Prolonged infusions of beta-lactam antibiotics

INTRODUCTION — Beta-lactam antibiotics demonstrate a time-dependent effect on bacterial eradication. Prolonged infusions attain the pharmacodynamic efficacy target defined for beta-lactam antibiotics more effectively than short infusions. Thus, a prolonged infusion administration strategy may improve microbiologic and clinical cure, especially when pathogens demonstrate higher minimum inhibitory concentrations (MIC). Prolonged infusion administration strategies for intravenous beta-lactam antibiotics may include either a continuous infusion (over the entire dosing interval) or an extended infusion (over 2 to 4 hours).

This article will discuss beta-lactam pharmacodynamics, clinical evidence available for prolonged infusion strategies, the benefits and risks, and recommendations for dosing of prolonged infusion beta-lactam antibiotics.

PHARMACOLOGIC BACKGROUND — Beta-lactam antibiotics kill sensitive bacteria by inactivating key enzymes involved in cell wall synthesis, termed penicillin-binding proteins. The specific beta-lactam agents have different half-lives, but all demonstrate time-dependent killing [1]. This refers to the phenomenon that the duration that the pathogen is exposed to the beta-lactam drug is the most important determinant of bacterial eradication and clinical response. This is in contrast to other antibiotic classes that exhibit concentration-dependent killing and post-antibiotic effect, such as aminoglycosides.

The duration of exposure is commonly measured as the percentage of the dosing interval that the concentration of free drug remains above the minimum inhibitory concentration (MIC) of the pathogen (%fT >MIC). The maximal bactericidal effect is achieved when the free drug concentration exceeds the pathogen’s MIC by approximately fourfold for 40 to 60 percent of the dosing interval [2]. The optimal targets for %fT >MIC vary depending on the specific drug (table 1). Maximizing the duration of exposure can be accomplished in three possible ways: increasing the dose, shortening the dosing interval, or prolonging the infusion time (figure 1) [3]. Prolonged infusion administration strategies for intravenous beta-lactam antibiotics may include either a continuous infusion (over the entire dosing interval) or an extended infusion (over 2 to 4 hours) as opposed to traditional, intermittent infusion times (over 30 to 60 minutes).

RATIONALE — Although high-quality, randomized controlled trials to support prolonged infusion beta-lactam strategies are not available, the rationale for this practice is in part based on pharmacodynamic principles and evidence of clinical benefit from observational studies and limited trials without evidence of additional toxicity. In addition, this practice is supported by a theoretical benefit of reduced emergence of drug resistance and, in some cases, measurable economic benefits.

Possible clinical benefit — Clinical data suggest that prolonged (extended or continuous) infusions of beta-lactams are at least equally effective as and, in certain circumstances, such as critical illness, more effective than traditional intermittent infusions for gram-negative infections. These studies are discussed in detail elsewhere. (See 'Clinical efficacy' below.)

Pharmacologic advantage

Less susceptible pathogens — Increases in pathogen resistance and rising MICs, as well as the limited number of new antibiotic agents introduced over the past 10 years, have forced clinicians to explore ways to maximize the administration of beta-lactams to achieve favorable clinical outcomes. Pharmacodynamic modeling suggests that patients with infections due to pathogens that demonstrate higher MICs but are still within the susceptible range would benefit from prolonged infusion strategies. (See 'Pharmacologic background' above.)

Predictive models of clinical response based on pharmacodynamic targets can help to identify optimal antibiotic dosing. These pharmacodynamic models allow inputs of different combinations of dose, interval, and infusion time to estimate the likelihood of attaining the targeted %fT >MIC. A predictive modeling approach also allows for tailoring to certain patient population characteristics or local patterns of drug resistance, specifically the local pathogen-antibiotic MIC distribution. Several studies utilizing such models have supported prolonged infusion strategies over traditional intermittent dosing for infections due to Pseudomonas aeruginosa and other gram-negative pathogens with higher MICs [3-5]. Estimates from pharmacodynamic models are often the basis for alternative dosing strategies that are then evaluated clinically. (See 'Clinical efficacy' below.)

Patients with altered pharmacokinetics — Critically ill or young patients, patients with cystic fibrosis, patients with malignancies, or obese patients may have creatinine clearance >120 mL/min. Critical illness can also lead to variable pharmacokinetics, such as augmented drug clearance, altered volumes of distribution, abnormal fluid balance, and/or changes in protein binding [6]. Achieving adequate serum levels of antibiotics in these populations can thus be extremely challenging; higher doses and prolonged infusions may be the best pharmacologic approach in such situations.

