Primary and secondary SBP prophylaxis

AASLD Guidelines recommend lifelong secondary spontaneous bacterial peritonitis (SBP) prophylaxis for all patients with prior SBP and primary prophylaxis for high-risk patients (ie, patients with advanced cirrhosis and low ascites protein (<1.5 g/dL) with either impaired renal function or hyponatremia). ^{Reference Biggins, Angeli and Garcia-Tsao4}

In recent years, this approach to SBP prophylaxis has been questioned. A 2020 Cochrane meta-analysis of 23 RCTs found no significant benefit of primary or secondary SBP prophylaxis on mortality or number of serious adverse events, however noted the underlying poor quality of included trials. ^{Reference Komolafe, Roberts and Freeman5} Concurrently, concerns regarding toxicity and antibiotic resistance have intensified. One retrospective Veterans Affairs (VA) study (n = 7,553) found that among patients who were admitted for a first SBP case, those who received primary prophylaxis had a higher proportion of drug-resistant organisms compared to those who received no primary prophylaxis. ^{Reference Badal, Silvey and Dragilev6} Another retrospective VA study (n = 11,381) found that patients who received secondary SBP prophylaxis had higher rates of SBP recurrence in multivariate analysis compared to those who received no prophylaxis. Of the 100 patients with available microbiologic data, those who received fluoroquinolone secondary prophylaxis had higher rates of fluoroquinolone-resistant isolates compared to those who received no prophylaxis. ^{Reference Silvey, Patel and Tsai7}

Given this data, clinicians are re-evaluating both primary and secondary prophylaxis. The ongoing ASEPTIC trial may provide clarity. ^{Reference Crocombe, Ahmed and Balakrishnan8} In the meantime, Markley et al. propose revising institutional SBP prophylaxis guidelines to reflect the limited evidence of benefit and growing evidence of harm. ^{Reference Markley and Bajaj9}

SBP prophylaxis in patients with cirrhosis and upper gastrointestinal bleed

In patients with cirrhosis and upper gastrointestinal bleeding (UGIB), antibiotic prophylaxis is the standard of care in Western countries. This practice is based on a meta-analysis of five studies of patients with cirrhosis and gastrointestinal bleeding (n = 534), in which short antibiotic durations (mostly oral norfloxacin) were associated with reduced rates of infections, bacteremia and/or SBP, and improved survival. ^{Reference Bernard, Grangé, Khac, Amiot, Opolon and Poynard10} The AASLD 2021 guidelines recommend intravenous ceftriaxone due to increasing fluoroquinolone resistance. ^{Reference Biggins, Angeli and Garcia-Tsao4}

Recent studies question this practice. A randomized controlled trial from India included 180 hemodynamically stable patients with Child-Pugh A cirrhosis admitted with UGIB and found no difference in infections at 5 days or 6-week mortality between those who received ceftriaxone or placebo. ^{Reference Gupta, Agarwal, Sharma, Gopi, Gunjan and Saraya11} One retrospective observational study (n = 790) in Japan included patients with esophageal variceal bleeding who received emergency endoscopic variceal ligation and found no differences in the primary composite outcome of 6-week mortality, 4-week rebleeding or 4-week SBP onset in those who received antibiotics compared to those who did not. ^{Reference Ichita, Shimizu and Goto12} A smaller Japanese retrospective cohort study (n = 150) of patients admitted with acute variceal bleed also found no differences in bacterial infection, in-hospital mortality, 5-day rebleeding rate, and 30-day emergency re-admission in those who received antibiotics compared to those who did not. ^{Reference Ueno, Kayahara and Sunami13} These findings may reflect improved modern endoscopic hemostasis techniques and UGIB survival and suggest that further work is needed to evaluate necessity of SBP prophylaxis in UGIB.

While awaiting more data, ASPs can also target SBP prophylaxis duration in patients with cirrhosis and UGIB. National guidelines recommend antibiotics only until bleeding resolves and the patient stabilizes, up to a maximum of 7 days, not a fixed 7-day duration. ^{Reference Biggins, Angeli and Garcia-Tsao4}

Enterococcal coverage

IAIs, including biliary tract infections, are commonly polymicrobial, but the significance of Enterococcus spp. in cultures or the need for empiric treatment is uncertain. Observational studies have not demonstrated a consistent benefit of empiric anti-enterococcal therapy in IAIs, especially those of community onset. ^{Reference Sanders, Tessier and Sawyer18} Furthermore, a meta-analysis found no improvement in treatment success or mortality when anti-enterococcal regimens were used for IAIs. ^{Reference Zhang, Yu and Chen19} In biliary infections specifically, although Enterococcus spp. are frequently isolated in bile cultures (13%–30%), ^{Reference Gomi, Solomkin and Schlossberg20} their pathogenic significance in mixed infections without bacteremia is unclear.

