Pharmacist-Led Empiric Piperacillin-Tazobactam De-escalation Among Patients With Community-Acquired Intra-abdominal Infections

Abstract

Background: Piperacillin-Tazobactam (PTZ) is often used to treat community-acquired intra-abdominal infections (CA-IAIs) despite common causative pathogens being susceptible to more narrow-spectrum agents. However, susceptibility to PTZ among these predominant pathogens has been declining. Antibiotic de-escalation to non-antipseudomonal beta-lactams whenever possible is an important strategy to prevent the development of resistance to PTZ. Objective: The purpose of this study is to assess PTZ length of therapy in patients with CA-IAI, by comparing patients who received a pharmacist-led intervention involving the de-escalation of PTZ to narrow-spectrum regimens with those who did not receive the intervention. Methods: A retrospective analysis was conducted among patients >18 years old and admitted with CA-IAI empirically placed on PTZ between January 1, 2022, through June 30, 2022 (pre-intervention group), and January 1, 2024, through June 30, 2024 (post-intervention group). A total of 246 patients were included in the pre-intervention group and 129 patients in the post-intervention group. The utilization of PTZ, hospital length of stay (LOS), and treatment-associated complications were assessed using linear and logistic regression model, respectively. Results: Compared with patients in the pre-intervention group, those in the post-intervention group had a mean 1.2-day reduction in PTZ length of therapy (2.3 vs 1.2 days, P < 0.001). There was no difference in LOS, (β = 0.001, 98% confidence interval [CI] −1.29 to 1.29; P = 0.477), hospital readmission within 30 days due to IAI (odds ratio [OR] = 0.85, 98% CI = 0.51 to 1.44; P = 0.56), treatment-associated complications during current hospitalization (OR = 0.77, 98% CI = 0.45 to 1.32; P = 0.35), development of Clostridium difficile-associated diarrhea (OR = 3.29, 98% CI = 0.77 to 22.4; P = 0.14), or medication toxicity (OR = 2.07, 98% CI = 0.79 to 6.08; P = 0.15). Conclusion and relevance: The use of narrow-spectrum antibiotics for the empiric treatment of CA-IAI-reduced PTZ length of therapy and did not result in adverse clinical outcomes.

Introduction

Intra-abdominal infections (IAIs) include a wide spectrum of pathological conditions, ranging from uncomplicated appendicitis to fecal peritonitis. ¹ Intra-abdominal infections are classified into 2 groups: community-acquired intra-abdominal infections (CA-IAIs) and health care-associated intra-abdominal infections (HA-IAIs). ¹ The CA-IAIs are an important cause of morbidity and mortality with rates ranging from 10% to 20%. ² This is frequently associated with poor prognosis, particularly in higher-risk patients. ² Early prognostic evaluation of complicated IAIs is important to identify high risk patients for more aggressive treatment. ³ Complications after CA-IAIs are also of concern with an estimated recurrence rate of 15% to 30%. ²

Effective management of IAIs relies on 3 key components: surgical source control, consideration of patient-specific factors, and the use of appropriate antimicrobial therapy. ³ Recent guideline recommends the use of narrow-spectrum agents such as ceftriaxone and metronidazole due to known susceptibility to common causative pathogens, which include Escherichia coli, Klebsiella spp., Proteus spp., Streptococci, and anaerobic bacteria (especially Bacteroides fragilis). However, Piperacillin-Tazobactam (PTZ) is oftentimes being used to treat CA-IAIs.^4,5 The overuse of broad-spectrum antibiotics remains a leading cause of antimicrobial resistance, and multi-drug-resistant Pseudomonas is classified by the Centers for Disease Control and Prevention as a serious threat of concern to human health, affecting approximately 33 000 hospitalized patients each year. ² Recommendations published by the Infectious Diseases Society of America (IDSA) and Surgical Infection Society (SIS) support the use of narrow-spectrum antimicrobial regimens for the majority of patients presenting with CA-IAIs. ⁴

There has been a significant decline in the susceptibility to PTZ by the predominant bacteria involved in CA-IAI. ⁵ As resistance becomes more prevalent, infections have become increasingly more difficult to treat. ⁵ An antimicrobial-based approach to treating IAIs involves a delicate balance between the optimization of empirical therapy, which has been shown to improve clinical outcomes, and the reduction of excessive antimicrobial use, which has been proven to decrease the rate of emergence of antimicrobial-resistant strains. ⁶

Pharmacists play a major role in antimicrobial stewardship and de-escalation of antibiotics, and multiple studies have shown that antimicrobial stewardship programs are effective for improving antibiotic resistance patterns. ⁷ The purpose of this study was to assess PTZ length of therapy in patients with CA-IAI, by comparing patients who received a pharmacist-led intervention involving the de-escalation of PTZ to narrow-spectrum regimens with those who did not receive the intervention. We additionally aimed to compare treatment-associated complications in adult patients treated for CA-IAI with antipseudomonal vs narrow-spectrum regimens.

