Empiric Choice of Antimicrobial Therapy in The Antibiotic-Conditioned Patient with Sepsis

Authors: Stanley Deresinski, M.D.

Antibiotic resistance is a growing plague in our healthcare facilities (Table 1).  While the possibility of resistance to one or more antimicrobial agents should be considered in all septic patients, those who have had recent antibiotic exposure are especially likely to be infected with resistant microorganisms.  This resistance may extend beyond the class of antibiotics which the patient had received.  Favorable outcomes in such patients requires prompt recognition of sepsis and the early empiric institution of effective antimicrobial therapy.  Choosing appropriate therapy in this setting is, of course, complicated, by the likelihood of resistance of the infecting pathogen to the antibiotics to which the patient has already been exposed, as well as to related antibiotic and, possibly, unrelated ones.  This article attempts to specifically address the choice of empiric antibiotic therapy in septic patients with recent or current antibiotic exposure. Prompt recognition of sepsis (Table 2), allowing immediate initiation of aggressive therapeutic interventions, is critical to patient outcome.  In some patients, the onset of sepsis may be insidious, leading to delayed recognition.  This is especially true in the elderly in whom such typical manifestations such as fever may be absent and in whom the only initial clue to the presence of sepsis may be an alteration in mental status.


The appropriate management of the septic patient requires aggressive attempts to identify its etiology and source, as well as evaluation of factors that are likely to affect the antimicrobial susceptibility pattern of the presumed pathogen(s).   For instance, previous exposure to particular antibiotics are known to be important risk factors for infection with resistant pathogens, such as methicillin-resistant Staphylococcus aureus (MRSA)Pseudomonas aeruginosa, or Stenotrophomonas maltophilia. Such resistance may extend to classes of antibiotics other than those to which the patient has been exposed (1,10,11,12).  Knowledge of organisms colonizing the patient prior to the onset of infection improves the likelihood of choosing appropriate empiric antibiotic therapy in the patient with bacteremia due to resistant Gram negative bacillary organisms (3). Consideration must be given to the immune competence of the patient, his or her exposure to healthcare facilities and activities, the likely source of infection, and any preexisting microbiological data, as well as factors likely to lead to intolerance to individual antibiotics (Table 3).

back to top


Bloodstream Infection

Recent recommendations from the American Society for Microbiology have addressed the optimal approach to the utilization of blood cultures, which should be obtained in all patients with sepsis (Table 4).  In addition, special procedures may be warranted in patients with suspected infection of vascular access devices. 

IV Catheter-Related Infection (2)

Further procedures are recommended for the patient with suspected infection of a central venous access device.  If the clinical setting dictates catheter removal (e.g., shock, local purulence), remove the catheter and perform semiquantitative or quantitative culture, as well as 2 peripheral blood cultures obtained by venipunctures.  In this setting, anaerobic cultures are not indicated - only aerobic cultures should be obtained.  Thus, each individual blood sample may be inoculated into two aerobic blood culture bottles.  If circumstances do not dictate immediate removal of catheter, one can do one of the following in order to definitively identify the catheter as the source of the infection:              

1.  Perform quantitative cultures on blood obtained from both the vascular access device and a peripheral vein: the presence of a central to peripheral ratio > 4:1 is indicative of catheter-related infection (CRI); OR           

2.   Determine the differential time to positivity between catheter blood and peripheral blood: CRI diagnosed if the difference is  > 2 hours, with the peripheral blood sample requiring the longer incubation. If, however, blood for culture is available only from the central venous catheter only, CRI infection can be diagnosed if the culture yields >100 CFU bacteria/ml or >25 CFU yeast cells/ml.

