Abstract
OBJECTIVE
The pharmacokinetics of β-lactam antibiotics favor administration via an extended infusion. Although literature to support extended infusion β-lactams exists for adults, few data are available in pediatrics, especially among patients with bacteremia. The purpose of this study was to compare clinical outcomes between extended and standard infusions in children with Gram-negative bacteremia.
METHODS
This retrospective chart analysis included hospitalized patients ages 0 to 18 years who received at least 72 hours of cefepime, meropenem, or piperacillin-tazobactam between January 1, 2013 and July 30, 2021. Clinical outcomes included duration of antibiotic therapy, hospital length of stay, readmission within 30 days, all-cause mortality, time to blood culture clearance, and time to normalization of inflammatory markers.
RESULTS
A total of 124 patients (51 extended infusion, 73 standard infusion) met criteria for evaluation. Duration of antibiotic therapy was shorter in the extended infusion group (6.6 days versus 10.2 days; p = 0.01). There were no differences in hospital length of stay, readmission rates, all-cause mortality, time to normalization of inflammatory markers, or time to blood culture clearance.
CONCLUSIONS
Use of extended infusion β-lactam antibiotics in children with Gram-negative bacteremia was associated with shorter durations of therapy and should be the preferred method of administration when feasible.
Keywords: bacteremia, β-lactam, extended infusion, outcomes, pediatrics
Introduction
Broad-spectrum β-lactam antibiotics are frequently used to treat Gram-negative bacteremia in children. Insufficient drug exposure is associated with poor clinical outcomes, including development of resistant bacteria.1–3 Antimicrobial dose optimization to maximize pharmacokinetics or pharmacodynamics of the drug is a potential way to improve outcomes.4,5 For β-lactam antibiotics in particular this means maximizing the time during which the concentration of the drug exceeds the minimum inhibitory concentration (fT>MIC) of the organism. Extending the duration of infusion increases the fT>MIC.
Many reports of extended infusion dosing strategies are available in adult populations. Studies reporting positive benefits suggest increasing the infusion time improves outcomes (e.g., lower mortality, decreased length of hospital stay, decreased time to defervescence).6–9 There are fewer reports of outcomes using the extended infusion strategy in children, and even fewer among children with confirmed Gram-negative bacteremia.10–12 We previously evaluated children who received extended infusion β-lactam antibiotics for greater than 72 hours for any indication, and identified decreased mortality among critically ill patients.13 However, given the limited number of patients with confirmed bloodstream infections in that cohort, we sought to expand our study population to evaluate if children with bacteremia would specifically benefit from an extended infusion dosing strategy. The objective of this study was to identify differences in outcomes comparing children with confirmed Gram-negative bacteremia who received standard (shorter) versus extended infusion β-lactam antibiotics.
Methods
We identified children younger than 18 years with confirmed Gram-negative bacteremia from January 1, 2013 through July 30, 2021, who received at least 72 hours of cefepime, meropenem, or piperacillin-tazobactam. Only patients who had Gram-negative organisms against which the included antibiotics would be expected to have activity were included. Sex, age, severity of illness markers (e.g., need for mechanical ventilation or vasopressors) blood culture results, white blood cell (WBC) count, C-reactive protein (CRP) values, mortality, readmission rates, and information related to the administered drugs (antibiotic name, dose, frequency, and duration) were collected. Pediatric Risk of Mortality 3 (PRISM-3) scores were calculated 12 hours after admission for patients in the pediatric intensive care unit. PRISM-3 is a widely used scoring system that predicts the risk of death in pediatric intensive care unit patients.14 Time to normalization of WBC and CRP was defined as the difference in days from the first high value to the first normal value following the high value. Data were obtained from the electronic medical record, followed by manual chart review to determine cause of death and reason for readmission. Data relative to inflammatory markers were