Despite this, broad-spectrum regimens that include enterococcal coverage are commonly used in biliary infections, often driven by national guidelines. The Tokyo Guidelines recommend anti-enterococcal therapy for patients with grade III community-acquired (ie, severe infections associated with organ dysfunction) or healthcare-associated acute cholangitis and cholecystitis. ^{Reference Gomi, Solomkin and Schlossberg20} Similarly, the IDSA guidelines suggest enterococcal coverage in patients with healthcare-associated IAIs, particularly in those with postoperative infections, prior exposure to cephalosporins, immunocompromising conditions, valvular heart disease or prosthetic intravascular devices. ^{Reference Solomkin, Mazuski and Bradley16}

However, traditional risk factors may inadequately predict biliary infections with enterococcal bacteremia. Mussa et al. examined 850 cases of bacteremia due to biliary infections from an existing Spanish cohort; 73 (8.5%) were due to Enterococcus spp. ^{Reference Mussa, Martínez Pérez-Crespo and Lopez-Cortes21} They developed a predictive risk score for enterococcal bacteremic biliary tract infections incorporating five variables: cholangiocarcinoma, biliary prosthesis, hospital-acquired infection, prior carbapenem use, and chronic kidney disease. This score demonstrated a negative predictive value (NPV) of 95% at a threshold of six or more points and outperformed the Tokyo Guidelines risk factors (NPV of 92.5%.) However, both demonstrated low positive predictive values (PPV) (Mussa et al: PPV 20.8%; Tokyo Guidelines: PPV: 11.2%).

While it is still unclear which patients truly benefit from empiric enterococcal coverage, ASPs can play a crucial role in ensuring that broad-spectrum regimens with enterococcal activity (eg, piperacillin-tazobactam) are not routinely used for low-risk patients. More definitive evidence is needed to guide when such coverage is necessary.

Anaerobic coverage

Anaerobic coverage is often reflexively administered in biliary tract infections but this practice has limited benefit. ^{Reference Beheshti, Graber, Goetz and Bluestone22} In a retrospective analysis of intraoperative biliary cultures, Strohäker et al found that only 10% of 365 patient cultures had anaerobes. ^{Reference Strohäker, Wiegand, Beltzer, Königsrainer, Ladurner and Meier23} Guidelines recommend anaerobic coverage only in cases of documented anaerobic bacteremia or in patients with biliary-enteric anastomoses. ^{Reference Solomkin, Mazuski and Bradley16,Reference Gomi, Solomkin and Schlossberg20}

This targeted approach is supported by clinical data. In a randomized controlled trial of 100 patients with cholangitis, Sung et al. compared ciprofloxacin monotherapy to combination therapy of ampicillin, ceftazidime, and metronidazole and found no difference in urgent procedural intervention or mortality. ^{Reference Sung, Lyon and Suen24} In a cohort of 87 patients with community-onset biliary tract infections without anaerobic bacteremia, Wu et al. found that definitive therapy without anti-anaerobic coverage was not associated with treatment failure, defined as relapse and 28-day mortality. ^{Reference Wu, Chuang, Wu, Lin and Wang25} In a cohort of patients with biliary tract infections, Simeonova et al. showed no significant difference in 30-day mortality or 90-day relapse between patients who received anaerobic coverage and those who did not. ^{Reference Simeonova, Daneman, Lam and Elligsen26} These findings support current guideline recommendations that routine anaerobic coverage has limited clinical value in biliary infections except in specific situations such as anaerobic bacteremia or presence of biliary anastomoses.

Amoxicillin-clavulanate dosing as step-down therapy for Enterobacterales bacteremia associated with IAI

Few randomized controlled trials for gram-negative bacteremia that reported use of oral beta-lactam agents for step-down therapy disclosed specific drugs and doses. ^{Reference Yahav, Franceschini and Koppel28–30} Observational studies signaled an increased risk of recurrence with stepdown to oral beta-lactam agents compared with fluoroquinolones or trimethoprim/sulfamethoxazole. ^{Reference Punjabi, Tien, Meng, Deresinski and Holubar31,Reference Alzaidi, Veillette and May32} Some hypothesize that inadequate oral beta-lactam dosing may be a contributing factor, though this has not yet been substantiated in clinical study.

