Ceftriaxone 1 g Versus 2 g Daily for the Treatment of Enterobacterales Bacteremia: A Retrospective Cohort Study

Abstract

Background: Ceftriaxone is a commonly used antibiotic for the treatment of susceptible Enterobacterales infections. There is currently limited clinical data on the optimal dose of ceftriaxone for Enterobacterales bacteremia. Objectives: To evaluate the rate of clinical failure of ceftriaxone 1 g versus 2 g daily in patients with Enterobacterales bacteremia. Methods: This was a retrospective cohort study of patients admitted to any of the 3 New York University Hospitals: Long Island, Tisch, or Brooklyn, with ceftriaxone-susceptible Enterobacterales bacteremia, receiving ceftriaxone 1 or 2 g daily from October 2019 to September 2020. The primary outcome was 90-day rate of clinical failure. Clinical failure was defined as escalation of therapy, relapse of infection, or all-cause mortality. Results: A total of 124 patients, 58% in the 1-g group and 42% in the 2-g group, were included. There was no statistically significant difference found in the primary outcome. The 90-day rate of clinical failure was 16.7% versus 9.6%, P = 0.260. There were no statistically significant secondary outcomes, although infection relapse rates at 90 days were numerically greater in the 1-g group (11.1% vs 1.9%, P = 0.078). Hypoalbuminemia was the only variable associated with an increased risk of clinical failure (odds ratio = 4.03; 95% confidence interval [CI] = 1.12-14.50, P = 0.033). Conclusion: In our exploratory findings, there was no statistically significant difference with the 90-day rate of clinical failure between ceftriaxone 1 g versus 2 g daily, although there was a numeric trend toward an increased rate of infection relapse within the 1-g group. Hypoalbuminemia was associated with an increased risk of clinical failure. Prospective studies are warranted to confirm these findings.

Introduction

Gram-negative bacilli (GNB) are estimated to be responsible for 25% to 50% of all bloodstream infections, with reported rates of GNB bacteremia mortality ranging between 27% and 38%.^1,2 Epidemiologic data have shown that organisms of the Enterobacteriaceae family, a family within the Enterobacterales order, such as Escherichia coli, Klebsiella pneumoniae, and Enterobacter cloacae have been among the most commonly isolated bacteria from patients with GNB infections.^1,3 Patients usually at risk of bloodstream infections include those with malignancies, urologic and immunologic disorders, intravascular catheters, and certain chronic conditions such as diabetes mellitus.^{4 -6} Ceftriaxone (CRO) is a commonly utilized antibiotic in practice for community-acquired infections, with its frequent use attributed to factors such as its pharmacokinetic (PK) profile and its spectrum of activity.^{7 -9} The extensive protein binding of CRO has allowed for an extended half-life and the convenience of once-daily dosing.^8,9 Generally, its PK profile also allows it to be dosed irrespective of renal and hepatic function.^10,11 Ceftriaxone is a well-tolerated antibiotic, even at higher doses, with activity against gram-positive and gram-negative organisms, excluding many gram-negative anaerobes, Pseudomonas aeruginosa, and Enterococcus species, allowing it to be a favorable and safe antimicrobial stewardship option compared with broader agents.^12,13

Usual doses of CRO range from 1 to 4 g daily depending on the indication, with higher doses of 2 g every 12 hours for conditions such as meningitis or for enterococcal endocarditis synergy.^14,15 Doses for bacteremia have been recommended as a range from 1 to 2 g daily in the limited guidance available. ¹⁶ To date, there are limited data on the optimal dose of CRO in the setting of gram-negative bacteremia, although prior PK data may have influence on individual dosing decisions in certain patient populations. ¹⁷ Globally recognized organizations responsible for setting laboratory standards and breakpoints such as The Clinical & Laboratory Standards Institute (CLSI) and The European Committee on Antimicrobial Susceptibility Testing (EUCAST) base their breakpoints on different CRO doses.^18,19 The Clinical & Laboratory Standards Institute bases their breakpoints for Enterobacterales on 1 g standard daily dosing of CRO while EUCAST bases their breakpoints on 2 g standard daily dosing.^18,19 The relationship of CRO with albumin has been of particular interest due to its extensive protein binding that may in fact have an impact on drug exposure. ²⁰ Prior literature has sought to determine the role of hypoalbuminemia in clinical outcomes with associations of lower rates of treatment success in patients with suboptimal dosing. ²⁰ There have been conflicting outcomes from various studies comparing CRO 1 to 2 g daily in different types of infections, with pneumonia being the leading diagnosis in many of these publications.^{20 -22}

