study showed that the intermittent infusion of imipenem (1 g every 8 hours) achieved a probability of target attainment of 88 percent for organisms with an MIC of 2 mg/L, using a target of drug concentration exceeding the MIC for 40 percent of the dosing interval.56 So, for organisms with MIC values of 2 mg/L or less, the intermittent infusion of imipenem had a 88 percent probability of reaching the predefined PD target for drug concentrations sufficient to inhibit bacterial growth for 40 percent of the dosing intervial. The probability of target attainment decreased for organisms with MICs >2 mg/L (less susceptible organisms). In the continuous infusion group, the probability of target attainment was 90 percent for organisms with an MIC of 2 mg/L and 86 percent for organisms with an MIC of 4 mg/L, using the target of 40 percent (drug concentration exceeding the MIC for 40 percent of the dosing interval).
Neither the Nicolau et al. nor the Sakka et al. studies related results of the pharmacodynamics analyses to patient outcomes.
One RCT and one retrospective cohort study reported on rates of resistance or development of resistance during the study periods.55,57,64 The trial prospectively evaluated susceptibility data (333 serial MICs) for the identified isolates,55,57 but the investigators reported that they did not observe any development of resistance during the study period in either group. The cohort study reported that no antibiotic resistance was observed during the treatment course in either group.64
Strength of Evidence
For KQ 2, we graded SOE as insufficient for clinical response, duration of mechanical ventilation, morbidity or mortality, and rates of antibiotic-related adverse events (Table 11). The main reason was the small number of studies with small numbers of patients, which generally resulted in unknown consistency and imprecision. In addition, aggregate risk of bias was medium or high for all outcomes for which we had any evidence.
Table 11
Strength of evidence for comparisons of continuous and intermittent infusion.
Key Question 3. Subgroup Analyses
We found no studies meeting inclusion criteria that answered any questions about the impact of using PK/PD measures or principles on either intermediate or health outcomes or adverse events for subgroups characterized by age, sex, race, ethnicity, renal dysfunction or need for dialysis, severity of illness, type of microorganism, or susceptibility patterns. Consequently, the SOE was insufficient for subgroup issues.
- Results - Pharmacokinetic/Pharmacodynamic Measures for Guiding Antibiotic Treatm...Results - Pharmacokinetic/Pharmacodynamic Measures for Guiding Antibiotic Treatment for Hospital-Acquired Pneumonia
Your browsing activity is empty.
Activity recording is turned off.
See more...