By: Andrew Vogler, PharmD; PGY1 Pharmacy Resident
Mentor: Daniel Hansen, PharmD; Clinical Pharmacy Specialist
Mercy Hospital Springfield
Approval Dates: June 1, 2021 – December 1, 2021
Approved Contact Hours: 1 hour
Take CE Quiz
There are approximately 2 million cases and 250,000 deaths annually from sepsis due to bacteremia with approximately 45% due to a gram-negative pathogen in North America and Europe.1 With the prevalence of gram-negative bacteremia so high and no current guideline discussing proper treatment regimens or duration, it is important to have a clear understanding of bacteremia before one can start a proper treatment regimen.2 Bacteremia is defined as bacteria in the blood and can often be asymptomatic or transient becoming a blood stream infection if the immune system becomes overwhelmed. Bacteremia should be differentiated from sepsis or septicemia. The Surviving Sepsis Campaign defines sepsis as, “life-threatening organ dysfunction caused by a dysregulated host response to infection.”3 Septicemia is a narrower term for sepsis that is caused by bacterial spread into the blood stream. Throughout this article the discussion of sepsis will be kept separate from the discussion of bacteremia, as they are not interchangeable terms.
Bacteremia can vary in source of infection and infectious pathogen. Though bacteremia can occur due to direct inoculation into the blood stream, it typically occurs as the result of an infectious pathogen spreading to the blood from another source. Bacteremia is classified by 3 main criteria: infectious pathogen, the source of infection, and whether the bacteremia is complicated or uncomplicated. The source of infection can either be primary or secondary. A primary bacteremia is caused from direct inoculation of pathogen into the bloodstream. A secondary bacteremia is caused by a pathogen entering the body from a site other than direct inoculation such as bacteremia secondary to pneumonia or urinary tract infection.4
During the initial Gram-stain phase, pathogen-based classification of bacteremia is typically either Gram-positive or Gram-negative. The most common cause of gram-positive bacteremia is Staphylococcus aureus (S. aureus), which is due to the organism’s ability to produce the enzyme coagulase, which can convert fibrinogen in the blood to fibrin causing the blood to clot.1 The infectious emboli then stick to different areas of the body like blood vessels or heart valves, making a bacteremia very difficult to clear. Other Gram-positive bacteria such as enterococcus and coagulase-negative staphylococcus can form biofilms making them difficult to treat, as well. In comparison to Gram-positive bacteria, Gram-negative bacteria do not produce coagulase and are often easier to treat, with patients often being able to clear infection with oral antibiotics and shorter durations of therapy.
The severity of bacteremia is classified as either complicated or uncomplicated based on the likelihood of a timely resolution of infection. To be considered an uncomplicated bacteremia, the patient must be afebrile within 72 hours of initial treatment, have a negative repeat blood cultures obtained 2-4 days after initial set, and not have endocarditis or metastatic infection. Complicated bacteremia is often treated for longer durations with IV antibiotics due to severity of illness, high inoculum of infection, lack of treatment response, seeding of infection, or a combination there of. Morpeth and colleagues looked at the rate of endocarditis in 2761 patient cases with species other than Haemophilus species, Actinobacillus actinomycetemcomitans, Cardiobacterium hominis, Eikenella corrodens, or Kingella species (non-HACEK) Gram-negative bacteremia. The study determined the risk of complicated bacteremia due to Gram-negative endocarditis is extremely low at approximately 1.8% with a high percentage of those patients having some sort of implanted endovascular device (29%).6
The Infectious Disease Society of America (IDSA) for the treatment of Methicillin-Resistant S. aureus (MRSA) bacteremia. They do not address bacteremia due to organisms other than MRSA. The rate of mortality for MRSA bacteremia with endocarditis (upwards of 37%) is higher than that of gram-negative bacteremia (12.5%).7 For MRSA bacteremia, IDSA recommends at least 14 days of IV anti-MRSA antibiotics following negative blood cultures. Historically Gram-negative bacteremia has been treated using IV antibiotics for 7 to 14 days, primarily based on expert opinion. The following discussion of the evidence supporting shorter treatment durations and opportunities for oral antibiotic therapy for uncomplicated Gram-negative bacteremia will better prepare the pharmacists in managing patients and in antibiotic stewardship.