Optimizing antibiotic utilization based on PK properties potentially reduces the risk of treatment failure as well as the selection for resistant organisms and prior literature does suggest that CRO dosed at 1 g daily may be inadequate in targeting optimal therapeutic levels for many patients.^17,23,24 Antimicrobial resistance continues to be a global health issue and optimizing doses of antibiotics is a critical factor of antimicrobial stewardship practices. ²⁵ Thus, the paucity of data on CRO dosing for bloodstream infections raises the need for further investigation.

Methods

This is a retrospective, cohort study of patients admitted to any of the 3 NYU hospitals: Long Island, Tisch, or Brooklyn, with CRO-susceptible Enterobacterales bacteremia receiving definitive therapy with CRO 1 or 2 g daily from October 2019 to September 2020. Definitive therapy was defined as ≥3 doses of CRO 1 or 2 g daily without any concomitant therapy with in vitro activity against the isolated organism. This study was approved by the NYU Langone Health Institutional Review board, and the requirement for informed consent was waived.

Patients were identified using electronic medical records and were eligible for inclusion if they were ≥18 years old, were on definitive therapy of CRO 1 or 2 g daily, and had at least 1 positive blood culture (Figure 1). Patients were excluded if the isolated organism was outside of the Enterobacterales order, received >72 hours of empiric antibiotic therapy other than CRO, had an interruption of CRO therapy, adjunctive antibiotic therapy with activity against the isolated organism, polymicrobial growth in the blood cultures, concomitant treatment of an unrelated infection, or the organism isolated was resistant to CRO. To limit the confounding effects of the Sars-cov-2 (COVID-19) pandemic, we also excluded any patient admitted with confirmed COVID-19 or 90-day mortality that could be attributed to COVID-19.

Figure 1.

Study population.

The primary outcome of this study was the 90-day clinical failure rate which was defined as meeting ≥1 of the following criteria: relapse of infection within 90 days from discharge, escalation of antibiotics after definitive CRO therapy due to clinical worsening, and death from any cause. The secondary outcomes were 30-day rate of clinical failure, the rate of clinical worsening requiring escalation of antibiotics, the 30- and 90-day rates of all-cause mortality, relapse of infection, and rate of Clostridioides difficile infection (CDI), the length of hospital stay, and adverse events leading to the discontinuation of CRO.

Data Collection and Study Definitions

The data points collected for each patient were age, gender, height, weight, body mass index (BMI), CRO dose, the isolated organism, suspected source, duration of CRO definitive therapy, total duration of intravenous inpatient antibiotics, intensive care unit (ICU) stay requiring >48 hours, hypoalbuminemia within the first 24 hours of admission and the lowest level recorded in that time frame, discharge antibiotics and total duration of all antibiotics given, length of hospital stay, CDI events, relapse of infection, escalation of therapy, and all-cause mortality.

Patients on CRO definitive therapy were not excluded if they were on other antibiotics with no in vitro activity against the isolated organism. Interruption of therapy was defined as >30 hours between doses of CRO, in order to give leniency to doses given late. COVID-19 was confirmed by an inpatient polymerase chain reaction swab and mortality attributed to COVID-19 was defined as mortality due to any cause with a confirmed COVID-19 diagnosis at the time of death. Relapse of infection was defined as a subsequent bloodstream infection or re-infection of the confirmed primary source by the same genus and species of the original infection at 30 and 90 days post-discharge. Any patient with in-hospital mortality and/or inpatient relapse of infection after definitive therapy was included as a failure. Any patient without 30- or 90-day follow-up information was deemed as successfully treated.

