Antibiotic treatment of sepsis

ANTIMICROBIAL THERAPY I

0025-7125/95 $0.00 + .20

ANTIBIOTIC TREATMENT OF SEPSIS Burke A. Cunha, MD

Bactcremia describes the presence of bacteria in blood. Implicit in this definition is that the bacteremia is transient and the bacteria are not actively multiplying in the blood stream. Septicemia is the term used for active bacterial replication in the blood stream, which may associate with systemic manifestations. Septicemia is accompanied by various degrees of organ dysfunction, the extent of which depends on inoculum size, virulence of the organism, underlying medical diseases, and the host's defense status. Patients with septicemia, therefore, vary in their clinical presentation and prognosis as a function of the factors mentioned. Definition of the clinical septicemia state is achieved when septicemia is accompanied by various degrees of organ dysfunction. 1. 3. 4, 12, 17, 26, 28 Unfortunately, with the advent of a variety of antimediator interventions that have become experimentally available during the last decade, a variety of definitions have been introduced primarily to broaden patient eligibility for these trials, The introduction of such terms as sepsis syndrome, systemic inflammatory response syndrome (SIRS), or SIRS/sepsis with or without multiple organ dysfunction syndrome (MODS) has only served to include a large population of nonsepticemic patients that are critically ill from a variety of causes. The problem with the use of such terms is the exclusion of septicemia as part of the definition, e.g., evidence of septicemia is not a criterion for inclusion. Excluding blood culture positivity or evidence of septicemia as a criterion results from the misinterpretation of clinical data from the Systemic Sepsis Cooperative Study Group, in which fewer than 50% of the pa-

Fro1J1 the Infectious Disease Division, Winthrop-University Hospital, Mineola; and the State University of New York School of Medicine, Stony Brook, New York

MEDICAL CLINICS OF NORTH AMERICA VOLUME 79 • NUMBER 3 • MAY 1995

551

552

CU"HA

tients had positive blood cultures. Timing and procedures were not uniform in this multicenter study. By excluding proof of septicemia by blood culture positivity, a variety of noninfectious disorders associated with temperature elevations and various degrees of organ dysfunction have been termed SIRS, which is a clinically unhelpful designation.5, 10,11,13,19,21,24,25 In the absence of septicemia, critically ill patients from a variety of causes, e.g., myocardial infarction, gastrointestinal hemorrhage, acute pancreatitis, and massive trauma, frequently have the physiologic parameters of sepsis syndrome. Because these hemodynamics and physiologic manifestations mimic sepsis but are not infectious disease based, they may be termed pscudoscpsis rather than grouped loosely based on severity as sepsis syndrome/SIRS. Various researchers have indicated that what is needed is a means to diagnose septicemia accurately before blood culture results are available. One way to do this at the present time is by acridine staining of buffy coat preparations in patients in whom septicemia is suspected. Although buffy coat smears are not always positive, at least results are rapidly available, and when positive, they define septicemia. 1R Blood culture positivity is a function of timing in the volume of blood drawn. Properly obtained and timed blood cultures should result in blood culture positivity in more than 90% of septicemic patients. Active bacterial multiplication defines septicemia, and methods of detecting microbial replication should be part of the inclusion criteria for studies. Much needless confusion could be avoided by eliminating imprecise definitions and severity scoring systems that exist primarily for study purposes and do not change the clinical approach or outcome in septicemic patients. I, 3, 5, 10,24

CLINICAL APPROACH TO THE PATIENT WITH SEPSIS

Patients presenting with septicemia and associated organ dysfunction should have cultures taken of relevant body fluids in addition to blood cultures, The clinical presentation of septicemia has been well described and is an adequate working diagnosis. During this initial phase before treatment is initiated, the clinician should eliminate conditions that physiologically or hemodynamically mimic sepsis (pseudosepsis), e,g., myocardial infarction, adrenal insufficiency, gastrointestinal hemorrhage, pulmonary emboli, and acute pancreatitis. By identifying patients with pseudosepsis, needless and potentially dangerous administration of unnecessary antibiotics may be avoided (Table 1).6,17,24, 2H After conditions simulating sepsis are eliminated, the clinician should be able to identify the organ system involved as the initial source or focus of infection, Patients presenting with septicemia and shock, e.g" septic shock, nearly always have a gastrointestinal, genitourinary, or pelvic source if an intravenous line infection is not a consideration. Once the presumed infectious focus is localized to a particular organ system, the clinician should determine whether the condition in the organ system