Prevalence of Infectious Pathogen
Gram-negative bacteremia is most often secondary to another source of infection. Likely pathogens for secondary bacteremia are variable based on source of infection and location of onset. From the results of three studies, Table 2 depicts the likely pathogens for community, hospital, and ICU acquired Gram-negative bacteremia.8-10 Manzoni and colleagues looked at 2,924 different microorganisms from 16 different hospitals in northern Italy over the course of 2 years for community acquired gram-negative bacteremia.8 The study found that the majority of community acquired Gram-negative bacteremia were caused by cephalosporin susceptible Escherichia coli in this study population. The most likely source for bacteremia with this organism is a urinary-tract infection, although the study did not report sources of infection.8 There is an increase in more drug-resistant organisms when infection onset occurs in the hospital and intensive care unit setting setting.9,10 Shorr and colleagues looked at data from 6,697 patient from 59 different hospitals in the United States with hospital acquired Gram-negative bacteremia defined as a first positive blood cultures drawn >2 days after admission.9 The study found a wider variety of infectious pathogens than was found in the study of community acquired bacteremia with only 18% of infections due to Escherichia coli and 56% being from various spp. For ICU acquired Gram-negative bacteremia, 2 studies by Sligl and colleagues looked at 18,146 admissions over a 5-year period from 1999 to 2003 and an 8-year period from 2004 to 2012 seeing consistent occurrence rates of infectious pathogen, over the 13 year period.5,10 Swamy and colleagues looked at 406 cases of Gram-negative bacteremia and broke down source of infection by percent, the results can be found in figure 1.17 Swamy and colleagues found the majority of gram-negative bacteremia were due to a urinary source, which remains consistent to the other studies discussed. They found the majority of community acquired Gram-negative bacteremia were due to Escherichia coli. Based on prevalence data presented, the majority (up to 90% in community acquired) of Gram-negative bacteremia is the result of bacteria from the Enterobacterales (formerly Enterobacteriaceae) family such as E. coli, Klebsiella, Enterobacter, and Citrobacter.8
Intravenous vs Oral Treatment
Once an infection is suspected, empiric therapy is initiated, and cultures should be obtained. As blood cultures begin to result, one may begin targeting therapy to treat the source of infection. A clinician often does not know they are treating a bacteremia until the blood cultures result. Once a causative pathogen is identified, treatment considerations such as de-escalation to oral therapy can be considered. The transition of a patient’s antibiotic therapy from IV to oral is not only a cost avoidance measure for the hospital system and the patient, provided it is done appropriately. Conversion to oral antibiotics lowers the number IV administrations decreasing a patient’s risk for infection and often allows for earlier discharge. There is evidence to support patients transitioning to highly bioavailable antibiotics after 1 to 5 days of IV therapy for gram-negative bacteremia. Listed in table 3 are common highly bioavailable antibiotics listed from highest to lowest:
*IV dosing 400mg vs PO dosing 500mg accounts for decreased bioavailability
Uncomplicated MRSA bacteremia treatment must be IV for a duration of at least 14 days per IDSA guidelines.1 The transition to oral therapy for Gram-negative bacteremia can be considered for the treatment of Enterobacteriaceae. Two different studies found treatment failure for highly bioavailable antibiotics was 2% or less when evaluating uncomplicated Enterobacterales bacteremia of urinary source.12,13 One of the studies by Kutob and colleagues looked at the rate of treatment failure for 362 patients being treated with high, moderate, and low bioavailable oral antibiotics. The study showed treatment failure rates of 2% (n=106), 12% (n=179), and 14% (n=77), respectively. Treatment failure for the purpose of this study was defined as all-cause mortality or recurrent infection within 90 days of the initial episode of bacteremia. Levofloxacin was the only highly bioavailable antibiotics investigated, and all 3 groups received an average of 4.7 days IV therapy prior to oral conversion. These results are further bolstered by 2019 meta-analysis from Punjabi and colleagues, which investigated 2289 patients from 14 studies.16 The studies evaluated oral vs IV step-down therapy for Enterobacterales bacteremia. The analysis found 65% of patients transitioned to oral fluoroquinolone, 7.7% to TMP-SMX, and 