Lower Respiratory Tract Infection

There is no demonstrated value of bacterial cultures of respiratory secretions in the absence of clinical or radiographic evidence of lower respiratory infection in the evaluation of the septic patient (although knowledge of recently colonizing organisms may prove of value when the patient becomes septic). In addition, many pulmonary infiltrates in hospitalized patients have a non-infectious etiology. In the presence of pneumonia, attempts to obtain respiratory secretions for direct examination and culture should always be made, but these attempts should not inordinately delay the initiation of antibiotic therapy in the critically ill patient. If viral infection, such as influenza, is suspected, upper respiratory specimens may be obtained for rapid testing and/or culture.  Appropriate specimens, depending on the circumstances, may include expectorated sputum, or secretions obtained by endotracheal aspirate, bronchoalveolar lavage, or protected specimen brush. The latter three specimens should be cultured quantitatively or semiquantitatively. DFA testing of respiratory secretions for the presence of Legionella spp. may be of value. Suspected empyema fluid, which may require CT scanning for detection, should also be examined and cultured. Consideration may also be given, depending on the circumstance, to testing urine for the presence of antigens of S. pneumoniaeLegionella pneumophila serogroup 1, or evenHistoplasma capsulatum.  Serum antibody testing is likely to only have retrospective value, with the exception of testing for antibody to C. immitis, which should be performed when a relevant epidemiological history has been obtained. Testing serum for the presence of cryptococcal antigen may be indicated in some cases. In the severely immunocompromised host, infection with filamentous fungi, such as Aspergillus spp. may be the cause of pulmonary infection. While some believe that serum tests for serum galactomannan are useful in the diagnosis of invasive aspergillosis and that serum assays beta-glucan may have value for this and other fungal infections, definitive diagnosis often requires the use of invasive procedures. Pneumocystis jirovecii may be detected in expectorated sputum, often obtained by induction methods, but may also require the acquisition of specimens by bronchoscopy for detection. Particularly in immunocompromised patients, the possibility of infection with viruses such as RSV or influenza should be considered.

Urinary Tract Infection

The combination of leukocyte esterase and nitrite testing has a reasonable sensitivity and specificity in the diagnosis of urinary tract infection when the concentration of organisms in urine is high (13). The specificity of pyuria in patients with indwelling bladder catheters is likely to be low and such testing may be inadequately sensitive in the patient with low bacterial counts in the face of existing antibiotic therapy. Thus, urine culture should be obtained most septic patients. In addition, examination of a Gram stained smear of urine may provide some clues that help direct antimicrobial therapy, such as the presence of Gram positive bacteria or of yeast.  If yeast are seen on smear or are otherwise suspected to be present, incubation of the urine should continue for at least 48 hours, since growth of Candida glabrata may be delayed. Sterile pyuria is commonplace in patients with urinary catheters.


Helical CT scanning of the chest may be useful in excluding pulmonary embolism as a cause of pulmonary infiltrates and, with a different protocol, allow for the detection of unsuspected infiltrates. CT may also provide clues to the etiologic pathogen, as in the case of pulmonary aspergillosis. CT scanning may also identify the presence of pleural effusionsand empyema, as well as unsuspected intrapulmonary or mediastinal abscesses.

CT scanning is the preferred methodology for examining the abdomen and pelvis in patients with suspected intraabdominal pathology, although ultrasound examination may be preferred in the identification of acalculous cholecystitis. In the patient with suspected paranasal sinusitis, CT scanning is also the procedure of choice.  Nuclear medicine procedures may be of value in some individual cases.

back to top


General Management

The management of the patient with sepsis involves much more than antibiotic administration. Additional measures are critical to a favorable clinical outcome. Thus, source control is of importance in many patients. In those with severe sepsis or septic shock, additional critical measures may include early hemodynamic goal-directed resuscitation with fluid resuscitation and administration of vasoactive agents, administration of moderate dose corticosteroids and, in some patients, the use of recombinant activated protein C (5).

Source Control

Closed space infections require drainage and necrotic tissue requires debridement or resection and foreign bodies may require removal. Table 5 lists some infection source sites that may require such intervention.

