Editorial
June 12, 2024
Resolving the Dilemma on Continuous vs Intermittent β-Lactam Antibiotics in Sepsis
W. Joost Wiersinga, Michiel A. van Agtmael
JAMA. Published online June 12, 2024. doi:10.1001/jama.2024.10168
The Surviving Sepsis Campaign guidelines recommend that adults with sepsis or septic shock should receive prolonged infusion of β-lactams (after an initial bolus) vs conventional intermittent infusion.1 However, this is only a weak recommendation based on moderate quality of evidence. In this issue of JAMA, 2 major new contributions strengthen the evidence for this recommendation.2,3
Up to 40% of patients treated with β-lactam antibiotics in the intensive care unit (ICU) may not achieve antibiotic concentrations above the minimum inhibitory concentration (MIC) during 50% to 100% of the dosing intervals.4 In sepsis, many physiologic changes have pharmacokinetic-pharmacodynamic consequences, including increased cardiac output leading to increased drug clearance, and leaky capillaries necessitating intravenous volume resuscitation leading to an increased volume of distribution. Both of these changes will result in lower antimicrobial plasma concentrations, especially in the early phase before kidney failure sets in. Later in the clinical course, kidney failure or hepatic failure can decrease clearance, leading to increased plasma drug concentrations. β-Lactam administration via prolonged (with an infusion time of 4 hours) or continuous infusion will lead to sustained concentrations throughout the dosing interval, longer time above MIC, and improved bacterial eradication. But does a better pharmacokinetic target mean a better clinical outcome?
Numerous randomized clinical trials (RCTs) over the past 2 decades have addressed this question. One of the early trials, conducted in France and published in 2005, compared continuous infusion vs intermittent dosing of cefepime in critically ill patients.5 No differences in clinical outcomes (mean ICU stay, mortality) were observed between the groups; however, continuous infusion was associated with longer times above the MIC, reflecting more constant bactericidal activity.5 Multiple meta-analyses, including this trial and those performed subsequently, have reported reduced short-term mortality with continuous infusion of β-lactam antibiotics.6-8 The MERCY trial published in 2023 in JAMA found a nonsignificant 2% reduction in absolute 28-day mortality rate with continuous infusion of meropenem.8 However, questions regarding long-term outcomes and the emergence of antimicrobial resistance remain and must be addressed in order to better define the true clinical benefits of prolonged infusions.9
A significant addition to the evidence comes from the Beta-Lactam Infusion Group (BLING) III trial, a well-conducted, open-label, randomized trial in 104 ICUs in 7 countries, comparing 3498 patients receiving continuous infusions with 3533 patients receiving intermittent infusions (with loading dose) of β-lactam antibiotics (piperacillin-tazobactam or meropenem).2 Nearly 30 000 patients were screened for eligibility, of whom 21 840 were excluded, with receipt of antibiotics for more than 24 hours being the primary reason for exclusion. Enrolled patients had a mean age of 59 years and were predominantly male, with no mention of comorbidities. The percentage of patients with septic shock was not specified, but approximately 70% received vasopressors within the 24 hours prior to randomization. The respiratory tract was the most common source of sepsis, followed by intra-abdominal infections, with Escherichia spp, Klebsiella spp, Pseudomonas spp, and methicillin-sensitive Staphylococcus aureus being the most prevalent causative pathogens.
All-cause mortality within 90 days after randomization, the primary outcome of the study, occurred in 24.9% of those receiving continuous infusion of β-lactam antibiotics and in 26.8% of those receiving intermittent infusions. The odds of dying within 90 days was 0.91 for the patients in the continuous group, reflecting a mortality reduction of 1.9%, with a 95% CI ranging from a reduction of 4.9% to an increase in mortality of 1.1%. While lacking statistical significance (P = .08), the authors conclude that the confidence interval around the effect estimate includes the possibility of both no important effect and a clinically important benefit in the use of continuous infusions in this group of patients.
Clinical cure was higher in the continuous compared with the intermittent infusion group (1930/3467 [55.7%] and 1744/3491 [50.0%], respectively; absolute difference, 5.7% [95% CI, 2.4% to 9.1%]). Other secondary outcomes were not statistically different. Only piperacillin-tazobactam and meropenem were studied, and no data were provided for bacterial sensitivity to this empirical treatment.