The Defining Antibiotic Levels in Intensive Care Unit Patients (DALI) study assessed whether beta-lactam dosing in critically ill patients achieves drug concentrations associated with maximal activity and whether achievement of target concentrations correlate with clinical outcomes [6]. This prospective, multinational, pharmacokinetic point-prevalence study included 248 patients who received various beta-lactam antibiotics for treatment of an infection. The severity of illness in the study population was high (median APACHE II score 18). Thirty-three percent of infected patients received prolonged infusion dosing of beta-lactams. Patients who received prolonged infusion therapy were more likely to achieve the pharmacodynamic target (93 versus 80 percent of patients receiving intermittent infusions). Achieving an antibiotic concentration over the MIC for >50 percent of the dosing interval (>50%fT>MIC) was associated with positive clinical outcomes for bloodstream infections but not with lung and intra-abdominal infections. Another single-center study of 52 intensive care unit (ICU) patients receiving piperacillin-tazobactam for hospital-acquired pneumonia also demonstrated improved target attainment for patients who received extended infusions compared with intermittent infusions, although it did not demonstrate a difference in clinical outcomes [7]. These studies indicate that critically ill patients have highly variable serum drug levels and that prolonged infusions do assist in attaining pharmacodynamic targets.

Safety — Prolonged infusions will maintain higher serum and tissue drug concentrations than intermittent dosing, which may raise concerns about the potential for drug toxicity. However, beta-lactam drugs are well tolerated in general, and prolonged infusion strategies appear to have no more toxicity risk than intermittent dosing. Available data provide limited information about adverse events with prolonged infusion strategies. In studies that have examined adverse events, no significant differences in incidence of adverse reactions were recognized between prolonged and intermittent infusion strategies [8-10].

Renal toxicity has been observed with the combination of vancomycin and piperacillin-tazobactam, although the finding is not reliably replicated and a few studies suggest this observation may not be clinically relevant [11-15]. In one observational study of patients receiving the combination, administration of piperacillin-tazobactam as an extended infusion was not associated with higher rates of kidney injury compared with intermittent infusion, although it was associated with higher vancomycin troughs [16].

Reduced selection for drug resistance — Prolonged infusions provide shorter periods of time when serum and tissue levels of the antibiotic fall lower than the MIC. Thus, there is less bacterial regrowth occurring between doses and more rapid bacterial eradication. Rapid bacterial killing, reduced organism burden, and reduced time in the presence of sub-MIC antibiotic concentrations may result in reduced opportunities for pathogens to acquire new genetic elements or de-repress chromosomal resistance genes (eg, AmpC beta-lactamases). Additionally, prolonged infusion strategies can result in lower overall exposure to the antibiotic (ie, a lower overall daily dose) without a decrease in serum concentrations (eg, piperacillin-tazobactam) [17]. Thus, on a theoretical basis, prolonged infusion strategies may provide less selective pressure than short infusions and avoid the emergence of acquired drug resistance. Evidence of this theoretical benefit is an area for future study.

Cost benefit — Economic benefits of prolonged infusion strategies have been cited in several, single-center, observational studies [4,5,17,18]. As an example, manufacturer’s dosing of piperacillin-tazobactam for patients with normal renal function (30-minute infusion of 3.375 to 4.5 grams every 6 hours) results in a total of 13.5 to 18 grams of drug per day. In contrast, an extended infusion strategy (4-hour infusion of 3.375 grams every 8 hours) results in a total of 10.125 grams of drug per day. The extended infusion strategy saves essentially one dose of piperacillin-tazobactam per patient day which results in drug acquisition cost savings. One study estimated that extended infusion piperacillin-tazobactam employed in a 650-bed hospital could reduce annual acquisition costs by approximately 68,750 to 137,500 US dollars per year [4]. In addition to acquisition costs, estimates of cost savings due to reduced length of stay and reduced complication rates may also be significant. In an analysis that integrated extended infusion cefepime and meropenem into a ventilator-associated pneumonia (VAP) pathway in three ICUs, the particular dosing regimens resulted in higher per day antibiotic acquisition costs but were estimated to reduce post-VAP hospital costs by 40,000 US dollars per patient, largely because of shorter lengths of stay [19,20]. Finally, if the prolonged infusion protocol results in longer dosing intervals (eg, every 8 hours as opposed to every 6 hours), there may be savings realized in nursing labor costs due to one less administration event per patient day.

Ease of administration in the outpatient setting — Continuous infusions are easier to administer than intermittent infusions in the ambulatory setting with dedicated intravenous lines and portable, battery-powered infusion pumps. As an example, patients must access their line and change the cartridge on their pump only once in a 24-hour period, rather than every four hours with intermittent infusions. This is particularly convenient for patients who require prolonged durations of therapy delivered in the ambulatory setting (eg, outpatient parenteral antibiotic therapy for endocarditis or osteomyelitis). Continuous infusions in the outpatient setting are typically used for penicillin, nafcillin, and oxacillin. Many outpatient pumps can also be programmed to allow prolonged infusions of piperacillin-tazobactam and cefepime.

DRAWBACKS — Potential drawbacks to prolonged infusion beta-lactam dosing strategies include several logistical barriers, need for continued intravenous line access, and issues of compatibility with coadministered drugs.