Implications of Prior Antibiotic Exposure for Antibiotic Choice

The initial choice of empiric antibiotic therapy is critical to a favorable outcome in the patient with life-threatening infection (Table 6). An important predictor of inappropriate choice is the prior exposure of the patient to antibiotic therapy because of the likelihood that the etiologic agent is resistant to one or more antibiotics. An increased risk of antibiotic resistance exists for patients whose prior antibacterial exposure is as distant as three months. Prior antibiotic exposure may lead to colonization and infection with resistant organisms most often as a consequence of its selective pressure on previously targeted or non-targeted organisms. Examples of the latter include the emergence of enterococci as important pathogens in patients who have received cephalosporins and Candida infections in patients who have received any broad spectrum antibacterial therapy. Some classes of antibiotics may also increase the mutation rate of exposed organisms, and thus actively drive resistance, rather than passively serving as a selective force. This knowledge, as well as empirical experience, and the knowledge of the importance of the appropriate choice of such therapy to patient outcomes, dictate the choice of empirical anti-infective therapy in antibiotic exposed patients. To an important extent, however, choices are limited by our available therapeutic armamentarium.


It is important to recognize that pulmonary infiltrates in intensive care patients have a non-infectious etiology in many, possibly a majority. These etiologies include pulmonary contusion in trauma patients, pulmonary hemorrhage, infarction, and pulmonary edema.   

Recent recommendations for the treatment of patients take cognizance of the effects of recent antibiotic exposure on the risk of infection with antibiotic resistant pathogens (9). Table 7 lists the most commonly isolated bacteria in patients with hospital-acquired and ventilator-associated pneumonia.  These organisms and their susceptibility pattern dictate the initial choice of antibacterial therapy.  This initial empiric therapy should include an anti-pseudomonal beta lactam together with either an aminoglycoside or an anti-pseudomonal fluoroquinolone (Tables 89).  If the presence of MRSA is likely, linezolid should be used.  If ESBL-producing Gram negative bacilli are prevalent in the institution, a carbapenem is preferred over other beta-lactams as well as over fluoroquinolones. If legionellosis is suspected, a fluoroquinolone should be used rather than an aminoglycoside. If multidrug resistance in local Acinetobacter spp. includes resistance to carbapenems, a polymyxin (polymyxin B or colistin) or tigecycline may be necessary.

Intraabdominal Infection

Source control is critical to a favorable outcome in patients with intraabdominal infection.  Succeessful and urgent source control may require an open or laparoscopic procedure for abscess drainage or debridement or resection of infected and/or necrotic tissue.  Inadequate source control is likely the most likely reason for an unsuccessful outcome of treatment.   

The patient with intraabdominal infection with recent or current antibiotic exposure should be considered to be infected with organisms resistant those antibiotics. These infections may be caused by flora that includes such organisms as  Pseudomonas aeruginosaEnterobacter speciesProteus speciesmethicillin-resistant Staphylococcus aureus,enterococci, and Candida species (14).  The presence of anaerobic organisms, including those of the Bacteroides fragilis group, although less frequently encountered than in primary infection, must also be considered.  Thus the distinct difference in choice of antibiotic regimen when compared with the recommendations for pneumonia is the inclusion of agents with activity against anaerobes, particularly of the B. fragilis group. As a consequence, the initial choice of empiric therapy may involve the administration of agents in combination, such as the following regimens:   

               Imipenem or meropenem +  an aminoglycoside + vancomycin

               Piperacillin/tazobactam +  an aminoglycoside  

In addition, if vancomycin resistant enterococci are prevalent in the institution, linezolid may be preferred over vancomycin. If yeast are seen on gram stain, or if the patient is known to be colonized at multiple sites with Candida spp., empiric administration of antifungal therapy may be considered. This may consist of fluconazole if the patient has not recently been exposed to this drug and if fluconazole-resistant Candida spp. are not locally prevalent.  Alternatives include an amphotericin B preparation or an echinocandin. Therapy should subsequently be modified based upon microbiological results as they become available.