Statistics

Descriptive statistics (median [25th, 75th percentiles] for continuous variables; frequencies and percentages for categorical variables) were calculated separately by group (1 g vs 2 g daily of CRO). The 2 groups were compared using the chi-square test or Fisher’s exact test, as deemed appropriate, for categorical variables and the 2-sample t-test or Mann-Whitney test for continuous data.

Length of stay was analyzed by applying standard methods of survival analysis, where the data were stratified by the above groups. No data were considered censored as all patients were discharged alive. The groups were compared using the log-rank test. The median rates for each group were obtained from the Kaplan-Meier/Product-Limit Estimates and their corresponding 95% confidence intervals (CIs) were computed, using Greenwood’s formula to calculate the standard error.

Univariable logistic regression was used to assess certain risk factors in predicting clinical failure. Results were reported as odds ratios with corresponding 95% confidence intervals. Those factors that appeared to be associated with clinical failure in the univariate analysis were included in a multivariable logistic regression model. A receiver operating characteristic (ROC) curve was constructed to look at the final model’s ability to predict the outcome. A numerical measure of the accuracy of the model was obtained from the area under the curve (AUC), where an area of 1.0 signifies near-perfect accuracy, while an area of less than 0.5 indicates that the model is worse than just flipping a coin. The following was used as a guide for AUC: 0.9-1.0 Excellent, 0.8-0.9 Very good, 0.7-0.8 Good, 0.6-0.7 Average, and 0.5-0.6 Poor. The Hosmer and Lemeshow Goodness-of-Fit test was used to test how well the model fits the data.

The result was considered statistically significant at the P < 0.05 level of significance. All analyses were performed using SAS version 9.4 (SAS Institute Inc., Cary, NC).

Results

A total of 932 patients with an Enterobacterales isolate in the blood were assessed for inclusion. During chart review, 808 patients were excluded, and the most common reason for exclusion was patients did not meet the criteria for CRO definitive therapy (Figure 1). Seventy-two patients (58%) were included in the CRO 1-g group and 52 (42%) in the 2-g group. There were no statistical differences between the baseline characteristics other than patients in the 2-g group being treated with antibiotics for a longer duration of time (Table 1). In the CRO 1 g and 2 g groups, the median ages were 74 (59, 86) and 71 (62, 79.5), respectively (P = 0.210). The most common source of infection was the urinary tract (75% vs 61.5%). E. coli was the most common organism isolated (75% vs 75%) followed by Klebsiella species (17% vs 23%). The Klebsiella species included were pneumoniae, oxytoca, and ozaenae. Eighty-nine patients were subsequently given an oral antibiotic after initial intravenous therapy (Table 2). The most common classes used as step-down oral agents were third-generation cephalosporins (45.3% vs 30.6%) and fluoroquinolones (28.3% vs 47.2%) (Table 2).

Table 1.

Baseline Characteristics.