ANTIBIOTIC TREATMENT OF SEPSIS

553

Table 1. PSEUDOSEPSIS NONINFECTIOUS CONDITIONS MIMICKING SEPSIS

Hemodynamic mimics of sepsis Acute pancreatitis Anaphylaxis Spinal cord injury Adrenal insufficiency Clinical mimics of sepsis Hemorrhage Pulmonary embolism Myocardial infarction Pancreatitis Diabetic ketoacidosis with abdominal crisis SLE flare with abdominal crisis Ventricular pseudoaneurysm Massive aspiration/atelectasis Systemic vasculitis Hypovolemia SLE = Systemic lupus erythematosus. Data from Cunha BA: Sepsis and its mimics. Intern Med 13:48, 1992.

is one usually associated with septicemia/ septic shock. The esophagus, stomach, and small bowel except under unusual circumstances are rarely implicated in septic shock, but conditions involving the colon, e.g., diverticulitis, colonic perforation, and pericolonic abscess, are commonly associated with septicemia/septic shock. Patients with bacterial community-acquired pneumonias in normal hosts, for example, rarely, if ever, present in septic shock. (Fulminant pneumococcal sepsis in an asplenic and Klebsiella pneumonia in an alcoholic are recognized exceptions.) Antimicrobial therapy is based on either a likely pathogen(s) presenting in a clinical setting, e.g., the syndromic approach, or the usual resident flora of the involved organ. The microbes of the infected organ, not the nature of the organ insult or the severity of the illness, determine proper antibiotic coverage. Whether the patient is septicemic with minimal organ involvement or has multiple organ dysfunction extending to and including septic shock, antibiotic therapy is based on the site of infection and its usual flora. Novel therapeutic modalities, e.g., antimediator therapy including polyclonal antiendotoxin anticore antibodies, monoclonal antiendotoxin, antitumor necrosis factor-a therapies, and interleukin-l receptor antagonist therapies, have been useful in the basic science understanding of the role of inflammatory mediators in patients with an inflammation/infection but have not been shown to be clinically effective and show little future promise. 10, 12, 23 Because untreated septicemia may eventuate in septic shock, the critical clinical intervention remains appropriate antimicrobial therapy administered as early as possible. Because the use of corticosteroids in the treatment of septic shock has been as disappointing as antimediator therapy, the cornerstone of treatment remains antimicrobial therapy directed against the flora of the organ system primarily involved in the infectious process (Table 2).10,20,23

554

CUNHA

Table 2. ORGAN SYSTEM INVOLVEMENT AND CONDITIONS ASSOCIATED WITH SEPSIS Commonly Associated with Sepsis

Not Commonly Associated with Sepsis

Gastrointestinal source Liver Gallbladder Colon Intra-abdominal abscess Intestinal obstruction Terminal small bowel obstruction Instrumentation Genitourinary source Pyelonephritis Intranephritic/perinephritic abscess Stones Urinary tract obstruction Prostatitis/abscess Instrumentation of bacteriuric patient Renal insufficiency Pelvic source Peritonitis Pelvic abscess Intravascular source IV line sepsis Infected prosthetic device Acute bacterial endocarditis

Gastrointestinal source Pancreatitis Esophagitis Gastritis Small bowel disorders Genitourinary source Urethritis Cystitis* Cervicitis Vaginitis Catheter-associated bacteriuria Upper respiratory source Pharyngitis Sinusitis Bronchitis Otitis Lower respiratory source Community-acquired pneumonia (normal host) Lung abscess Skin/soft tissue source Osteomyelitis Uncomplicated wounds Cardiovascular source Peripheral vascular disease Subacute bacterial endocarditis Central nervous system source Meningitis (nonmeningococcal/without DIC) Brain abscess

'Except in nonleukopenic compromised hosts, e.g., patients with diabetes mellitus or on steroids. DIC ~ Disseminated intravascular coagulation. Oata from Cunha BA: Sepsis and its mimics. Intern Med 13:48, 1992.