Bloodstream Infection

In the patient with suspected bloodstream infection resulting from an infected central venous access device, the device should, in most circumstances, be removed. The initial choice of antibiotics should provide coverage of the pathogens listed in Table 8, while taking into account other recent or current antibiotic therapy. Staphylococci predominate among the causes of hospital-acquired bloodstream infection. An appropriate regimen might consist of a fluoroquinolone (ciprofloxacin or levofloxacin) together with an agent active against methicillin-resistant staphylococci (Tables 89). Consideration may also be given  to the addition of an another agent to broaden the spectrum against additional antibiotic resistant Gram negative bacilli.  If the patient is heavily colonized with Candida spp., or has clinical clues (e.g., chorioretinal lesions) suggestive of candidemia, antifungal therapy should be initiated.


If the urinary tract is believed to be the source of sepsis and a gram stain of urine reveals Gram positive cocci, the chosen antibiotic regimen should provide coverage of enterococci, including VRE is this organism is prevalent in the institution or if the infection has occurred in the face of  vancomycin therapy. If the morphology of the organism suggests the presence of staphylococci, especially in the patient with an indwelling bladder catheter, therapy should provide coverage of MRSA.

Sepsis of unknown source

The patient with sepsis of unknown source presents an especially difficult problem, because the potential range of pathogens is so broad. The extent and aggressiveness of antibiotic coverage depends, to perhaps a greater extent than with more defined clinical illnesses, upon the severity of illness, as well as the extent of prior antibiotic exposure.

back to top


In patients with complicated skin and skin structure infections, including surgical site infections not responding to antibiotic therapy, the possibility of necrotizing fasciitisshould be urgently explored. Skin lesions, including cellulitis, in severely immunocompromised patients may be due to fungi. Not everything that looks like and infection is: progressive ulcerative lesions, including ones appearing within a surgical wound, may be due to pyoderma gangrenosum.  

Clostridium difficile is a frequent cause of the systemic inflammatory response syndrome in patients who have received antibiotics (as well as in some who have not). Not all patiends with C. difficile-associated disease (CDAD) have diarrhea. Some may have ileus.  CDAD should be suspected in any febrile patient with marked leukocytosis.  Orally administered vancomycin is the recommended initial treatment for patients with severe CDAD.  If oral administration is not applicable, metronidazole should be given intravenously.

Subsequent Steps

Close clinical and microbiological monitoring of the patient is essential.  Antimicrobial therapy should be adjusted based on clinical response and microbiological data as it becomes available.  If possible, antimicrobial therapy should be deescalated.  If the patient does not appear to be responding to the chosen empiric therapy, a thorough reevaluation is indicated to identify other causes of the sepsis syndrome, both in terms of alternative pathogens and unsuspected sources of infection.  In addition, a search for non-infectious etiologies may be warranted.  In the responding patient, the duration of antibiotic therapy should be limited to that necessary.  In the case of ventilator-associated pneumonia, antibiotics may be discontinued on the eighth day in most cases, although some infections due to aerobic non-fermenting Gram negative bacilli (e.g., P. aeruginosa) may require a longer duration of therapy (9).  In intra-abdominal infections, antibiotic therapy should be continued until clinical signs of infection have resolved, white blood cell count is normal, and gastrointestinal function has returned (14).  Similarly, the duration of treatment of skin and skin structure infection, rather than depending upon some arbitrarily chosen number of days, should depend upon the resolution of clinical signs of infection - i.e., the appearance of the surgical wound.  It is generally recommended that antibiotics be administered for uncomplicated bloodstream infection for 10 to 14 days.  The duration of antibiotic administration in patients with complicated bloodstream infection depends upon the nature of the complication (7).

back to top


1. Alyaseen SA, Piper KE, Rouse MS, Steckelberg JM, Patel R. Selection of cross-resistance following exposure of Pseudomonas aeruginosa clinical isolates to ciprofloxacin or cefepime.  Antimicrob Agents Chemother 2005; 49:2543-5. [PubMed] 

2. Baron EJ, Weinstein MP, Dunne WM Jr, et al.  2005.  Cumitech 1C, Blood Cultures.  Coordinating ed., E.J. Baron.  ASM Press, Washington, D.C.