Materials and Methods

Patient Consent Statement

This study qualified for exemption after the Institutional Review Board review and did not require patient consent.

Study Design and Patient Population

A retrospective cohort study was conducted at Mayo Clinic in Florida. This study was approved by the Institutional Review Board. A report was generated from the electronic medical record of patients on PTZ between January 1, 2022, and June 30, 2022, and January 1, 2024, and June 30, 2024. Patients identified from the generated report were systematically screened through chart review to determine eligibility based on predefined inclusion criteria, until the target sample size was reached. Patients were eligible for inclusion if they were adults >18 years old and admitted with CA-IAI (defined as infection in the abdominal cavity outside of a health care setting or within 48 hours of hospital admission) empirically placed on PTZ between January 1, 2022, through June 30, 2022 (defined as the pre-intervention group), and January 1, 2024, through June 30, 2024 (defined as post-intervention group). Patients were excluded if they meet 1 or more of the following: age >70 years old with concomitant malignancy, if they had a concomitant infection requiring the use of PTZ, pregnant patients, diagnosed with sepsis or septic shock (defined as the presence of acute organ dysfunction ≥2 new SOFA points plus evidence of infection), received a solid organ or bone marrow transplant, received intravenous (IV) antibiotics in the past 90 days, had a diagnosis of cholangitis, pancreatitis, isolated Enterococcus or Pseudomonas aeruginosa from intra-abdominal fluid, abscess cultures, and/or blood cultures in the last year.

Data Collection

Patient, infection, and treatment characteristics were collected. This included patient demographic information on admission such as gender, height, and weight. Data collected regarding infection characteristics include empiric antibiotic on admission, type of IAI, and microbial susceptibility. Pharmacist-led intervention was implemented into the pharmacist verification workflow. All pharmacists were responsible for conducting chart reviews to identify patients who met the inclusion criteria and to recommend de-escalation to intravenous ceftriaxone plus either oral or intravenous metronidazole. During the post-intervention period, physicians were notified via email regarding the ongoing research to promote awareness and support timely de-escalation of PTZ within 24 to 48 hours for CA-IAIs, based on pharmacist recommendations. The stepwise algorithm pharmacist(s) used for de-escalation can be found in Figure 1.

Figure 1.

Pharmacist intervention.

Study Outcomes

The primary outcome was to evaluate the utilization of PTZ in the treatment of CA-IAIs by assessing the length of therapy (Table 3). The secondary outcome was to compare hospital length of stay (LOS) and treatment-associated complications during current hospitalization. Treatment-associated complications were defined as post-operative infections including surgical site infection or development of new intra-abdominal abscess, readmission to the hospital within 30 days due to IAI (diverticulitis, appendicitis, infectious colitis, peritonitis, liver/hepatic abscess, cholecystitis, and intra-abdominal abscess), development of Clostridium difficile-associated diarrhea, and medication toxicity such as acute kidney injury, hepatotoxicity, or hematologic abnormalities (Table 3).

Table 3.

Primary and Secondary Outcomes.