SELECTION OF APPROPRIATE ANTIBIOTIC COVERAGE BY ORGAN INVOLVEMENT

Antibiotics selected for the septicemic patient should be appropriate to cover the presumed pathogens as determined by the primary site of organ involvement and relevant epidemiologic factors, i.e., whether the infection is acquired in the community or nosocomially. Patients with septicemia may present with a spectrum of disease eventuating in septic shock, making early parenteral therapy with one or more antibiotics given in the full therapeutic doses essential. Fully recommended doses of the antibiotic selected should be administered because organ perfusion may be suboptimal during infection, and there are few data available on the pharmacokinetics and disposition of antibiotics in critically ill individuals. Intravenous administration is optimal, and intramuscular administration should be avoided. Oral therapy is not an option in the critically ill patient. Adequate blood levels are easily achieved with intravenously administered antibiotics, and attention should be paid to

ANTIBIOTIC TREATMENT

or

SEPSIS

555

the tissue concentrations of the antibiotic at the primary site of infection. Many patients with a gastrointestinal, genitourinary, or pelvic source of infection have processes, e.g., abscesses or perforations, requiring surgical intervention. Surgery indicated should be performed as soon as possible and is essential to control or eradicate the infection.'" 14, 1.0 Antibiotic dosing should take into account the patient's hepatic and renal function. Patients with renal insufficiency may be given renally eliminated antibiotics in a dosage reduced proportionally to the degree of renal dysfunction. Alternately, hepatically eliminated antibiotics with the same spectrum may be given to patients with renal insufficiency. Patients with hepatic insufficiency may receive either renally eliminated antibiotics with the proper spectrum or hepatically eliminated antibiotics at reduced dosage. In patients with severe hepatic and renal dysfunction, antibiotics with a wide toxic-to-therapeutic ratio should be used in reduced dosage. Dosing regimens should take into account whether the antibiotic kills by time-dependent kinetics, e.g., f3-lactam antibiotics, or concentration-dependent kinetics, e.g., aminoglycosides. For this reason as well as cost-effectiveness, metronidazole for intra-abdominal sepsis should be given as 1 g intravenously every 12 hours rather than on an every-6-hours basis. Little is to be gained by exceeding the recommended full dose of an antibiotic in the critically ill patient, but underdosing should be avoided as clinical failure may result because the distribution of antibiotic in these individuals is not certain. There is no optimal regimen for treating septicemic patients; antibiotics that are effective against the offending microorganisms are all equally efficacious as long as the organism is sensitive in vivo and the antibiotic is given intravenously in full therapeutic doses (Table 3).2,8,22,27 For most infections, mono therapy is more than adequate, and there is no need to add "additional anti-gram-negative coverage" to a perfectly adequate single-drug regimen. The indications for double-drug therapy are relatively few and include the treatment of serious Pseudomonas aeruginosa infections, the empiric treatment of the febrile leukopenic compromised host, and arguably the treatment of intra-abdominal infections. Adding additional antibiotic coverage because the patient is severely ill has no basis in infectious diseases, and this approach is to be discouraged because multiple antibiotics do not cure patients more quickly or effectively. The use of additional unnecessary antibiotics is wasteful and potentially dangerous because additional antibiotics increase the possibility of superinfection, the emergence of resistant organisms, and the potential for adverse drug reactions. The patient's infection should be treated and not the setting as frequently occurs in intensive care units. Antibiotic therapy should ordinarily not be continued for more than 2 weeks unless there are reasons for continuing coverage beyond this time frame. Patients responding rapidly to intravenous antibiotics that are not in intensive care units may be switched to oral antibiotics as soon as clinically feasible. After clinical defervescence, which usually occurs 3 to 5 days after admission to the hospital, switch to an oral antibiotic with an equivalent spectrum, not necessarily of

c.n c.n

'"

Table 3. EMPIRIC ANTIBIOTIC THERAPY BASED ON ORGAN SYSTEM INVOLVEMENT AND PROBABLE PATHOGENS Usual Pathogens Intra-abdominal pelvic source Escherichia coli, Klebsiella, Biliary tract enterococci

Coverage Not Needed

Preferred Therapy

Staphylococcus aureus, B. tragilis

APP, cefoperazone, P:BLI Metronidazole/clindamycin plus aminoglycoside or aztreonam or quinolone or cefipime or P:BLI Metronidazole/clindamycin plus aminoglycoside or aztreonam or quinolone or cefipime or P:BLI Metronidazole/clindamycin plus aminoglycoside or aztreonam or quinolone or cefipime or P:BLI