27.2% to oral beta-lactam, and again showed overall treatment failure for transitioning patients to oral antibiotic was low when using a highly bioavailable antibiotic. The results did find that recurrence of infection occurred more often when transitioned to oral beta-lactam than fluoroquinolone (OR 2.15; 95% CI, 0.93-4.99), however inadequate dosing was cited as a possible reason for this finding. All of these studies evaluated transitioning patients to oral after day 3-4 of IV therapy. While these studies show positive results for fluoroquinolone efficacy in this setting, it is noteworthy that resistance amongst Enterobacterales to fluoroquinolones is increasing limiting their use. In addition, black box warnings around toxicities of these drugs make them less than optimal choices in many patients. Fortunately, newer data have shown positive results for alternative agents as well. Two retrospective studies found that the rate of treatment failure for oral beta-lactams were similar to oral fluoroquinolones.13,14 The first study by Rieger and colleagues, looked at 241 patients with uncomplicated urinary Enterobacterales bacteremia treated with oral antibiotics. The study found no statistically significant difference in treatment failure between IV only and IV to oral treatment (3.8% vs 8.2%; p=0.19). Treatment failure was defined as a change in antibiotic regimen due to worsening clinical status, escalation back to IV antibiotics from oral, or readmission for the same infection within 30 days of discharge. The primary oral regimens used were, ciprofloxacin (65.3%), oral beta-lactams (19%) and trimethoprim-sulfamethoxazole (9.1%).13 The second study was by Mercuro and colleagues.14 The study reviewed 224 patients with uncomplicated urinary Enterobacterales bacteremia comparing clinical success of oral beta-lactam step-down therapy vs oral fluoroquinolone. Rates of clinical success were found to be similar among both groups (86.9% vs 87.1%; p>0.05) with higher rates of therapy completion in the beta-lactam group (91.7% vs 82.1%; p=0.049).14 A prospective study by Sutton and colleagues released in 2020 was much larger and evaluated 4089 patients who received an oral beta-lactam compared with a highly bioavailable fluoroquinolone or trimethoprim-sulfamethoxazole (TMP-SMX). The study found a 30-day mortality rate of 3% (n=29) vs 2.6% (n=82) and a recurrence rate of 1.5% (n=14) vs 0.4% (n=12), respectively.15 Based on the findings of these studies, a highly bioavailable fluoroquinolone should be considered an adequate choice for step-down therapy for an uncomplicated Enterobacterales bacteremia of urinary source after at least 2-days IV therapy. In addition, it appears that in many cases an oral beta-lactam can be considered an acceptable, side-effect minimizing substitution to a highly bioavailable fluoroquinolone provided the dosing regimen is optimized based on pharmacokinetic parameters and the patient has completed 3-4 days of IV therapy.
Evidence for the use of oral antibiotics outside of treating Enterobacterales is lacking. Fluoroquinolones are the only oral agents with reliable activity against Pseudomonas aeruginosa due to high intrinsic resistance to oral beta-lactams. This means there are significantly less options for oral step-down therapy.16 As a result, there is an overall lack of evidence to support routine transition to oral therapy for MDR-bacteremia including Pseudomonas aeruginosa.16 However, in a study by Fabre and colleagues of 249 patients treated for uncomplicated urinary Pseudomonal bacteremia, 17 (6.8%) transitioned to an oral fluoroquinolone.17 A reported median time to transition was 5 days after initiating therapy. All 17 patients had source control, defined as the removal of infected hardware or devices, resolution of biliary or urinary obstruction, or drainage of infected fluid collections, and no difference in outcomes were reported. Thus, while routine transition of all patients with Pseudomonas bacteremia would not be recommended, the high bioavailability of fluoroquinolones along with the small retrospective study by Fabre and colleagues does support consideration of oral fluoroquinolones in uncomplicated urinary pseudomonal bacteremia where source control is achieved, and the patient has a rapid clinical response to antibiotics. The use of oral fluoroquinolone step-down therapy for pseudomonal bacteremia should be made on a case-by-case basis. There is also a lack of evidence to support the use of oral antibiotics for non-urinary source uncomplicated Gram-negative bacteremia. The meta-analysis by Punjabi and colleagues cited 6 studies which evaluated non-urinary sources of bacteremia in addition to urinary sources. These studies reported positive outcomes supporting the use of oral therapy for bacteremia of any source, but the results of these studies are likely skewed as the majority (>60%) of cases were secondary to a urinary tract infection.18