limited to patients with at least 2 values measured on separate days. Data relative to blood culture clearance were limited to patients who had at least 1 repeat blood culture after the initial positive culture. Mortality was defined as all-cause mortality within 30 days of completing the antibiotic course. Readmission was defined as a readmission within 30 days of completing the antibiotic course, limited to patients who received antibiotics and had a potential infection-related diagnosis. A determination regarding infection-related diagnoses was made based on unanimous agreement among a critical care physician, infectious diseases physician, and infectious diseases pharmacist. Readmissions related to viral infections, planned admissions (e.g., for chemotherapy), or any indication that did not clearly require antibiotics were excluded for analysis. Standard infusions were defined as infusion duration of 30 minutes or less, whereas extended infusions were administered over 3 to 4 hours. Cefepime was dosed as 50 mg/kg intravenously (IV) every 8 hours. Meropenem was dosed as 20 mg/kg IV every 8 hours, unless the patient was being treated for a central nervous system infection, in which case the dose was increased to 40 mg/kg IV every 8 hours. Piperacillin-tazobactam was dosed as 75 mg/kg IV every 6 hours prior to July 1, 2018, and then was updated to 100 mg/kg IV every 8 hours when the extended infusion protocol was implemented. All doses were the same, regardless of infusion type. Proportions of patients with various demographic or medical characteristics who received extended versus standard infusion were compared using χ2 tests or Fisher exact tests. Patients receiving therapy prior to implementation of the extended infusion protocol (July 1, 2018) received standard infusions, whereas patients after implementation of the protocol received extended infusions as the standard of care. Continuous or numeric variables (e.g., age and duration of therapy) were compared by type of infusion using Mann-Whitney tests. Non-parametric tests were used because of the skewed distribution of many of the variables. Logistic regression was performed to evaluate the odds ratio for mortality after adjusting for PRISM-3 risk of mortality. Statistical analysis was performed using IBM SPSS Statistics version 20.0.
Results
A total of 124 children met criteria for study inclusion; 73 received a standard duration of infusion and 51 received an extended duration of infusion. Demographics are described in Table 1. Both groups were similar, except for younger age in the extended group. Severity of illness, determined by need for mechanical ventilation and/or vasopressors, as well as PRISM-3 probability of death scores, were similar between infusion types. The overall risk of death was low in both groups.
Table 1.
Patient Demographic and Clinical Characteristics
| Standard infusion (n = 73) | Extended infusion (n = 51) | p value | |
|---|---|---|---|
| Male, n (%) | 50 (68.5) | 37 (72.5) | 0.627 |
| Age, median (IQR), yr | 5 (2.1–11.5) | 1.7 (1–10.5) | 0.024 |
| Received mechanical ventilation, n (%) | 10 (13.7) | 7 (13.7) | 0.997 |
| Received vasopressors, n (%)* | 9 (12.3) | 11 (21.6) | 0.169 |
| PRISM-3 probability of death, median (IQR)† | (n = 30) 1.41% (0.4–6.3) | (n = 25) 2.2% (0.3–12.5) | 0.417 |
* For example, epinephrine or norepinephrine
† Limited to patients in the pediatric intensive care unit; score calculated 12 hours following admission
There were 127 organisms identified among the 124 children; 3 patients had polymicrobial infection. The most common organisms were Klebsiella pneumoniae (24%), Escherichia coli (19%), Enterobacter cloacea (18%), and Pseudomonas aeruginosa (16%). Although we did not specifically collect resistance patterns for each organism, overall resistance is low at Children's Wisconsin based on our annual antibiograms.
A comparison of clinical outcomes between the 2 infusion strategies is presented in Table 2. There were no differences in hospital length of stay, readmission rates, or mortality. The odds of death within 30 days for patients receiving extended versus standard infusion, when adjusted for PRISM-3 probability of death score, was 0.19 (95% CI, 0.01–4.73; p = 0.313). Times to WBC and CRP normalization were similar. Time to blood culture clearance occurred at 25 hours (median) for both groups. Duration of antibiotic therapy was shorter in the extended infusion group (6.6 days versus 10.2 days; p = 0.01).
Table 2.