Logistical barriers — Administration of prolonged infusions require utilization of the intravenous pump for longer periods of time, which may be problematic for patients with limited intravenous access or lower levels of nursing care. As an example, patients on a medical ward may need to ambulate to the bathroom or leave the unit for procedures, which can lead to interruptions in the infusion and impaired drug delivery. Additionally, nurses must be aware of the need to flush intravenous line tubing at the end of the infusion to allow for complete drug administration. Prolonged infusions also may result in higher intravenous catheter utilization for additional venous access. Risks of catheter-associated infections and other potential complications must be weighed with benefits of prolonged infusions for individual patients. Central catheters should not be placed for the sole reason of delivering prolonged infusions unless medical providers believe that the benefit to the individual patient outweighs the risks of central catheter placement. (See "Central venous access: Device and site selection in adults", section on 'Factors influencing catheter selection'.)

Compatibility — Administering other intravenous medications through the same intravenous line used for a prolonged infusion of beta-lactams can be a challenge because of compatibility issues. Clinical pharmacists can assist nursing to help resolve medication scheduling and compatibility issues; however, if shifting medication administration times does not alleviate compatibility issues, reversion to intermittent infusion may be appropriate. Standard textbooks for compatibility may be used as a reference [21,22].

A common compatibility challenge occurs with piperacillin-tazobactam and vancomycin coadministration because compatibility is dependent on the vancomycin concentration. Of note, branded piperacillin-tazobactam (Zosyn) is formulated with EDTA, which allows for expanded compatibility with gentamicin, amikacin, and Lactated Ringer's solution. The physical compatibility of vancomycin has been evaluated with both branded and generic piperacillin-tazobactam (which does not contain EDTA) in concentrations typical for prolonged infusions [23,24]. Vancomycin concentrations of less than or equal to 7 mg/mL were compatible using simulated Y-site administration with piperacillin-tazobactam 33.75 mg/mL up to 90 mg/mL. Individual institutions have different practices for compounded concentrations of vancomycin and variable piperacillin-tazobactam products. Clinicians, nurses, and pharmacists should evaluate compatibility based on available products prior to attempting to use Y-site infusions of these drugs together.

Stability — If drugs are administered over a prolonged period of time, they must be stable over that time, and so not all drugs are appropriate for prolonged infusions. Stability of beta-lactams can be influenced by the type of intravenous fluid used to reconstitute the drug, the concentration of the final solution, and the storage temperature. Some beta-lactam agents, such as carbapenems, are not stable at room temperature for long durations. The product information for branded meropenem (Merrem) notes that meropenem prepared for infusion in normal saline is stable for 1 hour at room temperature and up to 15 hours refrigerated [25]. However, studies examining stability have shown that meropenem prepared for infusion is stable for up to 12 hours at room temperature and up to 24 hours refrigerated (at concentrations up to 20 mg/mL), values that are consistent with stability documentation on generic product information [26-29]. Ampicillin and ampicillin/sulbactam are often limited to intermittent infusions because of their short half-lives; however, individual institutions may have extended stability data to allow for safe administration of continuous ampicillin and ampicillin/sulbactam, particularly for outpatient home infusion therapy [30].

CLINICAL EFFICACY — Clinical data supporting prolonged infusion beta-lactam antibiotic administration resides largely in observational or non-randomized prospective studies, predominantly in adults. A clinical benefit of prolonged infusion strategies has not been definitively demonstrated in randomized controlled trials. Possible reasons that existing randomized controlled trials have not yet demonstrated benefit include small sample size and low study quality. Other limitations have been the inclusion of heterogeneous patient populations, including patients infected with highly susceptible pathogens and low severity of illness, in whom traditional dosing strategies may not need to be improved upon. Finally, there have been variable prolonged and intermittent dosing strategies used in the available studies which make compiling data in systematic reviews and meta-analyses problematic. Well-designed, adequately powered randomized controlled trials are needed to provide definitive evidence that prolonged infusion beta-lactam antibiotics provide clinical benefit.

Adult populations

Pooled data — The totality of the pooled data suggests that prolonged (extended or continuous) infusions of beta-lactams are at least equally effective and, in most studies that combine data from multiple investigations, more effective than traditional intermittent infusions for gram-negative bacterial infections and in adults with sepsis [8,9,31-35]. In observational data, the greatest benefit of prolonged infusions has been primarily in critically ill patients with higher severity of illness and in patients with P. aeruginosa infection [3-5,8,36]. However, the data overall are limited by small samples sizes, heterogeneity of patient populations and dosing strategies, and other methodologic flaws. (See 'Clinical efficacy' above.)