3. Blot S, Depuydt P, Vogelaers D, Decruyenaere J, De Waele J, Hoste E, Peleman R, Claeys G, Verschraegen G, Colardyn F, Vandewoude K. Colonization status and appropriate antibiotic therapy for nosocomial bacteremia caused by antibiotic-resistant gram-negative bacteria in an intensive care unit.  Infect Control Hosp Epidemiol 2005; 26:575-9. [PubMed] 

4. Chastre J, Fagon JY.  Ventilator-associated pneumonia.  Am J Resp Crit Care Med 2002; 165:867-903. [PubMed] 

5. Dellinger RP, Carlet JM, Masur H, Gerlach H, Calandra T, Cohen J, Gea-Banacloche J, Keh D, Marshall JC, Parker MM, Ramsay G, Zimmerman JL, Vincent JL, Levy MM; Surviving Sepsis Campaign Management Guidelines Committee. Surviving Sepsis Campaign guidelines for management of severe sepsis and septic shock.  Crit Care Med 2004; 32:858-73. [PubMed] 

6. Gaynes R, Edwards JR, and the National Nosocomial Infections Surveillance System.  Overview of nosocomial infections caused by Gram-negative bacilli.  Clin Infect Dis 2005; 41:848-54. [PubMed] 

7. Mermel LA, Farr BM, Sherertz RJ, Raad II, O'Grady N, Harris JS, Craven DE; Infectious Diseases Society of America; American College of Critical Care Medicine; Society for Healthcare Epidemiology of America. Guidelines for the management of intravascular catheter-related infections.  Clin Infect Dis 2001; 32:1249-72. [PubMed] 

8. National Nosocomial Infections Surveillance (NNIS) System Report, data summary from January 1992 through June 2004, issued October 2004.  Am J Infect Control 2004; 32:470-85. [PubMed] 

9. Niederman MS, Mandell LA, Anzueto A, Bass JB, Broughton WA, Campbell GD, Dean N, File T, Fine MJ, Gross PA, Martinez F, Marrie TJ, Plouffe JF, Ramirez J, Sarosi GA, Torres A, Wilson R, Yu VL; American Thoracic Society. Guidelines for the management of adults with community-acquired pneumonia. Diagnosis, assessment of severity, antimicrobial therapy, and prevention.  Am J Respir Crit Care Med. 2001; 163:1730-54. [PubMed] 

10. Nouer SA, Nucci M, de-Oliveira MP, Pellegrino FL, Moreira BM.  Risk factors for acquisition of multidrug-resistant Pseudomonas aeruginosa producing SPM metallo-beta-lactamase.  Antimicrob Agents Chemother 2005; 49:3663-7. [PubMed] 

11. Paterson DL.  "Collateral damage" from cephalosporin or quinolone antibiotic therapy. Clin Infect Dis. 2004;38 Suppl 4:S341-5. [PubMed] 

12. Radberg G, Nilsson LE, Svensson S.  Development of quinolone-imipenem cross resistance in Pseudomonas aeruginosa during exposure to ciprofloxacin.  Antimicrob Agents Chemother 1990; 34:2142-7. [PubMed] 

13. Ramakrishnan K, Scheid DC.  Diagnosis and management of acute pyelonephritis in adults.  Am Fam Physician. 2005;71:933-42. [PubMed] 

14. Solomkin JS, Mazuski JE, Baron EJ, Sawyer RG, Nathens AB, DiPiro JT, Buchman T, Dellinger EP, Jernigan J, Gorbach S, Chow AW, Bartlett J; Infectious Diseases Society of America. Guidelines for the Selection of anti-infective agents for complicated intra-abdominal infections.  Clin Infect Dis 2003; 37:997-1005. [PubMed] 