Primary outcome	Pre-intervention N = 117	Post-intervention N = 129	P-value
Length of therapy, days (mean ± SD)	2.3 ± 4.0	1.2 ± 2.0	<0.001
Secondary outcomes	Pre-intervention (N = 246)	Post-intervention (N = 129)	P-value
Length of stay, days (mean ± SD)	4.8 ± 5.9	4.8 ± 4.4	0.477
Hospital readmission 30 days, n (%)	45 (38.5)	45 (34.9)	0.597
Readmission reason, n (%)
Abdominal abscess/cyst	6 (5.1)	1 (0.8)	0.056
Abdominal pain	1 (0.9)	1 (0.8)	1.00
Anal fissure	0 (0.0)	1 (0.8)	1.00
Appendicitis	5 (4.3)	2 (1.6)	0.262
Ascites	1 (0.9)	2 (1.6)	1.00
Cholecystitis	5 (4.3)	0 (0.0)	0.023
Colitis	5 (4.3)	1 (0.8)	0.105
Diverticulitis	10 (8.5)	10 (7.8)	0.821
Pancreatitis	0 (0.0)	1 (0.8)	1.00
Sepsis	4 (3.4)	1 (0.8)	0.194
Small bowel obstruction	1 (0.9)	1 (0.8)	1.00
Clostridium difficile 30 days	2 (1.7%)	7 (5.4)	0.176
Treatment-associated complication, n (%)	41 (35.0)	38 (29.5)	0.412
Intra-abdominal abscess, n (%)	14 (12.0)	21 (16.3)	0.365
Intestinal obstruction, n (%)	3 (2.6)	3 (2.3)	1.00
Gangrene, n (%)	0 (0.0)	2 (1.6)	0.499
Intestinal perforation, n (%)	22 (18.8)	10 (7.8)	0.013
Acute kidney injury, n (%)	6 (5.1)	13 (10.1)	0.160

Statistical Analysis

All statistical analysis was performed using BlueSky Statistics software v.10.3. The primary outcome was analyzed using linear regression, and the secondary outcomes were analyzed using logistic regression. Categorical variables were analyzed using Fisher’s exact test, and continuous variables were summarized using the Wilcoxon rank-sum test with the sample mean. P-values less than 0.05 were considered statistically significant. We determined that 100 patients were needed to detect a difference with α = 0.05 and 80% power.

Results

Demographics

During the 2-year period, we identified and screened 1,087 patients. A total of 712 patients did not meet the inclusion criteria. Out of the 375 patients who met inclusion, 246 patients were included in the pre-intervention group and 129 patients in the post-intervention group (see Figure 2). There was no difference observed in baseline characteristics between groups (see Tables 1 and 2). Patients in the pre-intervention group were empirically treated with PTZ and were not de-escalated and patients in the post-intervention group were de-escalated if started on PTZ and empirically treated with ceftriaxone/metronidazole if recommendations were accepted by the primary team.

Figure 2.

Flowchart summarizing patient inclusion and exclusion screenings.

Table 1.

Patient Characteristics.

Characteristics	Pre-intervention (N = 246)	Post-intervention (N = 129)	P-value
Age, years (mean ± SD)	63.2 ±16.6	62.5 ± 16.0	0.616
Race, n (%)			0.103
White	100 (85.5)	113 (87.6)
Asian	2 (1.7)	6 (4.7)
Black or African American	12 (10.3)	4 (3.1)
Hispanic or Latino	1 (0.9)	3 (2.3)
Sex, n (%)			0.897
Female	68 (58.1)	73 (56.6)
Male	49 (41.9)	56 (43.4)
BMI, kg/m2 (mean ± SD)	27.5 ± (6.5)	27.8 ± (6.6)	0.705
CrCl, mL/min (mean ± SD)	98.2 ± (79.5)	85.5 ± (42.3)	0.167
eGFR mL/min (mean ± SD)	75.5 ± (21.1)	74.9 ± (20.3)	0.507
Scr mg/dL (mean ± SD)	1.1 ± (1.1)	1.2 ± (1.5)	0.210
Antibiotics allergies, n (%)
Sulfa	21 (17.9)	68 (52.7)	<0.001
Penicillin and ertapenem	9 (7.7)	18 (14.0)	0.012
Metronidazole	0 (0.0)	4 (3.1)	0.124
Macrolides and clindamycin	5 (4.3)	2 (1.6)	0.262
Fluoroquinolones	9 (7.7)	9 (7.0)	1.00
Cephalosporins	1 (0.9)	2 (1.6)	1.00
Piperacillin/tazobactam	1 (0.9)	1 (0.8)	1.00
Vancomycin and doxycycline	3 (2.6)	0 (0.0)	0.106

Abbreviations: BMI, body mass index, classification based on the World Health Organization; CrCl, creatinine clearance (Cockcroft-Gault equation); eGFR, estimated glomerular filtration rate; Scr, serum creatinine.

Table 2.

Infection Characteristics.