Variable	Ceftriaxone 1 g (N = 72)	Ceftriaxone 2 g (N = 52)	P-value
Age (years), median (IQR)	74 (59, 86)	71 (62, 79.5)	0.211
Male, n (%)	31 (43.1)	22 (42.3)	0.934
Height (cm), mean (SD)	166.33 ± 9.81	166.97 ± 11.38	0.738
Weight (kg), mean (SD)	73.10 ± 15.73	74.13 ± 19.54	0.746
BMI (kg/m2), mean (SD)	26.45 ± 5.41	26.55 ± 6.49	0.921
ICU stay ≥48 hours, n (%)	8 (11.1)	8 (15.4)	0.484
Charlson Comorbidity Index, median (IQR)	5.5 (4, 8)	5 (3.5, 6.5)	0.544
History of solid/hematologic malignancy, n (%)	28 (38.9)	21 (40.4)	0.867
HTN, n (%)	48 (66.7)	40 (76.9)	0.214
HLD, n (%)	36 (50)	29 (55.8)	0.526
DM, n (%)	30 (41.7)	21 (40.4)	0.886
CVA, n (%)	10 (13.9)	6 (11.5)	0.700
CAD, n (%)	23 (31.9)	10 (19.2)	0.114
MI, n (%)	5 (6.9)	2 (3.9)	0.698
GERD, n (%)	15 (20.8)	7 (13.5)	0.289
HF, n (%)	13 (18.1)	9 (17.3)	0.914
PVD, n (%)	10 (13.9)	7 (13.5)	0.946
Asthma, n (%)	3 (4.2)	6 (11.5)	0.164
COPD, n (%)	8 (11.1)	3 (5.8)	0.356
CKD ≥ stage 3, n (%)	23 (31.9)	23 (44.2)	0.162
Liver disease, n (%)	3 (4.2)	5 (9.6)	0.278
Autoimmune disease, n (%)	5 (6.9)	4 (7.7)	1.000
HIV, n (%)	1 (1.4)	2 (3.9)	0.571
Prior organ transplant, n (%)	3 (4.2)	4 (7.7)	0.452
Neutropenia, a n (%)	2 (2.8)	2 (3.9)	1.000
Hypoalbuminemia on admission,b,c n (%)	12 (18.2)	4 (8)	0.115
Albumin levels on admission (g/dL), d median (IQR)	3.1 (2.7, 3.6)	3.3 (3, 3.7)	0.059
qSOFA score ≥ 2, n (%)	23 (31.9)	27 (51.9)	0.080
Duration of antibiotics (days), median (IQR)	13 (10.5, 14)	15 (13.5, 18)	<0.001
Suspected source of infection, e n (%)			0.150
• Urinary	54 (75)	32 (61.5)
• Biliary	4 (5.6)	8 (15.4)
• Prostate	1 (1.4)	0 (0)
• Intra-abdominal	4 (5.6)	6 (11.5)
• Joint	0 (0)	1 (1.9)
• Unknown	9 (12.5)	5 (9.6)

Abbreviations: BMI, body mass index; CAD, coronary artery disease; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease; CVA, cerebrovascular accident; DM, diabetes mellitus; GERD, gastroesophageal reflux disease; HF, heart failure; HIV, human immunodeficiency virus; HLD, hyperlipidemia; HTN, hypertension; ICU, intensive care unit; IQR, interquartile range; MI, myocardial infarction; PVD, peripheral vascular disease; qSOFA, quick sequential organ failure assessment.

^aNeutropenia was defined as having an inpatient absolute neutrophil count of <1500 cells/µL.

^bHypoalbuminemia was defined as albumin <2.5 g/dL within the first 24 hours of admission.

^cEight patients did not have recorded albumin levels on admission.

^dThe lowest albumin level within the first 24 hours was recorded (116/124 patients included).

^ePercentages do not equal exactly 100% due to rounding of percentages.

Table 2.

Oral Step-down Therapy.

Oral Antibiotic Class, n (%)a,b	Ceftriaxone 1g (N=53)	Ceftriaxone 2g (N=36)
1st generation cephalosporins	3 (5.7)	1 (2.8)
2nd generation cephalosporins	3 (5.7)	0 (0)
3rd generation cephalosporins	24 (45.3)	11 (30.6)
Fluoroquinolones	15 (28.3)	17 (47.2)
Penicillins	6 (11.3)	5 (13.9)
Other c	2 (3.8)	2 (5.6)

^aEighty-nine patients on definitive CRO therapy were subsequently given oral antibiotics