In a meta-analysis of 29 studies, including 18 randomized controlled trials, comparing infusions strategies of piperacillin-tazobactam, cephalosporins, and carbapenems, pooled data demonstrated lower mortality (relative risk [RR] 0.66, 95% CI 0.53-0.83) with extended or continuous compared with intermittent infusions [9]. Similarly, a systematic review and meta-analysis of six studies evaluating carbapenems and/or piperacillin-tazobactam for moderate to severe gram-negative bacillary infections found a lower mortality risk with prolonged compared with intermittent infusions (RR 0.59, 95% CI 0.41-0.83) [8]. Randomized trials have had mixed results, but they demonstrate at least equivalent efficacy of continuous infusions [31,32,37,38]. In a trial from 25 intensive care units (ICUs) across Australia, New Zealand, and Hong Kong, 443 patients with severe sepsis were randomly assigned to receive continuous versus intermittent infusion of the selected beta-lactam (piperacillin-tazobactam, ticarcillin-clavulanate, or meropenem) [37]. There were no differences in ICU-free days at 28 days (18 versus 20 days) or clinical cure rates at 14 days (52 versus 50 percent) with continuous versus intermittent infusion. Approximately one quarter of each group was receiving renal replacement therapy, which may have attenuated the difference in drug levels achieved with the two-drug administration strategies. In contrast, in a trial of 140 patients with severe sepsis in two ICUs in Malaysia, none of whom were on renal replacement therapy, clinical cure rates at 14 days were higher among the patients randomly assigned to continuous versus intermittent therapy with piperacillin-tazobactam, cefepime, or meropenem (56 versus 34 percent) [38]. There were no differences in ICU-free days at 28 days (20 versus 17 days). A separate meta-analysis of adults with sepsis included 22 randomized controlled trials and 1876 patients from a variety of countries. This review reported a significant association with lower all-cause mortality among patients who received prolonged beta-lactam infusions (RR 0.70, 95% CI 0.56-0.87) [36]. Trials published after 2015 have generally reported more favorable effects on mortality than earlier trials [39].

Studies of individual beta-lactam agents are discussed below. Subgroup analysis in two meta-analyses discussed above suggested that the prolonged infusions of penicillins, including piperacillin-tazobactam, were associated with a mortality benefit; one of two suggested significant mortality benefit with carbapenems. Neither meta-analysis showed significantly improved mortality with prolonged infusions of cephalosporins, mostly likely due to the smaller sample [9,36]. Nevertheless, outcomes appear at least comparable between prolonged and intermittent infusions of cephalosporins.

Piperacillin-tazobactam — Several, mainly observational studies have demonstrated the efficacy of extended infusions of piperacillin-tazobactam for gram-negative bacterial infections and have suggested that they are associated with at least equivalent, if not superior, outcomes compared with standard, intermittent infusions. In a systematic review and meta-analysis of five randomized trials and nine observational studies, prolonged infusion of piperacillin-tazobactam was associated with a higher clinical cure rate (nine studies, OR 1.88, 95% CI 1.29-2.73) and a lower mortality rate (11 studies, OR 0.67, 95% CI 0.50-0.89) compared with intermittent infusion [40]. However, heterogeneity between comparison groups and limitations inherent in observational studies decrease confidence in the findings of many of these studies.

Some observational studies suggested lower mortality with extended infusion of piperacillin-tazobactam [4,41,42]. However in some cases, baseline differences compared with the intermittent infusion cohorts suggested that confounding variables could have affected the observation [4,41,42]. As examples:

●In a single center, retrospective study of critically ill adult patients with P. aeruginosa infection, 102 patients who were treated with an extended infusion of piperacillin-tazobactam (3.375 grams infused over 4 hours every 8 hours) were compared to a historical group of 92 patients who received intermittent infusions (3.375 grams infused over 30 minutes every 6 hours) [4]. Among patients with a greater severity of illness (APACHE II scores ≥17) at presentation, the extended infusion strategy was associated with a lower 14-day mortality (12 versus 32 percent with intermittent infusions) and shorter length of stay (21 versus 38 days). Extended infusion was not associated with any difference in outcomes for patients with lesser severity of illness. Of note, the comparator, intermittent infusion group was given a lower dose than typically recommended for treatment of P. aeruginosa infections.

●In the Retrospective Cohort of Extended-Infusion Piperacillin-Tazobactam (RECEIPT) multicenter study of adult patients with gram-negative bacterial infections, 186 patients receiving extended infusion piperacillin-tazobactam were compared with 173 patients receiving intermittent infusions of comparator beta-lactam antibiotics (ie, cefepime, ceftazidime, imipenem, meropenem, doripenem, or piperacillin-tazobactam) [41]. Multivariate analysis demonstrated reduced odds of mortality in the extended infusion group. However, there were important baseline differences between the groups. Although severity of illness scores, hospital and ICU lengths of stay, and antibiotic durations were comparable, fewer patients in the extended-infusion group were infected with P. aeruginosa, received concomitant intravenous aminoglycosides, or had positive respiratory cultures.

In contrast, some observational studies that have not suggested a mortality benefit may have evaluated patient populations that would be less affected by optimizing antibiotic levels [43-45]. As examples:

●A cohort study of 129 patients with gram-negative bacillary infections suggested equivalent outcomes but failed to demonstrate improved clinical outcomes with extended infusion piperacillin-tazobactam compared with intermittent infusion (4.5 grams every 6 hour dosing) [43]. Patients in this study were infected with pathogens with low MICs (ie, more susceptible pathogens) and low severity of illness scores. Only a minority had P. aeruginosa infections. These factors, along with the small study size, may explain its failure to demonstrate a beneficial effect [44].

