Invited Commentary 

Infectious Diseases

July 2, 2024

Optimizing the Value of β-Lactam Antibiotics Through Extended Infusion

Miranda So

JAMA Netw Open. 2024;7(7):e2418196. doi:10.1001/jamanetworkopen.2024.18196

In the decade since the World Health Organization (WHO) declared antimicrobial resistance (AMR) a global health threat—the so-called postantibiotic era when common infections would be life-threatening and when routine surgcal procedures, transplantation, and cancer therapy cannot be supported due to lack of effective antimicrobials—great strides have been made to mitigate against this scenario.¹ The US government’s National Strategy for Combating Antibiotic-Resistant Bacteria spurred substantial cross-sectorial, interagency collaboration. Furthermore, recognizing the inter-connectivity between the health of the planet, the environment, humans, and animals, the One Health concept is gaining momentum.¹ However, despite efforts from local and international agencies championing for investment to develop new agents, as well as promoting antimicrobial stewardship interventions, enhanced surveillance, and infection control measures, the effect of AMR on the morbidity and mortality of the world’s most at-risk populations continue to demand global efforts.²

Although antimicrobial innovation is critical against AMR, it requires multimillion-dollar investment in research and development, followed by regulatory approval. To ensure new agents are accessible to countries and regions of all income levels, extraordinary schemes have to be in place to maintain incentivization for pharmaceutical research.¹ Therefore, generating new agents involves complex processes which can potentially further exacerbate preexisting global inequalities. Conversely, an equally critical strategy is to make the most of the antimicrobials we have available. Applying knowledge of antimicrobials’ pharmacokinetic and pharmacodynamic (PK/PD) properties to optimize their effectiveness—getting the best bacterial-killing bang for the antibiotic buck—has been advocated by antimicrobial stewardship and infectious diseases clinicians. For β-lactam antibiotics, the PK/PD characteristic is time-dependent killing. This property is optimized by maximizing the percentage of time within a dosing interval the concentration of unbound drug is above the minimum inhibitory concentration (MIC) of the pathogen.³ Modeling data suggest that extended infusion (EI; administering the same dose over 3 hours) can overcome a higher MIC by circumventing the peaks and troughs of antibiotic concentrations associated with short (30-minute) intermittent infusions (IIs).³ There is consensus among experts on how EI can be applied, with PK/PD targets defined for each β-lactam class.³ Recently, antimicrobial susceptibility testing standards have incorporated EI recommendations in the interpretation of antimicrobial susceptibility testing results.⁴ In the 2022 to 2023 Clinical Laboratory Standards Institutes updates, an isolate with MIC of 16 µg/mL is categorized as susceptible dose-dependent (SDD), predicated on the dose of piperacillin-tazobactam being 4.5 grams every 8 hours administered over 4 hours or 4.5 grams every 6 hours over 3 hours.⁴ Intermittent infusion would only be recommended for isolates with MIC of less than or equal 8 µg/mL, which are categorized as susceptible.⁴ As for supporting evidence, multiple systematic reviews and meta-analyses have been conducted to evaluate the benefits of EI. Among hospitalized patients with confirmed gram-negative infections, compared with II, EI was associated with mortality benefits (risk ratio [RR], 0.67 [95% CI, 0.55-0.81]; P = .001; I² = 4.52%) but there was no difference in clinical cure rate, intensive care stay, or hospital length of stay.⁵ Among patients with febrile neutropenia, however, a mortality benefit was not observed (pooled RR, 0.83 [95% CI, 0.47-1.48]).⁶ A common limitation was that the majority of included studies were observational or open-label designs, which may be a source of bias. Beyond mortality, many relevant clinical, safety, and microbiological end points remain underexplored and underrecognized.

Elsewhere in JAMA Network Open, Karaba et al⁷ sought to compare the outcomes of EI with II in patients with gram-negative bloodstream infections. In this retrospective cohort study with 4861 patients admitted to 24 US hospitals between January 1 and December 31, 2019, the primary outcome was mortality at 90 days from blood culture collection.⁷ Secondary outcomes were recurrent infection with the same bacteria, emergence of an antibiotic-resistant organism, and treatment-associated adverse events. All secondary outcomes were censored at 90 days. To address confounding, the EI group and II group underwent 1:3 nearest-neighbor propensity score matching without replacement. The analytical method to evaluate the association between infusion strategy and the outcomes among matched patients was multivariable regression. From the full study population, 352 in the EI group and 1056 in the II group were matched for analysis. EI was associated with lower mortality (adjusted odds ratio [aOR], 0.71 [95% CI, 0.52-0.97]; P = .03), compared with II. Notably, in the stratified analysis, mortality benefit was only observed in those with severe illness and/or infected with an isolate with MIC in the intermediate or SDD range (aOR, 0.06 [95% CI, 0.01-0.66]; P = .02). This benefit was not observed in patients infected with more susceptible isolates.

This study by Karaba et al⁷ supports prior knowledge about EI of β-lactam, which has been taught in clinical education and implemented in many centers in recent years. By improving how β-lactams are administered, EI can overcome isolates with a higher MIC. The study also supports that EI may have an advantage over II in patients with more severe infections. In turn, reliance on non–β-lactam antibiotics, which are often second- or third-line therapy, may be reduced. Importantly, Karaba and colleagues⁷ addressed several gaps in knowledge for antimicrobial stewardship clinicians and health care safety practitioners. First, the study population included those with severe immune compromise, such as patients with bone marrow transplant or solid organ transplant and patients with neutropenia, all of whom are in great need of effective antimicrobials. Second, the authors evaluated treatment-related adverse events which are pertinent balancing measures for quality improvement initiatives. EI was associated with a higher odds of vascular catheter complications compared with II (aOR, 3.14 [95% CI, 1.66-5.96]; P < .001). Hospitals may consider including preventative measures and nursing education in the EI implementation bundle. Third, although underpowered, the investigators explored the association between infusion strategy and treatment-emergent resistance, defined as a recurrent infection caused by an isolate with at least 4-fold the MIC after a 7-day blood culture negative interval. The proportion of patients who experienced this outcome was 7.2% (8 of 111) in II vs in 2.9% (1 of 35) in EI (P = .35), a finding that warrants additional studies.

This study by Karaba et al⁷ has limitations. Confounding by indication is a concern, as patients who are more seriously ill or had an isolate with a higher MIC may be more likely to receive EI. Propensity score matching can address this concern by improving the comparability of the 2 groups according to their probability of exposure to treatment choice, conditional on covariates. However, propensity scores cannot account for unmeasured covariates. Thus, selection bias may still exist as it is not possible to fully surmise the clinician’s decision-making process through patient records. It was unclear if any centers already had standardized policy for EI. Clinicians practicing in centers where EI policy and procedure are in place may prescribe it more readily than those working in centers where EI is an exception. Conversely, patients admitted to centers where the use of EI is uncommon may be more likely to experience EI-related complications involving venous catheters.

Although EI is not novel, Karaba et al⁷ characterized the circumstances in which EI will be advantageous, while exploring important balancing measures. In turn, their data will help patients and clinicians to get more out of the so-called antibiotic buck.