^bThere was no statistically significant difference

^cOther antibiotic classes included sulfonamides and tetracyclines

There was no statistically significant difference observed in the primary outcome of clinical failure between CRO 1 g and 2 g daily (Table 3). The 90-day clinical failure rate was 16.7% versus 9.6%; P = 0.260, respectively. The factors that contributed to the primary outcome were assessed individually in the secondary outcomes along with other endpoints (Table 3). There were no statistically significant differences in any of the secondary outcomes although there was a numerically greater rate of infection relapse in the 1-g group versus the 2-g group at 30 and 90 days (6.9% vs 1.9% and 11.1% vs 1.9%, respectively). In all cases of relapse, there were no unresolved source control conflicts noted during chart review. In the 9 patients who relapsed at 90 days, 8 of them had a urinary source while 1 had an unknown source of infection.

Table 3.

Outcomes.

Outcomes	Ceftriaxone 1 g (N = 72)	Ceftriaxone 2 g (N = 52)	P-value
Primary outcomes, n (%)
90-day clinical failure	12 (16.7)	5 (9.6)	0.260
Secondary outcomes
30-day clinical failure, n (%)	10 (13.9)	5 (9.6)	0.472
Escalation of antibiotics, n (%)	2 (2.8)	1 (1.9)	1.000
30-day mortality, n (%)	4 (5.6)	3 (5.8)	1.000
30-day relapse, n (%)	5 (6.9)	1 (1.9)	0.399
30-day rate of CDI, n (%)	2 (2.8)	1 (1.9)	1.000
90-day mortality, n (%)	5 (6.9)	3 (5.8)	1.000
90-day relapse, n (%)	8 (11.1)	1 (1.9)	0.078
90-day rate of CDI, n (%)	3 (4.2)	1 (1.9)	0.639
Length of stay (days), median (95% CI)	6 (5, 7)	7 (6, 7)	0.457
ADR resulting in discontinuation of CRO, n (%)	0 (0)	0 (0)	1.000

Abbreviations: ADR, adverse drug reaction; CDI, Clostridioides difficile infection; CI, confidence interval; CRO, ceftriaxone.

A univariate analysis was conducted to assess the risk factors predicting 90-day clinical failure (Table 4). Hypoalbuminemia defined as albumin <2.5 g/dL and a BMI of ≥25 kg/m² were the only variables with a statistically significant result. In a multivariate analysis (Table 4), hypoalbuminemia was the only factor associated with clinical failure, after adjusting for BMI and CRO dose, with patients being 4 times more likely to have a 90-day clinical failure compared with those without hypoalbuminemia (odds ratio = 4.03; 95% confidence interval = 1.12-14.50, P = 0.033).

Table 4.

Risk Factors Predicting 90-Day Clinical Failure.

Variables	Univariate analysis		Multivariate analysis a
	OR (95% CI)	P-value	OR (95% CI)	P-value
CRO dose (1 g vs 2 g)	1.88 (0.62-5.71)	0.265	1.50 (0.46-4.90)	0.499
Hypoalbuminemia b	5.40 (1.62-18.01)	0.006	4.03 (1.12-14.50)	0.033
BMI ≥ 25	0.34 (0.12-0.98)	0.047	0.52 (0.16-1.68)	0.277
BMI ≥ 30	0.40 (0.09-1.84)	0.237
qSOFA score ≥ 2	1.81 (0.65-5.07)	0.268
ICU stay ≥ 48 hours	1.55 (0.39-6.13)	0.533
CKD Stage ≥ 3	0.91 (0.31-2.66)	0.869
Transitioned to oral cephalosporins	1.44 (0.51-4.10)	0.495
Transitioned to oral fluoroquinolones	0.34 (0.07-1.59)	0.171
Transitioned to oral penicillins	0.61 (0.07-5.06)	0.644

Abbreviations: AUC, area under the curve; BMI, body mass index; CI, confidence interval; CKD, chronic kidney disease; CRO, ceftriaxone; ICU, intensive care unit; OR, odds ratio; qSOFA, quick sequential organ failure assessment.