●A retrospective study of 553 ICU patients who received piperacillin-tazobactam for ≥72 hours reported that extended infusion (4.5 gram loading dose followed by 3.375 grams every eight-hour dosing) was associated with a comparable mortality rate but a shorter length of hospital stay compared with intermittent infusion [45]. However, P. aeruginosa was isolated from approximately 5 percent of patients overall, and the majority of study patients (>65 percent) did not have microbiologically confirmed infections.

Administration of piperacillin-tazobactam as a continuous infusion over 24 hours also has similar pharmacologic and clinical outcomes to intermittent infusion comparators as demonstrated by small, single center, randomized studies [17,46,47]. As an example, an open-label, randomized study in a US community hospital evaluated 262 patients with complicated intra-abdominal infections randomized to receive continuous infusion piperacillin-tazobactam (13.5 grams over 24 hours) or intermittent infusion (3.375 grams over 30 minutes every 6 hours) [46]. There were no differences in clinical and microbiologic cure between the two groups.

Carbapenems — Extended or continuous infusions of carbapenems are associated with similar mortality rates compared with traditional intermittent infusions, but may result in other benefits, as suggested by limited data [48]. In a single-center randomized, open-label trial in 240 adult ICU patients, administration of meropenem by continuous infusion resulted in similar mortality (16 percent) and clinical cure rates that were not statistically different (83 versus 75 percent) compared with intermittent infusion, but microbiologic success rates (90 versus 78 percent) were higher, and ICU stays and duration of therapy were shorter with continuous infusion [49].

Observational data suggest generally similar findings [8,50-55]. As an example, a retrospective study of patients with hematologic malignancies and febrile neutropenia compared outcomes between 76 patients who received extended infusion meropenem (1 gram over 4 hours every 8 hours) with 88 who received conventional intermittent infusion meropenem (1 gram over 30 minutes every 8 hours) [56]. Extended infusion was associated with greater day 5 treatment success (defined as fever resolution, improvement in infectious symptoms, absence of bacteremia, and no need for additional antibiotics), but there were no differences in length of stay and 100-day mortality between the two infusion strategies.

A multicenter, double-blind, randomized study performed in 2022 compared outcomes of 607 ICU patients with sepsis (most of whom had septic shock and required mechanical ventilation) who received either continuous or intermittent infusion meropenem [48]. Continuous infusion was not associated with a difference in the composite outcome of all-cause mortality and the emergence of drug-resistant bacteria at day 28. Antibiotic-free days, days free from the ICU, and 90-day mortality were also not different between groups. Subgroup analyses performed for hypothesis generation also did not detect significant effects among patients with pathogens with a high MIC to meropenem. The study population was from Russia, Croatia, Italy, and Kazakhstan and had high baseline rates of carbapenem resistance (35 percent in continuous infusion group versus 30 percent in the intermittent infusion group). The emergence of drug resistance was common in both groups (24 percent continuous versus 25 percent intermittent). Thus, the choice of meropenem was not optimal for several participants, and findings may not be generalizable to populations with lower baseline resistance. Most participants had respiratory infections; only 10 percent had confirmed bacteremia. Although large, the trial may have been underpowered to detect a benefit in patients with invasive infection (eg, gram-negative bacteremia). Importantly, this large trial did not detect differences in rates of toxicity or harm. Overall, this study was unable to prove that continuous infusion of meropenem was more effective than intermittent infusion but was limited in applicability. Although not statistically significant, the overall direction of effect was in favor of continuous infusion. Therefore, addition of this trial to prior pooled data may or may not alter the results of future meta-analyses studying the effects of prolonged infusions of beta lactams. In the meantime, we continue to favor extended infusion meropenem given its lack of harm and possibility of benefit.

Cephalosporins — Extended continuous infusions of cephalosporins are associated with similar outcomes as traditional intermittent infusions. A meta-analysis of 10 randomized trials and 1 non-randomized trial did not demonstrate a difference in mortality or clinical cure when comparing third- and fourth-generation cephalosporins administered via extended or continuous infusion versus intermittent infusion [57]. The total daily dose of antibiotics in the extended or continuous infusion arms were substantially lower than in the short infusion groups in many of the included studies, yet clinical outcomes were not different.

A subsequent retrospective study compared outcomes in patients with positive respiratory or blood cultures with a gram-negative organism before and after a hospital-wide change from intermittent to extended infusion of cefepime (2 grams over 30 min to 2 grams over 4 hours, each every 8 to 24 hours) [5]. Overall mortality, lengths of stay, and hospital costs were similar between the groups. However, in a subgroup analysis of 87 patients with infections due to P. aeruginosa, mortality was lower among patients who received prolonged infusion cefepime (3 versus 20 percent with intermittent infusion) as was ICU length of stay (8 versus 19 days).

Intraoperative continuous infusion cefazolin has been proposed to maintain optimal concentrations during surgeries requiring the use of cardiopulmonary bypass. A retrospective cohort study compared outcomes between 284 patients who received cefazolin administered as a weight-based intermittent dose every two hours and 232 patients who received cefazolin administered as a 1 gram per hour continuous infusion during cardiac surgery [58]. All patients received a weight-based loading dose prior to incision. Continuous infusion was associated with a reduction in superficial surgical site infections (0.4 percent versus 2.8 percent in the intermittent infusion group); however, the study was underpowered to detect a significant difference in overall SSIs.