Comparison of Clinical Outcomes by Infusion Type *
| Standard Infusion (n = 73) | Extended Infusion (n = 51) | p value | |
|---|---|---|---|
| Hospital length of stay, median (IQR), days | 14.1 (8.2–23.7) | 11 (6.3–19.3) | 0.105 |
| Duration of antibiotic therapy, median (IQR), days | 10.2 (5.6–14.3) | 6.6 (4.3–9.9) | 0.010 |
| Readmission within 30 days, n (%) | 9 (12) | 9 (18) | 0.408 |
| All-cause mortality within 30 days, n (%) | 3 (4.1) | 2 (3.9) | 1.0 |
| Time to WBC count normalization, median (IQR), days | 4.1 (1.1–9.4) | 3.1 (0.9–6) | 0.607 |
| Time to CRP normalization, median (IQR), days | 6.3 (3.5–10.3) | 4.1 (2.5–7.7) | 0.269 |
| Time to blood culture clearance, median (IQR), hr | 24.7 (13.9–34.8) | 24.5 (16.8–39.7) | 0.708 |
CRP, C-reactive protein; WBC, white blood cell
* Standard infusion administered over ≤30 minutes; extended infusion administered over 3 to 4 hours
Discussion
At our institution, we implemented a protocol to administer select β-lactam antibiotics (cefepime, meropenem, and piperacillin-tazobactam) via extended infusion as the standard of care, regardless of indication. We previously reported an association with reduced mortality among critically ill patients who received extended versus standard infusions.13 At that time, however, there were a limited number of patients with confirmed bacteremia. For this evaluation, we expanded the study period to capture a larger number of patients with confirmed Gram-negative bacteremia.
Although many studies report on the benefits of extended infusion in adults, limited data exist in children, especially relative to children with confirmed Gram-negative bacteremia. To our knowledge, only 2 other publications exist that report on extended infusion β-lactam antibiotics in children with Gram-negative bacteremia. Beauchamp et al12 described the use of extended infusion β-lactams in children with confirmed Gram-negative bacteremia. Their study included 67 children who received cefepime for Gram-negative bacteremia and did not identify any differences in outcomes between extended and standard infusions. Combination antibiotic therapy was allowed; more patients in the standard infusion group received concomitant antibiotics. Shabaan et al15 evaluated 102 infants with Gram-negative late-onset sepsis who received meropenem and found a lower mortality rate, improved microbiologic eradication, and higher rates of clinical improvement among children receiving extended versus standard infusion. It is unclear in the methods if concomitant antibiotics were allowed.
We did not identify any difference in outcomes related to hospital length of stay, normalization of inflammatory markers, or time to blood culture clearance. Although not statistically significant, 2 days of decreased time to CRP normalization and 3 days of less hospitalization may be clinically relevant. All patients, regardless of infusion time, cleared their bloodstream infection in about a day. This is not surprising, however, because a significant portion of bacteremia in children, including in our study, is caused by either pyelonephritis or translocation. These patients generally do quite well, so it would be difficult to find a benefit when these patients are properly treated. However, given the presumed pharmacokinetic and pharmacodynamic benefits of extended infusions, difficulty reliably predicting the future clinical trajectory of a patient on initial presentation, and the benefits of reducing variability in nursing care, we continue to support administering β-lactam antibiotics during a prolonged period of time when line time and compatibility issues do not exist.
All-cause mortality was low in our population. There were only 5 deaths that occurred within 30 days of completing antibiotic therapy during the study period. We performed a manual review of all 5 deaths to determine the reported cause of death. Only 1 of the 5 was potentially related to failure of antibiotic therapy; this patient received a standard infusion. The remaining deaths were related to complications of underlying disease or voluntary withholding of life-sustaining measures. Notably, we were not powered to detect a difference in mortality because of the low number of deaths in both groups. Low reported mortality rates in the literature among children with Gram-negative bloodstream infection are consistent with our results.16,17 Therefore, evaluation of mortality in pediatric patients with bloodstream infections may not be an appropriate measure of success.