^aHosmer-Lemeshow goodness-of-fit test (χ² = 0.573, df = 3, P < 0.903); AUC = 0.707.

^bHypoalbuminemia was defined as albumin <2.5 g/dL within the first 24 hours of admission.

Discussion

In this retrospective study, we found no statistically significant difference assessing the 90-day clinical failure rate between the use of CRO 1 g versus 2 g daily for treating Enterobacterales bacteremia.

Previous literature to date shows conflicting evidence on optimal CRO dosing for various types of infections being evaluated, which warrants comparative studies on specific infections.^{20 -23} None of the previous studies focused on bacteremia treatment with CRO specifically. This study was aimed to evaluate the optimal dose of CRO in patients with CRO-susceptible Enterobacterales bacteremia. To the authors’ knowledge, this is the first study assessing clinical failure comparing 2 doses of CRO in this patient population. It was not feasible to match the 2 groups due to the already limited population included. The majority of our patients were infected with E. coli or Klebsiella species (94%) which are consistent with the predominant organisms in prior literature. ³ Urinary sources were the most commonly suspected source of infection (69%).

Prior comparative studies primarily or exclusively included patients with pneumonia when assessing outcomes of CRO doses.^{20 -22} Urinary tract infections were the second most common infections included in the studies by Segev et al ²² and Ackerman et al. ²⁰ Hasegawa et al ²¹ studied CRO doses in patients with community-acquired pneumonia. They retrospectively assessed the clinical cure rates of CRO 1 g versus 2 g in pneumonia exclusively and concluded that 1 g was non-inferior to 2 g. ²¹ Segev et al ²² published a randomized controlled trial assessing the clinical efficacy of CRO 1 g versus 2 g in community-acquired infections. This study included 10 patients with bacteremia and the overall result was no statistical difference between the groups. Ackerman et al ²⁰ published a retrospective study assessing clinical failure in critically ill patients, comparing CRO 1 g versus 2 g. In contrast to Segev et al ²² and Hasegawa et al, ²¹ they found a statistical difference in clinical failure between the groups, favoring the 2-g group. ²⁰ Along with our primary endpoint, we found no statistically significant differences in any of the secondary outcomes, although we noted a trend of a greater rate of infection relapse in the 1-g group, of which 88% were from a urinary source which may have represented inadequate microbiological eradication or asymptomatic colonization. Due to the heterogeneity of the prior study populations and disease states, it is difficult to make a direct comparison of our study to the prior studies on whether this outcome was expected or not. Our study also differs significantly from the prior studies, having no patients with a suspected pulmonary source of infection. Our limited sample size may have had a role in the ability to detect a difference between the groups, especially subgroups like critically ill patients where we found no statistically significant differences compared to Ackermen et al. ²⁰

The role of hypoalbuminemia in CRO outcomes was of interest, so we stratified patients having hypoalbuminemia if their albumin levels were <2.5 g/dL in the first 24 hours of admission versus those without. Studies with a focus on hypoalbuminemia largely included critically ill patients, as this population exhibits PK alterations to a greater extent.^26,27 The majority of patients included in our study were not critically ill but due to the limited data available, we extrapolated the concept from previous studies in critically ill patients to assess if it may influence noncritically ill patients with hypoalbuminemia as well. Ceftriaxone is highly bound to plasma proteins (85%-95%), and hypoalbuminemia has displayed PK alterations which may affect the clinical outcomes in patients.^20,26 Despite a higher free fraction of CRO, increases in the volume of distribution and clearance with CRO were observed in patients with hypoalbuminemia. ²⁶ Schleibinger et al ²⁷ evaluated CRO PK in critically ill patients with hypoalbuminemia and found that CRO 2 g once daily maintained therapeutic concentrations during the entire dosing interval. In our study, hypoalbuminemia was associated with clinical failure. Although not statistically significant, the number of patients with hypoalbuminemia was numerically greater in the 1-g group (18.2% vs 8%, P = 0.115). The increased volume of distribution and clearance of CRO in the setting of hypoalbuminemia may play a critical role in underdosing patients with bacteremia, especially in patients receiving 1 g. ²⁶ The trend of hypoalbuminemia being more predominant in the 1-g group may have influenced the previously mentioned trend in rates of relapse due to possible subtherapeutic concentrations.