Characteristics	Pre-intervention (N = 246)	Post-intervention (N = 129)	P-value
Infection type, n (%)			0.373
Diverticulitis	37 (31.6)	47 (36.4)
Appendicitis	23 (19.7)	14 (10.9)
Infectious colitis	23 (19.7)	28 (21.7)
Cholecystitis	25 (21.4)	27 (20.9)
Hepatic abscess	4 (3.4)	9 (7.0)
Peritonitis	5 (4.3)	4 (3.1)
Microbes, n (%)
Staphylococcus	0 (0.0)	5 (3.9)	0.061
Streptococcus	0 (0.0)	1 (0.8)	1.00
Klebsiella	0 (0.0)	1 (0.8)	1.00
Fusobacterium	0 (0.0)	2 (1.6)	0.499
Enterobacter	0 (0.0)	1 (0.8)	1.00
Escherichia coli	1 (0.9)	4 (3.1)	0.373
Bacteroides	0 (0.0)	1 (0.8)	1.00
Clostridium	0 (0.0)	1 (0.8)	1.00
Arachnida	0 (0.0)	1 (0.8)	1.00
Cutibacterium	0 (0.0)	1 (0.8)	1.00
Susceptibility, n (%)	0 (0.0)	8 (6.2)	0.007
Post-op infection, n (%)	1 (0.9)	0 (0.0)	0.476
Current malignancy, n (%)	18 (15.4)	24 (18.6)	0.611
Hepatotoxicity, n (%)	5 (4.3)	8 (6.2)	0.577
Significant liver disease, n (%)	7 (6.0)	15 (11.6)	0.179
Significant renal disease, n (%)	18 (15.4)	23 (17.8)	0.732
History of multidrug-resistant organism, n (%)	2 (1.7)	0 (0.0)	0.225
Source control >24 h, n (%)	33 (28.2)	31 (24.0)	0.471
ID consulted, n (%)	7 (6.0)	9 (7.0)	0.801
Significant intra-abdominal disease, n (%)
History of abdominal surgery	5 (4.3)	4 (3.1)	.740
Recurrent C. diff/obstruction	5 (4.3)	2 (1.6)	.262
Chronic abdominal pain	6 (5.1)	3 (2.3)	.316
History of colitis/cholecystitis	4 (3.4)	11 (8.5)	.114
History of diverticulitis	9 (7.7)	23 (17.8)	.022
History of pancreatitis	1 (0.9)	4 (3.1)	.373
History of gastritis	7 (6.0)	3 (2.3)	.200
Irritable bowel disease	1 (0.9)	3 (2.3)	.624
History of gastrointestinal cancer	1 (0.9)	2 (1.6)	1.00

Definitions: Hepatotoxicity defined as ALT or AST >3× upper limit of normal (ULN) with symptoms of liver injury, or ALT or AST >5× ULN without symptoms.

Significant liver disease was defined as clinical or laboratory evidence of chronic hepatic dysfunction, including cirrhosis, hepatic insufficiency, or persistent elevations in liver enzymes (alanine aminotransferase [ALT], aspartate aminotransferase [AST], alkaline phosphatase [ALP], or total bilirubin) greater than 2 to 3 times the upper limit of normal (ULN), in the absence of another identifiable cause.

Significant renal disease was defined as chronic kidney disease (CKD) stage 3 or higher, indicated by an estimated glomerular filtration rate (eGFR) less than 60 mL/min/1.73 m² for 3 months or more. Conditions such as persistent proteinuria, structural kidney abnormalities, or a history of dialysis or kidney transplantation.

Abbreviation: ID, infectious disease.

Main Outcomes

There was a statistical difference in the length of therapy of PTZ between the pre-intervention and post-intervention groups (β = −1.6, 98% confidence interval [CI] −1.94 to −0.37; P = 0.0043) (Table 3). Hospital LOS (β = 0.001, 98% CI = −1.29 to 1.29; P = 0.99), hospital readmission within 30 days due to IAI (OR = 0.85, 98% CI = 0.51 to 1.44; P = 0.56), treatment-associated complications during current hospitalization (OR = 0.77, 98% CI = 0.45 to 1.32; P = 0.35), development of C. diff-associated diarrhea (OR = 3.29, 98% CI = 0.77 to 22.4; P = 0.14), and medication toxicity (OR = 2.07, 98% CI = 0.79 to 6.08; P = 0.15) were not statistically significant (Table 3).