15. Wisplinghoff H, Bischoff T, Tallent SM, Seifert H, Wenzel RP, Edmond MB. Nosocomial bloodstream infections in US hospitals: analysis of 24,179 cases from a prospective nationwide surveillance study.  Clin Infect Dis 2004;39:309-17. [PubMed]

back to top



Table 1. Antibacterial resistance in the U.S. January 1998 – June 2004 (ICARE/AUR) (8)

Organism Resistant to: Mean % - ICU Mean % - non-ICU
S. aureus Methicillin 52.9% 46.0%
Staphylococcus – coagulase negative Methicillin 76.6% 65.7%
Enterococcus spp. Vancomycin 13.9% 12.0%
P. aeruginosa Levofloxacin 35.3% 30.5%
P. aeruginosa Imipenem 19.1% 12.3%
P. aeruginosa Ceftazidime 13.9% 8.8%
P. aeruginosa Piperacillin 17.5% 11.6%
Enterobacter spp. 3rd Gen Ceph 27.7% 21.0%
Enterobacter spp. Carbapenem 0.7% 1.0%
K. pneumoniae 3rd Gen Ceph 6.2% 5.80%
E. coli 3rd Gen Ceph 1.3% 1.50%
E. coli Fluoroquinolone 7.3% 8.20%
S. pneumoniae Penicillin 18.9% 18.2%
S. pneumoniae Ceftriaxone 7.5% 7.60%

Table 2. Definition of Sepsis

The standard definition of sepsis requires the presence of a systemic inflammatory response syndrome (SIRS) together with evidence of infection.  SIRS is characterized by the presence of two or more of the following: 

• Temperature >38°C or <36°C

• Heart rate >90 beats per minute

• Respiratory rate >20 breaths per minute or PaCO2 <32 mm Hg

• WBC >12,000 cell/mm3, <4000 cells/mm3, or >10% immature forms

Table 3. Evaluation of the Patient: General Information

In selecting empiric antibiotic therapy, the clinician should seek answers to the following:

• Is the patient immunocompromised and, if so, what is the nature of their immunocompromise?

• Has the patient been hospitalized or resided in a long-term care facility in the previous three months?

• Did the current episode of sepsis occur while the patient was hospitalized?

• Has the patient received an antibiotic within the previous three months?  If so, which ones and for how long?

• What is the likely source of infection? 

• What invasive procedures has the patient recently undergone?

• What devices (intravascular, bladder, etc.) have been utilized in the care of the patient?

• What microbiological data (colonization or infection) are available from cultures of recently obtained specimens from the patient?

• What is known of the local (community, long term care facility, hospital, hospital unit) microbial ecology with regard to epidemic organisms and antimicrobial resistance patterns?

• Is the patient allergic to any antibiotics?  If so, what was the manifestation of the allergic reaction?

• Does the patient have underlying organ system abnormalities that may predispose to early onset adverse reactions to particular antimicrobials?

• Does a directed physical examination provide evidence for the source or metastatic foci of infection?

Table 4. American Society for Microbiology Blood Culture Recommendations


• For skin antisepsis, chlorhexidine products and tincture of iodine appear to be equivalent and both are superior to povidone-iodine. 

• Two to four blood cultures are necessary for optimal detection of bacteremia and fungemia.  In the setting of sepsis with an urgency to initiating antimicrobial therapy, two specimens provide adequate sampling. 

• “No single commercially available system or culture medium has been shown to be best suited for the detection of all potential blood pathogens.”

• The use of antibiotic-binding resins or activated charcoal may improve the yield of microorganisms in patients receiving antibiotics, but at the price of an increased rate of contamination.

• The use of two aerobic bottles in place of one aerobic and one anaerobic bottle may be beneficial in cases in which anaerobic bacteremia is unlikely. 

• A laboratory policy mandating a second blood culture (if only one has been ordered) as a reflexive test is warranted.