Novel beta-lactams (cefiderocol, advanced beta-lactam combinations) — Most of the advanced beta-lactams that target multidrug-resistant gram-negative pathogens were evaluated with and are administered using prolonged-infusion strategies. As examples, cefiderocol and meropenem-vaborbactam are administered with extended infusions over three hours [59,60]. In contrast, the manufacturer specifies intermittent dosing with 30-minute infusion times for imipenem-cilastatin-relebactam [61]. This difference is likely related, in part, to imipenem-cilastatin-relebactam stability limitations of two hours at room temperature; available pharmacokinetic models also suggest that relebactam activity best followed a concentration-time curve (AUC/MIC) and thus would not be affected by infusion duration [62].

Pharmacodynamic data support the use of prolonged infusion ceftolozane-tazobactam [63,64], and one small observational study of patients with P. aeruginosa infections suggested that extended infusion of ceftolozane-tazobactam was associated with a lower likelihood of treatment-emergent resistance to ceftolozane-tazobactam [65]. The manufacturer recommends that ceftazidime-avibactam be infused over two hours; population pharmacokinetic models show that two-hour infusions have a higher likelihood of target attainment than a 30-minute intermittent infusion [66,67]. However, observational data on patients with Klebsiella pneumoniae carbapenemase (KPC)-producing K. pneumoniae infections suggest that a three-hour infusion is associated with a mortality benefit [68]. As a result, some institutions use a three- rather than two-hour infusion of ceftazidime-avibactam.

Other beta-lactams — The efficacy of continuous versus intermittent infusion oxacillin has also been assessed in a single-center, retrospective cohort study [69]. Patients with methicillin-susceptible Staphylococcus aureus endocarditis received either intermittent infusion oxacillin (2 grams over 30 minutes every 4 hours; n = 29) or continuous infusion oxacillin (12 grams continuously over 24 hours; n = 78). The primary outcome was microbiologic cure at 30 days, which was defined as no positive cultures within 30 days of the end of treatment. Clinical cure was not evaluated. Continuous infusion was associated with greater microbiologic cure, (94 versus 79 percent with intermittent infusion), but 30-day mortality rates, lengths of stay, time to defervescence, and clearance of bacteremia were not different.

Continuous infusion administration of ceftazidime and aztreonam, a monobactam with similar pharmacokinetic and microbiologic activity to ceftazidime, has been evaluated in small series using various total daily doses, mainly in patients with cystic fibrosis [70-75]. We do not administer ceftazidime as extended infusion, as pharmacokinetic/pharmacodynamics studies to inform the optimal dosing and schedule are limited. However, some institutions may employ extended infusion administration of ceftazidime using institution-specific dosing protocols.

We administer aztreonam over three hours when used in combination therapy with ceftazidime/avibactam for the treatment of antimicrobial resistant infections, such as for Stenotrophomonas or metallo-beta-lactamase producing carbapenem-resistant Enterobacterales [76]. Similarly, the preferred treatment of carbapenem-resistant Acinetobacter infections includes high dose ampicillin-sulbactam with a prolonged infusion strategy [76]. We recommend consultation with an infectious diseases specialist for treatment of multidrug-resistant pathogens.

Pediatric populations — Prolonged infusion beta lactam studies in children are very limited but suggest at least similar, if not improved, clinical outcomes compared with traditional dosing.

A 2012 systematic review located one small randomized controlled clinical trial, five pharmacokinetic studies, two pharmacodynamic studies using Monte Carlo simulation, one case series, and seven case reports [77]. Studied agents included ceftazidime, meropenem, aztreonam, and various penicillins, and no definitive conclusions could be made. The single randomized controlled trial included in the review evaluated continuous infusion ceftazidime in 14 children with cystic fibrosis and P. aeruginosa lung infection and demonstrated similar clinical outcomes to intermittent infusions [78]. In contrast, in a subsequent single-center randomized trial of 102 neonates with late-onset sepsis due to gram-negative bacteremia or meningitis in Egypt, extended infusion of meropenem (over four hours) resulted in lower neonatal mortality rates (14 versus 31 percent) compared with conventional dosing (over 30 minutes) [79]. Seven-day bacterial eradication and clinical improvement rates as well as duration of respiratory support were also more favorable with extended-infusion dosing. This survival benefit has not been observed in other studies. A retrospective chart analysis of children who received extended versus intermittent infusion piperacillin-tazobactam, cefepime, or meropenem found that bone marrow transplant and critical care patients who received extended infusions showed fewer readmissions and lower all-cause mortality [80]

Descriptive implementation studies of hospital-wide protocols have demonstrated that extended infusion strategies are feasible in pediatric populations; however these studies did not provide analysis of clinical outcomes [80-82].