In our study, durations of therapy were almost 4 days shorter in the extended infusion group. However, we are skeptical this is attributable to receiving extended infusion antibiotics given potential confounders specific to our institution. First, we implemented in-person antibiotic stewardship rounds (“handshake stewardship”) in 2018. As part of those rounds, we provided education regarding durations of therapy for all indications, including treatment of bacteremia. We specifically promoted a study by Yahav et al18 published in 2019 that found 7 days of therapy was non-inferior to 14 days for the treatment of hospitalized adults with uncomplicated Gram-negative bacteremia. Furthermore, we participated in a quality improvement project (November 2018 to November 2020) that sought to reduce unnecessarily long durations of therapy for a select number of disease states, including bacteremia. However, Cutro et al19 also found shorter durations of therapy in adult patients receiving extended versus standard infusion piperacillin-tazobactam for the treatment of presumed sepsis, so our finding is not unreasonable. We did not specifically perform a formal cost analysis; however, shorter durations of therapy could potentially translate into cost savings for patients and hospitals. Additionally, shorter durations could reasonably be expected to result in fewer adverse effects, including development of antibiotic resistance.
One strength of our study is inclusion of only patients with confirmed Gram-negative bacteremia, which we further limited to organisms where the study drug was expected to have activity. Although we had a relatively low number of patients in each group with resulting low power, the study was fairly robust for a pediatric population, where infection due to bacteremia is not common. Furthermore, uptake of administration of these 3 antibiotics via extended infusion is high (> 90%) at our institution.13 Additionally, we performed a manual chart review of patients who died or were readmitted within 30 days of completing the antibiotic course to evaluate if either was related to failure in antibiotic therapy. Although 5 patients died, only 1 could be potentially related to failure of antibiotic therapy. Among the readmissions, after chart review we were able to limit our data analysis to those indications that were potentially infection related, strengthening the meaningfulness of the analysis.
There are limitations to our study. First, this was a retrospective cohort at a single center, with interventions unrelated to infusion type occurring during the study period (e.g., handshake stewardship, practices relative to improving central line–associated bloodstream infection rates). We defined time to normalization of inflammatory markers as the first high value to the first normal value after a high value, although there is no standard for measurement, and others may define differently. Furthermore, we did not evaluate concomitant antibiotics, which could have affected the outcomes, although double coverage of Gram-negative organisms is not routine at our institution. We also attempted to collect data on body temperature values, but the data were incomplete and could not be evaluated. Additionally, we relied on the electronic medical record to categorize patients as having received extended or standard infusions. Future prospective studies may be able to overcome these limitations.
Conclusion
Use of extended infusion β-lactam antibiotics in children with Gram-negative bacteremia was associated with shorter durations of therapy, although with possible confounders present. No differences in other outcomes, including hospital length of stay, readmission rates, or mortality, were identified.
ABBREVIATIONS
- CRP
C-reactive protein
- fT>MIC
free time above minimum inhibitory concentration
- IV
intravenously
- PRISM-3
Pediatric Risk of Mortality 3
- WBC
white blood cell
Funding Statement
Disclosures. The authors declare no financial interest in any product or service mentioned in the manuscript, including grants, equipment, medications, employment, gifts, and honoraria. All authors had full access to the data in the study and take responsibility for the integrity of the data and accuracy of the data analysis.
Footnotes
Disclosures. The authors declare no conflicts.
Ethical Approval and Informed Consent. Given the nature of this study, the project was exempt from institution review board/ethics committee review.