• Blood obtained for culture from vascular access devices should always be paired with another sample obtained by venipuncture. 

• The clinical status of patient should be the primary guide to the timing of blood cultures – there is no evidence that particular (or any) intervals between sampling is beneficial.

• Twenty to thirty ml of blood should be obtained per blood culture.

• The blood should be placed in an amount of liquid medium sufficient to achieve a 5- to 10-fold dilution.

• With the use of automated continuous monitoring systems, 5 days of incubation is adequate, including when infection with organisms such as Brucella or the HACEK group are suspected.

Table 5. Examples of source control.

Infection Procedure
Thoracic empyema Tube drainage
Purulent pericarditis Tube drainage, pericardectomy
Mediastinitis Drainage, debridement
Vascular access device Removal
Septic phlebitis Vein resection
Soft tissue abscess Incision and drainage
Necrotizing fasciitis Debridement (usually repeated)
Myonecrosis Resection, amputation
Intraabdominal abscess Percutaneous or surgical drainage
Infected pancreatic necrosis Debridement, resection
Intestinal infarction, perforation Resection
Cholangitis with obstruction Relief of obstruction
Acute cholecystitis Cholecystectomy
Pyelonephritis with obstruction Relief of obstruction
Renal, perirenal abscess Percutaneous, surgical drainage
Pyometritis Removal of intrauterine device, resection

Table 6. Some Implications of prior antibiotic therapy.

Antibiotic(s) Administered Some Bacterial Organisms Not Covered*
Levofloxacin Anaerobes, MRSE, MRSA, S. maltophilia, Enterococcus spp., many P. aeruginosa
Ciprofloxacin Anaerobes, MRSE, MRSA, S. maltophilia, Enterococcus spp., Streptococcus spp., manyP. aeruginosa
Ceftriaxone Anaerobes, MRSE, MRSA, P. aeruginosa, S. maltophilia, Enterococcus spp., ESBL producers, AmpC producer, many P. aeruginosa
Cefepime Enterococcus spp., ESBL producers, many P. aeruginosa
Piperacillin/tazobactam MRSE, MRSA, S. maltophilia, VRE
Imipenem/cilastatin or Meropenem MRSE, MRSA, S. maltophiliaE. faecium
Cefepime + (Ciprofloxacin or an Aminoglycoside) Anaerobes, MRSE, MRSA, S. maltophilia, Enterococcus spp.
Piperacillin/tazobactam + (Ciprofloxacin or an Aminoglycoside) MRSE, MRSA, S. maltophilia, VRE
Imipenem/cilastatin +  (Ciprofloxacin or an Aminoglycoside) MRSE, MRSA, S. maltophilia, VRE, E. faecium

*This following table provides some insight into that process, but it does not include consideration of all antibiotic resistant organisms.  Examples of omissions include Acinetobacter spp., (because of its highly varible resistance patterns),  and Candidaspp., each of which should always be considered as potentially etiologic in these patients.

Table 7. Hospital-acquired and ventilator-associated associated pneumonia (4,6).

Organism Frequency
Pneumonia (Gaynes) Ventilator-Associated Pneumonia (Chastre)
Pseudomonas aeruginosa 18.1% 24%
Acinetobacter spp. 6.9% 8%
Stenotrophomonas maltophilia   2%
Enterobacteriaciae   14%
Haemophilus spp.   10%
Staphylococcus aureus 27.8% 20%
Streptococcus spp.   8%
Streptococcus pneumoniae   4%
Anaerobes   1%
Fungi   1%
K. pneumoniae 7.2%  
Enterobacter spp. 10.0%  
S. marcescens 4.7%  

Consider a viral etiology (e.g., RSV or influenza), particularly in immunocompromised patients.

Table 8. Antibiotic Matrix.