POTENTIAL INDICATIONS — Indications for the use of prolonged infusion strategies with beta-lactams are not established for most routine scenarios. However, prolonged infusion approaches are included specifically in guidance for treatment of resistant gram-negative pathogens [76]. Pharmacodynamic modeling suggests that the situations in which prolonged infusion strategies would offer a clinical benefit over traditional intermittent strategies are infections with organisms that have a high, but still susceptible-range MIC to the chosen beta-lactam. Limited data from clinical studies suggest that patients with a high severity of illness or with P. aeruginosa infection are more likely to benefit from prolonged infusion strategies. (See 'Pharmacologic advantage' above and 'Clinical efficacy' above.)

Thus, populations with the most potential clinical benefit from prolonged infusion strategies include patients with an elevated risk of drug-resistant pathogens and/or patients with severe infections who may have altered pharmacokinetics. These include patients with:

●Structural lung disease (including cystic fibrosis)

●Frequent healthcare exposures

●Prior repeated antibiotic exposures

●Critical illness with severe infection (including central nervous system infections [83,84], necrotizing fasciitis, neutropenic fever [34,85,86], and burns)

●Infections due to pathogens with high intrinsic resistance and predilection for developing acquired resistance during therapy (eg, P. aeruginosa, Burkholderia cepacia, Acinetobacter baumannii)

●Infections that require advanced beta-lactam agents that target multidrug-resistant pathogens (eg, ampicillin-sulbactam, aztreonam, ceftazidime-avibactam, meropenem-vaborbactam, ceftolozane-tazobactam, cefiderocol) [76]

When piperacillin-tazobactam, meropenem, imipenem, cefepime, or ceftolozane-tazobactam is chosen for treatment in such patients, we suggest a prolonged infusion dosing strategy. Ceftazidime-avibactam, cefiderocol, and meropenem-vaborbactam are typically administered as prolonged infusions. In particular, we favor prolonged infusions for critically ill patients with gram-negative bacterial infections, for patients with infections due to gram-negative bacteria that have elevated but susceptible minimum inhibitory concentrations (MIC) to the chosen agent, and for patients with multidrug-resistant gram-negative bacterial infections. Our practices match those of expert guidelines for treatment of multidrug-resistant gram-negative bacterial infections [76].

In some cases, logistical issues can outweigh the potential benefit of prolonged infusions. Thus, the decision to use a prolonged infusion dosing strategy in individual patients must also take into account whether intravenous access is limited (since a prolonged infusion would utilize an intravenous catheter for a longer time), whether other administered drugs are compatible for co-infusion with the antibiotic, and whether nursing staff can accommodate the potential extra care with a prolonged infusion (see 'Drawbacks' above). Clinicians are encouraged to engage local pharmacy experts and nursing staff to discuss feasibility and compatibility challenges for individual patients. Consultation with an expert in infectious diseases is also valuable when dealing with infections involving pathogens with limited antibiotic susceptibility.

Some hospitals have implemented policies to encourage the use of prolonged infusions even in patients with low severity of illness, a policy that can have an economic benefit. (See 'Institutional implementation' below.)

ADMINISTRATION

Dosing — Optimal dosing for prolonged infusion of beta-lactams has not been well established. Our recommendations on dosing for adult patients with adequate intravenous access (table 2) are based on data from pharmacodynamic models [1,5,63,67,69,87-90]. However, dosing should be reviewed on a case-by-case basis, as such models are based on minimum inhibitory concentration (MIC) distribution assumptions that may not match local patterns. Beta-lactams are also dose-adjusted for renal impairment.

We limit use of prolonged infusion carbapenems to three-hour infusions and to the indications discussed above due to impaired stability of imipenem and meropenem at room temperature; although data are emerging to suggest that stability may be adequate, it is uncertain whether infusions longer than three hours (eg, continuous infusions) offer any clinical or practical benefit. (See 'Potential indications' above.)

Penicillins (eg, oxacillin, nafcillin) that are stable at room temperature for greater than 24 hours can be used in the outpatient setting with continuous infusion pumps. These agents are dosed by summing the total daily dose and then extending the infusion over 24 hours. As an example, nafcillin dosed intermittently at 2 grams every 4 hours with 30-minute infusions would be continuously infused as 12 grams over 24 hours.

Loading doses — Loading doses are sometimes used for patients with sepsis to try to achieve therapeutic drug levels more rapidly and are recommended in this setting by the Surviving Sepsis Campaign Guidelines [91]. For piperacillin-tazobactam, a 3.375 g to 4.5 g loading dose can be given over 0.5 hours. This approach may be particularly useful in institutions with higher P. aeruginosa MIC distributions to achieve faster target attainment. The optimal time to start the subsequent extended infusion is 4 hours later in patients with CrCl >20 mL/min and 8 hours later in patients with CrCl <20 mL/min. While data from various studies are mixed, a systematic review and meta-regression found increased clinical cure rates in patients who received loading doses [88,92,93].

Drug monitoring — We do not routinely monitor drug levels in patients receiving prolonged beta-lactam infusions. Although therapeutic drug monitoring of beta-lactams may help better achieve target drug levels in patients receiving nontraditional prolonged dosing regimens to maximize clinical efficacy, the clinical utility of these tests in this setting has not been proven [94]. Turnaround time and cost are further considerations in assessing their value.