One barrier to maximizing antibiotic doses is the fear of adverse effects, although CRO is generally a tolerable antibiotic even at doses required for meningitis. ¹³ Turnier et al ¹³ assessed the tolerability of high-dose CRO (ie, daily dosage ≥4 g or ≥75 mg/kg) in central nervous system infections with only 1 of 196 cases resulting in discontinuation due to an adverse drug reaction. The studies conducted by Segev et al ²⁰ and Ackerman et al, ²² along with our study, found no statistical difference in side effects/antibiotic-limiting side effects between the 2 groups either. Data continue to support the overall tolerability of CRO as well as the minimal differences in tolerability between CRO 1 g versus 2 g.

Optimizing antibiotics by accounting for PK variations has become a growing interest in treating infectious diseases, although studies may be limited due to funding and practicality. ²⁸ Antibiotic-specific PK and pharmacodynamic parameters are emphasized in critical care settings, especially for beta-lactams, due to frequent PK alterations that influence this class. ²⁹ Although many studies on PK are focused on critically ill patients, alterations in PK variables like albumin are evident in noncritically ill patients as well.^30,31 Another variable cited in the literature that may influence outcomes is the BMI of patients, notably a BMI of ≥30 kg/m².^32,33 Prior literature has displayed an association of increased rates of clinical failure in this patient population, although we found no statistically significant difference in clinical failure with this subgroup. These inconsistent finding could have likely been due to the limited population we had available within that group.^32,33 Along with EUCAST currently recommending breakpoints based on 2-g CRO daily dosing, certain global guidelines make recommendations for 2 g of CRO daily in contrast to certain Infectious Diseases Society of America guidelines that give 1 to 2 g ranges for other types of infections.^18,34,35 Clinical data are limited, but current extrapolated PK and clinical data along with our findings further argue that CRO 2 g standard dosing for most infections may be needed for optimal CRO target attainment in the general population, especially with limited access to beta-lactam therapeutic drug monitoring for many institutions.^{20 -22,26,36}

There are several limitations to this study. First, this was a retrospective cohort study relying on documentation solely. There was a limited sample size, but in order to limit confounders from other antibiotics administered as well as polymicrobial growth, it was necessary to exclude these patients, which is a strength of our study. This limited sample size may have contributed to the nonsignificant findings. Patients who did not have any follow-up data were considered as successfully treated, which may have missed readmissions at different institutions outside the organization’s network of hospitals. Allowing up to 72 hours of empiric antibiotics, which were broad antibiotics compared to CRO, may have contributed to the overall outcomes due to the evolving data that shorter durations of antibiotics may be adequate in certain patient groups. ³⁷ This time cut-off was necessary in order to give clinicians time for de-escalation based on the turnaround time for cultures and susceptibilities. Pharmacokinetic factors, other than albumin, which may certainly have a role in CRO dosing were not included, such as creatinine clearance. Patients with augmented renal function may potentially lead to subtherapeutic levels of CRO and could have contributed to the outcomes as well. ³⁸ Applicability of these findings needs to be taken with consideration as these are purely exploratory data due to the nature of the design.

Conclusion

The exploratory findings of this study resulted in no statistically significant differences between CRO 1 g and 2 g for the treatment of Enterobacterales bacteremia, although there was a trend of increased rates of infection relapse in the 1-g group. Hypoalbuminemia was associated with increased rates of clinical failure. Prospective, randomized, adequately sampled studies are warranted to confirm these findings.