Discussion

These findings contribute meaningfully to the growing body of evidence supporting the de-escalation of broad-spectrum antibiotics in favor of targeted, narrow-spectrum regimens for the treatment of CA-IAIs. Through the implementation of a pharmacist-led antimicrobial stewardship intervention, we observed a significant reduction in the duration of PTZ use without compromising clinical outcomes. This reduction not only reflects an advancement in antimicrobial stewardship practices but also offers considerable cost savings, as PTZ was found to be more expensive than the combination of ceftriaxone and metronidazole.

Importantly, our findings align with prior research, including the seminal work by Worden et al ⁵ by demonstrating no significant differences in key clinical outcomes such as treatment-associated complications, hospital LOS, readmissions, mortality, and the incidence of C. difficile infection between patients treated with PTZ and those receiving narrow-spectrum regimens. ⁴ The result of the study showed that there were no differences in 90-day treatment-associated complications between groups (antipseudomonal 15.1% vs narrow spectrum 11.3%, P = 0.296). ⁵ In addition, Beckermann et al ⁷ demonstrated in a quality improvement study; optimizing antibiotic management for patients with acute appendicitis that the rate of preoperative PTZ administration significantly decreased after the intervention (51.4% pre-intervention period vs 20.1% postintervention period, P < 0.001). ⁷ The rate of surgical site infections was similar in both groups (superficial surgical site infections = 1.4% pre-intervention period vs 0.8% postintervention period, P = 0.50; deep surgical site infections = 1.1% pre-intervention period vs 0.0% postintervention period, P = 0.06; and organ space surgical site infections = 3.1% pre-intervention period vs 3.0% postintervention period, P > 0.99). ⁷ Rates of 30-day readmission, reoperation, and C. difficile infection also did not differ between groups. ⁷ This study is distinct in its focus on ceftriaxone plus metronidazole as the recommended de-escalation regimen. Other narrow-spectrum antibiotic alternatives were not evaluated; we limited the evaluation to a single narrow-spectrum regimen to allow for a focused and consistent intervention across all eligible patients, thereby reducing variability and improving the interpretability of outcomes, and patients with complicated CA-IAIs were excluded to maintain a more targeted population.

It is also important to note that we found no difference in patient outcomes, LOS, hospital admission, mortality, or development of C. difficile between groups, which further supports the safety of using narrow-spectrum regimens for CA-IAI. The consistency of these findings across multiple studies reinforces the safety and efficacy of narrow-spectrum therapy in appropriately selected patients with CA-IAI.

Considering the increase in antimicrobial resistance and a heightened emphasis on patient safety, these findings are both timely and relevant. They support ongoing efforts to limit unnecessary exposure to broad-spectrum antibiotics and promote judicious antimicrobial use. ⁸ Future prospective studies are warranted to further validate these findings and to assess the long-term impacts of stewardship interventions on resistance patterns and institutional antibiotic utilization.

Several limitations should be considered. As a retrospective cohort study, the potential for selection and information bias exists, alongside a strong dependence on the accuracy and completeness of medical record documentation. Although documentation practices remained consistent over the study period, data collection was conducted by a single investigator to minimize interpretative variability. In addition, our analysis did not include patients treated with empiric antibiotic regimens other than ceftriaxone and metronidazole, which may limit generalizability.

Confounding variables not accounted for may have influenced primary outcomes and readmission rates. Furthermore, while the rate of treatment-associated complications appeared similar between groups, complications or hospital readmissions occurring outside our institution within the 90-day follow-up period may have been missed. Finally, the stringent exclusion criteria resulted in a substantial number of patients being excluded, potentially limiting the applicability of our findings to higher-risk populations. Given the relatively small sample size, our results should be interpreted with caution.

Despite these limitations, our study adds valuable insight to the limited existing literature supporting the use of narrow-spectrum agents in adult patients with CA-IAI. These findings reinforce the role of stewardship-driven interventions in optimizing antibiotic use while maintaining clinical effectiveness.

Conclusion

The result of our study depicts that a pharmacist-led interventions decreased PTZ utilization, and secondary outcomes were overall similar between the 2 groups. The use of narrow-spectrum antibiotics for the empiric treatment of CA-IAI did not lead to adverse clinical outcomes and may contribute to decreased incidence of antipseudomonal resistance. This suggests that pharmacists play a crucial role in mitigating antimicrobial resistance, highlighting the potential for strategic antibiotic stewardship to optimize therapeutic outcomes.