INSTITUTIONAL IMPLEMENTATION — The 2016 Infectious Diseases Society of America (IDSA) guidelines for implementation of antimicrobial stewardship programs make a weak recommendation for use of alternative dosing strategies for broad-spectrum beta-lactams among hospitalized patients in order to decrease costs [95]. Institutional policies that promote prolonged infusion strategies for empiric use of beta-lactams have been successfully employed at several medical centers [1,3,5,96].

Local hospital- or ICU-specific pathogen susceptibility (and minimum inhibitory concentration [MIC] distributions) should be considered when determining whether to incorporate prolonged infusion strategies into institution-specific guidelines for clinical syndromes and empiric therapies [3]. The clinical microbiology laboratory plays an important role in both reporting pathogen MIC values in a meaningful way for clinicians to interpret them, as well as monitoring for shifts in institutional MIC distributions. Laboratories should assist clinicians in identifying high-risk scenarios of elevated, but susceptible, MIC pathogens to target for optimized dosing strategies [94]. (See "Overview of antibacterial susceptibility testing", section on 'Interpretation of results'.)

Acute care hospitals or specific hospital units that frequently encounter drug-resistant gram-negative pathogens may be more apt to adopt prolonged infusion strategies as an antimicrobial stewardship initiative. Extended infusion piperacillin-tazobactam is the beta-lactam most commonly chosen for institutional policies due to high utilization of this drug and the acquisition cost savings associated with the switch to extended infusion.

Pharmacodynamic studies employing Monte Carlo simulation tailored to institutional MIC distributions allows for selection of the dosing strategies most likely to benefit the selected patient populations and is the preferred approach when planning an institutional dosing strategy [3]. If the approach is only to target certain hospital units, dosing can be challenging when transfers between units require adjustment of dosing strategies for a single patient.

Furthermore, careful weighing of the risks outlined above to the estimated benefits and plans to address the logistical implementation challenges must be addressed at the institutional level [96].

Several examples in the medical literature support the feasibility of facility- or unit-wide policies. Single-center, quasi-experimental studies of institutional protocols for extended infusion piperacillin-tazobactam have demonstrated similar clinical outcomes as with intermittent infusions, reductions in drug acquisition costs (13 to 29 percent), and reductions in total amount of drug received [97-99]. One study in two intensive care units (ICUs) in a community hospital also observed reduced lengths of ICU and overall hospital stay [97].

SUMMARY AND RECOMMENDATIONS

●Rationale – Beta-lactam antibiotics demonstrate a time-dependent effect on bacterial eradication. Infusing the antibiotic over a prolonged period of time (eg, over four hours at each dose or even continuously) can more effectively attain pharmacodynamic target levels compared with short, intermittent infusions. This pharmacologic advantage offers a potential clinical benefit to patients with altered pharmacokinetics or with infections due to less susceptible pathogens. (See 'Pharmacologic background' above and 'Pharmacologic advantage' above.)

●Clinical efficacy – Clinical data suggest that prolonged infusions of beta-lactams are at least equally effective as traditional intermittent infusions for gram-negative bacterial infections. Observational studies and meta-analyses of randomized trials suggest that prolonged infusions in critically ill patients are associated with a mortality benefit compared with intermittent infusions. (See 'Clinical efficacy' above.)

●Other benefits – Additional potential benefits of prolonged infusion strategies include cost savings, ease of administration in ambulatory settings, and a theoretical reduction in risk of acquired drug resistance. (See 'Rationale' above.)

●Drawbacks – Potential drawbacks to prolonged infusion beta-lactam dosing strategies include several logistical barriers, such as need for intravenous line access, drug stability at room temperature, and compatibility with coadministered intravenous drugs. (See 'Drawbacks' above.)

●Indications – Indications for the use of prolonged infusion strategies with beta-lactams are not well established for most routine scenarios, but prolonged infusions are specifically recommended for treatment of drug-resistant pathogens. The pharmacologic and clinical data indicate that patients who have an elevated risk of drug-resistant pathogens or who are critically ill in the setting of a severe infection are most likely to benefit. When piperacillin-tazobactam, meropenem, imipenem, cefepime, or ceftolozane-tazobactam is chosen for treatment in such patients, we suggest a prolonged infusion dosing strategy (Grade 2C). Ceftazidime-avibactam, meropenem-vaborbactam, and cefiderocol are typically administered as prolonged infusions. In particular, we favor prolonged infusion dosing for critically ill patients with gram-negative bacterial infections, for patients with infections due to gram-negative bacteria that have elevated but susceptible minimum inhibitory concentrations (MIC) to the chosen agent, and for patients with multidrug-resistant gram-negative bacterial infections. The decision to use this dosing strategy should also take into account logistical issues such as staffing or intravenous access availability. (See 'Potential indications' above.)

●Dosing – Optimal dosing for prolonged infusion of beta-lactams has not been established. Our preferred dosing for adult patients with adequate intravenous access (table 2) is based on data from pharmacodynamic models. Beta-lactams are dose-adjusted for renal impairment. (See 'Dosing' above.)