Jason Roberts, PhD, discusses insights from the BLING III trial, outlining its study design and primary outcomes. His analysis centers on the clinical benefits of continuous infusions while addressing the complexities of antibiotic administration in critical care. This is part 1 of a 2-part series.
In critical care settings, the optimal method of administering β-lactam antibiotics to septic patients has been a topic of ongoing contention. The international BLING III trial addressed this debate by comparing continuous and intermittent infusions of these antibiotics in 7,031 adult septic patients. The primary analysis revealed no statistically significant difference in 90-day mortality rates between the 2 infusion methods (24.9% vs. 26.8%, respectively; odds ratio (OR), 0.91). Although, the confidence interval suggests a potential for both no significant effect and a beneficial clinical impact with continuous infusions. These findings highlight the complexities in antibiotic administration strategies for critically ill patients.
Researcher Jason Roberts, PhD, started his career in critical care units at the Royal Brisbane Women's Hospital and transitioned to health and medical research over the past 15 years. He specializes in optimizing antimicrobial doses and is currently the director at the University of Queensland Centre for Clinical Research. Since beginning his PhD in 2005, Roberts has been deeply involved in the BLING research program (Beta Lactam Infusion Group), demonstrating his long-term dedication and expertise in this field.
This international, open label randomized clinical trial took place across 104 intensive care units in Australia, Belgium, France, Malaysia, New Zealand, Sweden, and the United Kingdom. Recruitment spanned from March 26, 2018, to January 11, 2023, with follow-up completed by April 12, 2023. Critically ill adults (≥18 years) treated with piperacillin-tazobactam or meropenem for sepsis were included.
What factors led to the decision to conduct an international, open-label trial across multiple countries? How did the diversity of ICUs and patient populations affect the study outcomes and generalizability?
To achieve adequate recruitment across multiple countries, the study organizers opted to conduct the research in an open-label format. This pragmatic decision aimed to involve a broad spectrum of ICUs and meet the study's recruitment needs effectively. The collaboration and support from investigators across different countries were pivotal and highly valued. Throughout BLING 2, strict adherence to specific patient inclusion criteria shaped the study's methodology and influenced its outcomes significantly.
Roberts explains how the diversity of ICUs and patient populations affected the study outcomes and generalizability,
“We deliberately aimed to be very broad with the types of ICUs that would be included from the countries, thinking that that would improve the generalizability of the results. And the patient population was still sepsis, people still had to align with that definition. Even if the individual case mix within the different ICUs could have been different, it's still more likely to be representative of the broader global sepsis population.”
Roberts reflects on the previous BLING I and BLING II study outcomes that influenced the outline of the BLING III trial because they were key factors in building the setting for BLING II to be successful.
BLING 1 and BLING 2 were pivotal randomized controlled trials conducted across multiple centers in Australia and Hong Kong. BLING 1, which involved four ICUs, aimed to assess the feasibility of a double-blind, double-dummy design in evaluating pharmacokinetic outcomes associated with continuous versus intermittent antibiotic infusion. BLING 2 expanded the scope to 26 ICUs, retaining the same design while focusing on ventilator-free days as the primary endpoint and examining clinical cures. These trials laid the groundwork for further investigation into continuous antibiotic infusion among critically ill patients.
“I think that the randomized control trial is very important but equally important is the systematic review and meta-analysis, and they should be both read together. And the cumulative evidence across all of the studies within the systematic review and meta-analysis is very consistent, favoring the use of prolonged infusions. We see the same results in the randomized control trial,” says Roberts.
The primary outcome of 90-day mortality showed a trend towards lower mortality with continuous infusion, although not statistically significant. How should clinicians interpret this finding, especially considering the confidence interval that suggests potential clinical benefit?
“Up until this point, for us to be able to randomize patients in our own intensive care units to continuous infusion, we couldn't have our own belief that intermittent infusion must be inferior because that would be unethical. So we were always very open-minded about this. I think that the nature of the point estimates for the odds ratio in the context of the confidence interval suggests that the likelihood is, is that this is a valuable intervention, which can save lives on certainly from a probabilistic perspective,” according to Roberts' perspectives
Out of 7202 participants randomly assigned, 7031 met the consent criteria for inclusion in the primary analysis (97.6%). Within 90 days, mortality rates were 24.9% (864 out of 3474) for patients receiving continuous infusion compared to 26.8% (939 out of 3507) for those receiving intermittent infusion, resulting in an absolute difference of -1.9% (95% CI, -4.9% to 1.1%) and an odds ratio of 0.91 (95% CI, 0.81 to 1.01; P = .08).
Roberts continues, “The final statistical evidence which comes through is through the posterior probabilities that came through from the systematic review and meta-analysis which said that there's a 99.1% probability that you're giving the drug by prolonged infusion because the systematic review is looking at continuous infusion and extended infusion studies, that there's a 99.1% likelihood that this intervention can reduce sepsis mortality.”
Clinical cure rates were higher in the continuous infusion group (55.7% (1930 out of 3467) compared to the intermittent infusion group (50.0% (1744 out of 3491), with an absolute difference of 5.7%. The primary outcome was all-cause mortality within 90 days after randomization.