Considerations for Duration of Treatment
Once targeted therapy has been chosen for an infection, a proper duration of therapy must be determined to reduce excessive use of antibiotics and risk of adverse events. Whether a bacteremia is complicated or uncomplicated as well as the source of infection are the primary factors in determining treatment duration. If the infection is complicated, an extended duration of treatment of up to 14 days or more should be considered following resolution of complicating signs and symptoms.19 For uncomplicated Gram-negative bacteremia, the majority of cases are derived from a urinary infection with a catheter related source the second most common, and then unidentified source.19 In the study by Swamy and colleagues discussed above in which the majority of Gram-negative bacteremia cases were the result of a urinary tract infection, the achievement of clinical response at the end of therapy for short (7 days or less), intermediate (8 to 14 days) and long (more than 14 days) courses of treatment for gram-negative bacteremia showed no difference in clinical responses. (78.6% vs 89% vs 80.6%, respectively; p=0.2). In addition, the study failed to find a correlation between identified pathogen type, source of infection (urinary vs non-urinary), and time to defervescence (≤72 hours, >72 hours) with clinical failure at the end of therapy. However, the study was underpowered and patients with delayed clinical response may require longer durations of treatment.19 Another study by Yahav and colleagues compared 7 vs 14 days for uncomplicated Gram-negative bacteremia.20 The study, which looked at a 90 days composite of all-cause mortality, relapse, suppurative, or distant complications, found a 7 day duration to be non-inferior to a 14 day duration of treatment (45.8% vs 48.3%). The majority of patients had a urinary sourced infection (68%) caused by a Enterobacterales (90%).20
Unlike data surrounding IV to oral conversion, treatment durations for multi-drug resistant pathogens such as Pseudomonas aeruginosa or Acinetobacter baumannii may be reduced. Of the studies cited above recommending a reduced duration of 7 days for uncomplicated bacteremia, there was a relatively low percent of patients included with multi-drug resistant pathogens.19,20 The study by Yahav and colleagues only evaluated 28 (4.6%) patients with pseudomonal bacteremia and 2 (0.3%) patients with Acinetobacter bacteremia.20 The study by Swamy and colleagues only included 7% of patients treated for a pseudomonal bacteremia and 4% of patients treated with an Acinetobacter bacteremia.19 A retrospective study by Fabre and colleagues of 249 patients with uncomplicated Pseudomonas bacteremia found patients treated for approximately 10 days had similar outcomes to those treated with longer durations.17 There are too few patients in these studies with MDR Gram-negative bacteremia to recommend a reduced duration of therapy for this patient population.
The Use of Follow-up Cultures
Follow-up cultures are necessary for adequate treatment duration for Gram-positive bacteremia. In GNB, the utility of follow-up cultures is more ambiguous. Canzoneri and colleagues looked at 383 cases of GNB where follow-up cultures had been drawn and found positive follow-up cultures for a Gram-negative bacteria in 8 cases.21 Only one of the positive cultures was indicative of a possible treatment failure, suggesting follow-up cultures for uncomplicated GNB are not needed.
The treatment of a GNB can range from 7 to 14 days. For complicated GNB a full 14-day duration following resolution of complicating factors would be ideal, as the risk for recurrence is likely high. For MDR pathogens such as Pseudomonas or Acinetobacter, there is evidence to support a reduced duration of 10-days IV antibiotics for uncomplicated bacteremia. Enterobacterales can be treated with a short 7-day course of either IV treatment for non-urinary sourced bacteremia or oral step-down therapy for urinary sourced bacteremia. Provided the patient sees clinical improvement, the use of follow-up blood cultures is not needed for GNB. The flow sheet in figure 2 depicts when to consider treatment duration reductions and IV to oral conversion for Gram-negative bacteremia based on infectious pathogen, source of infection, and complications of bacteremia. The use of shorter oral antibiotic regimens when appropriate will aid in better antibiotic stewardship and patient care.
Figure 2: Treatment Duration Flowsheet
*Recommendations should be considered on a case-by-case basis