The pharmacokinetic/pharmacodynamic (PK/PD) target of beta-lactams in Enterobacterales bacteremia is the percentage of time in a dosing interval where free serum drug concentration exceeds the minimum inhibitory concentration (MIC) of the pathogen. Modeling data for the most used doses of 875/125 mg orally twice daily or 500/125 mg orally three times daily suggest adequate exposure for Enterobacterales with amoxicillin MICs of < = 1 mg/L. A dose of 875/125 mg orally three times daily may be sufficient for amoxicillin MICs < = 2 mg/L. ^{Reference Narayanan, Mathers and Wenzler33,Reference Heil, Bork and Abbo34} However, current Clinical Laboratory Standards Institute (CLSI) breakpoints for Enterobacterales for amoxicillin-clavulanate define amoxicillin MICs of 8 mg/L or lower as “susceptible.” ³⁵ While CLSI provides a caveat that these recommendations are solely for the purposes of uncomplicated urinary tract infection treatment, clinicians often erroneously extrapolate these results to other infections, like IAI. Further complicating matters, commercially available susceptibility testing systems may not test MICs as low as 1–2 mg/L. Moreover, this MIC range bisects the wild-type MIC distribution for Enterobacterales, limiting the reliability of such testing. ^{Reference Narayanan, Mathers and Wenzler33}

In addition to 875/125 mg three times daily dosing, various other strategies to escalate dosing to achieve the PK/PD target in the setting of higher MICs have been explored, such as 875/125 mg orally twice daily with additional dose(s) (ie, 1 gram) of oral amoxicillin, and the use of extended-release formulations. However, increasing dosing may not be feasible due to gastrointestinal tolerability as well as saturable amoxicillin absorption—effectively limiting the maximum amount of amoxicillin exposure that can be achieved through oral administration. ^{Reference Navarro36–Reference Vree38}

A final practical issue is the limited availability of direct susceptibility testing, which can lead to the common misconception that ampicillin-sulbactam susceptibility can be used as a proxy because it predicts that of amoxicillin-clavulanate. In fact, ampicillin-sulbactam grossly underestimates amoxicillin-clavulanate susceptibility in Enterobacterales given that the two agents are fundamentally different, particularly regarding their beta-lactamase inhibitor. ³⁹

Taken together, these data suggest that a dosage of 875/125 mg orally three times daily appears to be optimal when amoxicillin-clavulanate is used as step-down therapy for bacteremia due to Enterobacterales with MICs < = 2 mg/L. It should be noted that much of the clinical data for E. bacteremia are in patients with a urinary infection source so further study in IAI is needed.

Amoxicillin-clavulanate dosing for non-bacteremic diverticulitis

As opposed to gram-negative bacteremia, the PK/PD correlate for efficacy in non-bacteremic IAI is less well understood. That said, two randomized controlled trials, the DIVER trial and the DINAMO trial, assessed amoxicillin-clavulanate as the sole therapy of uncomplicated diverticulitis. Both trials used amoxicillin-clavulanate dosed at 875/125 mg orally three times daily, though the authors do not explicitly mention why this higher dose was used. The DIVER trial compared outpatient management with 10 days of oral amoxicillin-clavulanate to inpatient management with parenteral antibiotic therapy. They found no difference in their primary end point of readmission and lower healthcare costs in the outpatient group. ^{Reference Biondo, Golda and Kreisler40} The DINAMO trial compared 7 days of amoxicillin-clavulanate to non-antibiotic treatment and found no difference in their primary end point of hospitalization and no difference in secondary end points of emergency department visits and pain control. ^{Reference Mora-López, Ruiz-Edo and Estrada-Ferrer41}

Based on the findings of the DINAMO trial and other studies, most patients with uncomplicated diverticulitis generally do not require antibiotics. For cases where amoxicillin-clavulanate is chosen, the DIVER trial supports a dose of 875/125 mg orally three times daily, although it should be acknowledged that twice-daily dosing in this setting has not been studied.