Abdul-Aziz and colleagues3 have included the results of BLING III in the latest meta-analysis of prolonged vs intermittent infusions of β-lactam antibiotics in critically ill adult patients with sepsis or septic shock. They included 18 RCTs with a total of 9108 patients of which 17 RCTs contributed to the primary outcome of all-cause 90-day mortality. Because of its size, the BLING III trial contributed 77.1% of patients included in this meta-analysis.2,3 The pooled estimated risk ratio (RR) for all-cause 90-day mortality for prolonged infusions of β-lactam antibiotics when compared with intermittent infusions was 0.86 (95% credible interval [CrI], 0.72-0.98), with a 99.1% posterior probability that prolonged infusions reduced 90-day mortality. Similarly, prolonged infusions of β-lactam antibiotics were also associated with reduced risk of ICU mortality (RR, 0.84 [95% CrI, 0.70-0.97]) and increase in clinical cure (RR, 1.16 [95% CrI, 1.07-1.31]). No differences were seen in the predefined secondary outcomes of microbiological cure, adverse events, and duration of ICU stay between both groups.
These results are consistent with previous meta-analyses.6,7,10 However, major individual RCTs have not been able to demonstrate a significant improvement of clinical outcome in sepsis in the ICU with the use of continuous β-lactam antibiotics over intermittent infusions.8,11,12 The effect appears modest, with most trials too small to detect a significant difference. Additionally, one would expect that most benefit is found in patients infected with a microorganism with high MIC (eg, Pseudomonas spp) and augmented creatinine clearance during the early phase of sepsis. Following pharmacokinetic-pharmacodynamic principles, the risk of a sub-MIC antibiotic level is increased in those situations.7 In BLING III, the median serum creatinine concentration was around 170 μmol/L, where low antibiotic levels are not expected.
In the meta-analysis, prespecified subgroup analysis did not show a difference between continuous vs intermittent infusions for all-cause 90-day mortality between meropenem vs piperacillin-tazobactam, culture-positive vs culture-negative infection, kidney replacement therapy vs no kidney replacement therapy, lung infection vs other infections, sepsis vs septic shock, or male vs female participants.3 In contrast to previous reports highlighting the primary advantage of prolonged infusion over intermittent infusion in gram-negative bacterial infections,7,10 this newest meta-analysis found no distinction between gram-negative and gram-positive infections. MIC measurements or estimates (so called epidemiological cutoff values) in the isolated pathogens were not provided. In this subgroup of infections caused by pathogens with high MICs you would expect added value from continuous infusion.
Are there reasons for not implementing prolonged/extended or continuous infusions? Most β-lactams have a half-life of 1 to 2 hours and continuous infusion will improve their pharmacokinetic target. The longer the half-life, the less benefit from continuous infusion. In the case of ceftriaxone, for example, with a half-life of 8 hours, intermittent twice-daily dosing will have a comparable pharmacokinetic effect. Incompatibility of intravenous comedications with the antibiotic makes administration over a second intravenous entry necessary.
Feasibility and cost-effectiveness were not considered in BLING III, but the total daily antibiotic dose is similar and probably less nursing time is needed. One should be aware that some β-lactams have a limited pharmaceutical stability. For meropenem, the stability is about 8 hours at room temperature, so it is not suitable for 24-hour infusion, thus affecting bedside logistics. For patients with sepsis needing ICU care, continuous infusion should preferably start on admission at the emergency department. These changes in workflow should not result in unintended delays and will require time to implement. None of the studies included in the meta-analyses showed an increase in toxicity or increase in development of resistance.2,7,8 In intermittent β-lactam dosing, at the end of the dosing interval, there is a risk of low concentrations at the site of infection. Although acquisition of resistance is multifactorial, low β-lactam levels might increase resistance rates (and failure).13 However, if continuous β-lactam antibiotic infusions are widely implemented in an ICU, monitoring resistance rates before and after implementation will be important.
As the authors of the BLING III study point out in their discussion, although nonsignificant, the observed reduction in mortality of around 2% with the use of a loading dose followed by continuous infusion represents a number needed to treat of 50 patients to prevent 1 death. All secondary outcomes and the findings of the novel meta-analysis point in the same direction. A main challenge for follow-up studies will be to identify subgroups of patients that will benefit most from this treatment option. Studies should explore the role of MIC-based dosing in patients with sepsis and augmented renal clearance (which can easily be measured with an 8-hour creatinine clearance). In other words, in those patients with a high renal clearance and infected with a pathogen with a high MIC, increased exposure by continuous infusion could be convincingly superior.
The authors of the BLING III trial should be applauded for a well-conducted study, with sites across 3 continents with almost 800 local study investigators making it one of the largest RCTs in the field of intensive care medicine.2,14 BLING III is a beautiful and impressive example of true team science that accelerates progress in the field. Despite the statistical nonsignificance of its primary end point, clinical guidelines are likely to use this new landmark study and accompanying meta-analysis to strengthen their recommendation to use continuous β-lactam antibiotics over intermittent dosing in adult patients with sepsis in the ICU.1,15