References
- 1.Rybak M. Pharmacodynamics: relation to antimicrobial resistance. Am J Med . 2006;119(6):S37–S44. doi: 10.1016/j.amjmed.2006.04.001. [DOI] [PubMed] [Google Scholar]
- 2.Odenholt I, Gustafsson I, Lowdin E, Cars O. Suboptimal antibiotic dosage as a risk factor for selection of penicillin-resistant Streptococcus pneumonia in vitro kinetic model. Antimicrob Agents Chemother . 2003;47(2):518–523. doi: 10.1128/AAC.47.2.518-523.2003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Gould IM, MacKenzie FM. Symp Ser Soc Appl Microbiol . 31. 2002. Antibiotic exposure as a risk factor for emergence of resistance: the influence of concentration; pp. S78–S84. [PubMed] [Google Scholar]
- 4.Connors KP, Kuti JL, Nicolau DP. Optimizing antibiotic pharmacodynamics for clinical practice. Pharm Anal Acta . 2013;4(3):1–8. [Google Scholar]
- 5.MacVane SH, Kuti JL, Nicolau DP. Prolonging beta-lactam infusion: a review of the rationale and evidence, and guidance for implementation. Int J Antimicrob Agents . 2014;43(2):105–113. doi: 10.1016/j.ijantimicag.2013.10.021. [DOI] [PubMed] [Google Scholar]
- 6.Vardakas K, Voulgaris G, Maliaros A, et al. Prolonged versus short-term intravenous infusion of antipseudomonal β-lactams for patients with sepsis: a systematic review and meta-analysis of randomised trials. Lancet Infect Dis . 2018;18(1):108–120. doi: 10.1016/S1473-3099(17)30615-1. [DOI] [PubMed] [Google Scholar]
- 7.Dow R, Rose W, Fox B, et al. Retrospective study of prolonged versus intermittent infusion piperacillin-tazobactam and meropenem in intensive care unit patients at an academic medical center. Infect Dis Clin Pract . 2011;19(6):413–417. [Google Scholar]
- 8.Przybylski D, Reeves D. Extended versus intermittent infusions of cefepime for the treatment of febrile neutropenia. Infect Dis Clin Pract . 2018;26(3):154–158. [Google Scholar]
- 9.Bauer KA, West JE, O'Brien JM, Goff DA. Extended-infusion cefepime reduces mortality in patients with Pseudomonas aeruginosa infections. Antimicrob Agents Chemother . 2013;57(7):2907–2912. doi: 10.1128/AAC.02365-12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Knoderer CA, Karmire LC, Andicopulos KL, Nichols KR. Extended infusion piperacillin/tazobactam in children. J Pediatr Pharmacol Ther . 2017;22(3):212–217. doi: 10.5863/1551-6776-22.3.212. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Walker MC, Lam WM, Manasco KB. Continuous and extended infusions of beta-lactam antibiotics in the pediatric population. Ann Pharmacother . 2012;46(11):1537–1546. doi: 10.1345/aph.1R216. [DOI] [PubMed] [Google Scholar]
- 12.Beauchamp LC, Nichols KR, Knoderer CA. Outcomes of extended infusion cefepime in pediatric patients. Infect Dis Clin Pract . 2019;27(5):283–287. [Google Scholar]
- 13.Zembles TN, Schortemeyer R, Kuhn EM et al. Extended infusion of beta-lactams is associated with improved outcomes in pediatric patients. J Pediatr Pharmacol Ther . 2021;26(2):187–193. doi: 10.5863/1551-6776-26.2.187. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Pollack MM, Patel KM, Ruttimann UE. PRISM III: an updated pediatric risk of mortality score. Crit Care Med . 1996;24(5):743–752. doi: 10.1097/00003246-199605000-00004. [DOI] [PubMed] [Google Scholar]
- 15.Shabaan A, Nour I, Elsayed Eldegla H et al. Standard versus prolonged infusion of meropenem in neonates with gram-negative late-onset sepsis. Pediatr Infect Dis J . 2017;36(4):358–363. doi: 10.1097/INF.0000000000001445. [DOI] [PubMed] [Google Scholar]
- 16.Al-Hasan MN, Huskins WC, Lahr BD et al. Epidemiology and outcome of Gram-negative bloodstream infection in children: a population based study. Epidemiol Infect . 2011;139(5):791–796. doi: 10.1017/S0950268810001640. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Levy I, Leibovici L, Drucker M et al. A prospective study of Gram-negative bacteremia in children. Pediatr Infect Dis J . 1996;15(2):117–122. doi: 10.1097/00006454-199602000-00006. [DOI] [PubMed] [Google Scholar]
- 18.Yahav D, Franceschini E, Koppel F et al. Seven versus 14 days of antibiotic therapy for uncomplicated Gram-negative bacteremia: a noninferiority randomized trial. Clin Infect Dis . 2019;69(7):1091–1098. doi: 10.1093/cid/ciy1054. [DOI] [PubMed] [Google Scholar]
- 19.Cutro SR, Holzman R, Dubrovskaya Y et al. Extended infusion versus standard infusion piperacillin-tazobactam for sepsis syndromes at a tertiary medical center. Antimicrob Agents Chemother . 2014;58(8):4470–4475. doi: 10.1128/AAC.02759-14. [DOI] [PMC free article] [PubMed] [Google Scholar]