Amoxicillin-clavulanate dosing in non-bacteremic biliary infections, acute appendicitis, or secondary peritonitis

No randomized trials currently exist for amoxicillin-clavulanate as sole therapy for biliary infections, acute appendicitis, or secondary peritonitis; however, data do exist for stepdown therapy.

In complicated IAI with secondary peritonitis, one randomized controlled trial compared stepdown therapy with oral amoxicillin-clavulanate dosed at 800/114mg twice daily to moxifloxacin after piperacillin-tazobactam initial therapy. ^{Reference Malangoni, Song, Herrington, Choudhri and Pertel42} Only 2 of the 315 clinically evaluated patients had bacteremia. Clinical cure rates were similar between the moxifloxacin and beta-lactam group. However, durations of therapy of up to14 days were allowed, and median durations used, including duration of parenteral therapy, were not disclosed. Thus, it is unclear how much oral step-down therapy contributed to efficacy.

In medical management of acute appendicitis, two randomized controlled trials exist for oral amoxicillin-clavulanate step-down therapy. In the COMMA trial, O’Leary et al. used parenteral amoxicillin-clavulanate with step-down to oral amoxicillin-clavulanate dosed at 500/125 mg orally three times daily in both the antibiotic-only and surgical arms. ^{Reference O’Leary, Walsh and Bolger43} In the ASAA trial, Ceresoli et al. used ertapenem with step-down down to oral amoxicillin-clavulanate dosed at 875/125 mg orally three times daily. ^{Reference Ceresoli, Pisano and Allievi44} Of note, while ASAA and several randomized trials that primarily used fluoroquinolone step-down therapy showed non-inferiority of antibiotics-only compared with surgical management, the COMMA trial (which used the lower amoxicillin-clavulanate dose) found inferiority of an antibiotics-only approach in a recent meta-analysis. ^{Reference De Almeida Leite, Seo and Gomez-Eslava45} These data suggest that 875/125 mg orally twice daily may be acceptable for oral step-down therapy in complicated IAI, with possible preference for 875/125 mg orally three times daily for step-down therapy in non-surgical management of appendicitis.

Pharmacokinetics and pharmacodynamics

Metronidazole is thought to exhibit concentration-dependent killing. ^{Reference Sprandel, Drusano, Hecht, Rotschafer, Danziger and Rodvold48} Its half-life is 8–12 hours and serum concentrations at 12 hours remain above the in vitro MIC for most anaerobic pathogens. Given its high peak levels, low protein binding, prolonged half-life, active metabolite, and post-antibiotic effect of greater than 3 hours, there is strong PK/PD justification for 500 mg twice daily dosing. ^{Reference Sprandel, Drusano, Hecht, Rotschafer, Danziger and Rodvold48,Reference Lamp, Freeman, Klutman and Lacy49}

Alternative dosing in anaerobic infections in adults

Clinical data also support 500 mg twice daily metronidazole dosing. One retrospective study of 200 patients with anaerobic or mixed infections (84% IAI) found no difference in clinical cure rates when either twice or three times daily regimens were used. ^{Reference Soule, Green and Blanchette50} Additionally, two retrospective studies focusing on anaerobic bacteremia reported no significant differences in 30-day mortality between the two dosing strategies. ^{Reference Shah, Adams, Merwede, McManus and Topal51,Reference Shah, Adams and Clarke52}

Twice daily dosing may also reduce the risk of adverse effects associated with cumulative metronidazole exposure, including peripheral neuropathy, encephalopathy, leukopenia, and pancreatitis. ^{Reference Shah, Adams and Clarke52,Reference Goolsby, Jakeman and Gaynes53}

Alternative dosing for appendicitis in the pediatric population

Although PK/PD data is limited in pediatric patients, a dosing regimen of every 6–8 hours has been historically used. ^{Reference Child, Chen and Mistry54} However multiple centers favor once-daily dosing for appendicitis. In a prospective study of children with appendicitis, metronidazole administered once daily at 30 mg/kg per dose achieved AUC target attainment for Bacteroides fragilis with an MIC of 2 mcg/mL or less. ^{Reference Child, Chen and Mistry54} A pediatric prospective randomized trial comparing once daily metronidazole and ceftriaxone to ampicillin, gentamicin, and clindamycin for perforated appendicitis found no differences in abscess rates or wound infections, while noting that the once-daily regimen was more cost-effective. ^{Reference St Peter, Tsao and Spilde55}