Class/Target                                                                Antibiotics
Anti-Pseudomonal b-lactam Cefepime Piperacillin/tazobactam Imipenemmeropenem    
Antipseudomonal Fluoroquinolone Ciprofloxacin Levofloxacin      
Antipeudomonal Aminoglycoside Tobramycin Amikacin Gentamicin    
Anti-MRSA Linezolid Daptomycin Vancomycin Quinupristin/dalfopristin Tigecycline
Anti- Enterococcus(not VRE) Ampicillin; piperacillin/ tazobactam Vancomycin Linezolid Daptomycin Tigecycline
Anti-B. fragilis Metronidazole, Piperacillin/tazobactam Imipenem or meropenem Tigecycline  
Anti-VRE Daptomycin   Linezolid  Quinupristin/dalfopristin*  Tigecylcline  
Anti-MDR-Acinetobacter Imipenem; meropenem Sulbactam Polymyxin B; Colistin Tigecycline  
Anti- S. maltophilia Trimethoprim/ sulfamethoxazole Ticarcilliin/clavulanic acid Ceftazidime Levofloxacin Tigecycline
Anti-Candida Fluconazole Caspofungin;micafungin;anidulafungin Amphotericin B Voriconazole  

*Not active vs. E. faecalis

Table 9. Empiric choice of antibiotic therapy in the antibiotic-conditioned septic patient.

    HAP/VAP 1 Intra-Abdominal Bloodstream Unknown
Anti-Pseudomonal ß-Lactam 2,3   + 4 + +
Anti-Pseudomonal Aminoglycoside or Fluoroquinolone   + +/- +/- +/-
Anti-MRSA and/or Anti-Enterococcal       + +/-
     Vancomycin or     +/- 5    
     Linezolid or   +/-      
     Daptomycin       +  
Anti-Candida   +/-6 +/-7 +/-7 +/-8

(1)If Stenotrophomonas maltophilia is suspected, e.g., in the patient colonized with the organism or if the infection occurred in the face of carbapenem therapy, usetrimethoprim/sulfamethoxazole.

(2)If ESBL-producing Gram negatives are prevalent, a carbapenem is indicated.

(3)If multidrug-resistant Acinetobacter spp. is prevalent, a carbapenem is a good choice.  If the local Acinetobacter spp. are also resistant to carbapenems, colistin may be used; tigecycline is also often active in vitro.

(4)If cefepime is chosen, metronidazole should be added to the regimen.

(5)If VRE is prevalent, use linezolid.  If linezolid resistance also prevalent, use daptomycin.

(6)If Aspergillus infection is suspected, use voriconazole with or without caspofungin.

(7)If patient has not received fluconazole and if the frequency of recovert if fluconazole-resistant yeasts is minimal in the hospital, fluconazole may be used.  Otherwise, caspofungin or lipid-associated amphotericin B is recommended.  

(8)If patient at risk of filamentous fungal infection, voriconazole or lipid-associated amphotericin B are indicated, otherwise, see footnote (7).


Masashi Narita, M.D.: Sterile Pyuria

O'Grady NP, et al.  Guidelines for evaluation of new fever in critically ill adult patients: 2008 update from the American College of Critical Care Medicine and the Infectious Diseases Society of America". Crit Care Med 2008;36(4):1330-1349.

Solomkin JS, Mazuski JE, Bradley JS, Rodvold KA, Goldstein EJC, Baron EJ, O'Neill PJ, Chow AW, Dellinger EP, Eachempati SR, Gorbach S, Hilfiker M, May AK, Nathens AB, Sawyer RG, Bartlett JG.  Diagnosis and Management of Complicated Intra-Abdominal Infection In Adults and Children: Guidelines by the Surgical Infection Society and the Infectious Disease Society of America.  Clin Infect Dis 2010;50:133-164.

Vincent JL.  Is the current management of severe sepsis and septic shock really evidence based?  PLoS Medicine 2006;3(9):1488-1491.

No longer recommended? Steroids and insulin for septic shock. N Engl J Med 2008.

Guided Medline Search For Recent Reviews


Clinical Manifestations