Penicillin Allergic Metronidazole/clindamycin plus aminoglycoside or aztreonam or quinolone or cefipime Metronidazole/clindamycin plus aminoglycoside or aztreonam or quinolone

Liver

Coliforms, Bacteroides tragilis

S. aureus

Terminal ileum/colon

COliforms, B. tragilis

S. aureus

Pelvis

COliforms, B. tragilis

S. aureus

S. aureus, B. tragilis

Ampicillin plus aminoglycoside, P:BLI or APP

Vancomycin plus aminoglycoside or aztreonam

Prostate

Coliforms, enterococcus, Pseudomonas aeruginosa Coliforms, enterococcus

Urologic instrumentation

Coliforms, P. aeruginosa

S. aureus, B. tragilis, P. aeruginosa S. aureus, B. tragilis

TMP-SMX or quinolone or P:BLI P:BLI or quinolone or aminoglycoside or aztreonam

Quinolone or TMP-SMX or doxycycline Quinolone or aminoglycoside or aztreonam

Common coliforms, B. tragilis

Cefipime or 36cc vancomycin plus aminoglycoside or aztreonam or quinolone Nafcillin or clindamycin

Imipenem or vancomycin plus aminoglycoside or aztreonam or quinolone Clindamycin or vancomycin or teicoplanin

Genitourinary source Bladder/kidneys

Intravascular source IV lines IV grafts/shunts

S. aureus, S. epidermidis, Klebsiella, Enterobacter, Serratia S. aureus

Gram-negative bacilli

Metronidazole/clindamycin plus aminoglycoside or azthreonam or quinolone Metronidazole/clindamycin plus aminoglycoside or aztreonam or quinolone

TMP-SMX = Trimethoprim/sulfamethoxazole; P:BLI = penicillin-beta-Iactamase inhibitor (tazobactam/piperacillin or sulbactam/ampicillin); APP icillin.

0=

antipseudomonal pen-

ANTIIlIOTIC TREATMDJT OF SEPSIS

557

the same class or type, is recommended if there is no gastrointestinal contraindication.7, 16,22,27 SUMMARY

Early, appropriate antibiotic therapy is critical to the treatment of the septicemic patient. The degree of organ dysfunction, underlying medical conditions, and physiologic abnormalities are important prognostic factors but are not important in initial antibiotic selection, Initial empiric therapy should be directed against the resident flora of the organ, which is primarily involved in the infectious process. Blood cultures should be obtained in all patients for the initiation of antibiotic therapy, and methods should be employed for the early detection of septicemia, Other conditions that mimic sepsis, e.g., pseudosepsis, should be ruled out initially to avoid an incorrect diagnosis and unnecessary antibiotic therapy. Monotherapy and fully recommended doses of antimicrobial drugs delivered by the intravenous route as soon as the diagnosis is established remain the cornerstone of therapy in treating the septic patient. Monotherapy with an antibiotic of the appropriate spectrum is more than adequate to treat the great majority of septicemic patients. Double-drug therapy is recommended to treat febrile leukopenic compromised hosts, serious P. aeruginosa infections, and selected cases of intra-abdominal sepsis. At the present time, corticosteroids and mediator therapy have no place in the treatment of the septic patient. References 1. Barron RL: Pathophysiology of septic shock and implications for therapy, Clin Pharm 12:829, 1993 2, Barza M: Pharmacologic principles. In Gorbach SL, Bartlett JG, Blacklow NR (eds): Infectious Diseases. Philadelphia, WB Saunders, 1992, p 147 3. Bone RC: Sepsis and its complications: The clinical problem. Crit Care Med 22:8, 1994 4. Bryan C, Reynolds K, Brenner E: Analysis of 1,186 episodes of gram-negative antimicrobial therapy. Rev Infect Dis 5:629, 1983 5. Corriveau CC, Danner RL: Antiendotoxin therapies for septic shock. Infect Agents Dis 2:44, 1993 6, Cunha BA: Sepsis and its mimics. Intern Med 13:48, 1992 7. Cunha BA: Third-generation cephalosporins, Clin Ther 14:616, 1992 8. Cunha BA: Antimicrobial therapy. In Cunha BA (ed): Infectious Diseases in the Elderly, Lilleton, MA, PSG Publishing, 1988, p 330 9. Eliopoulos GM, Moellering RC Jr: Principles of antibiotic therapy. Med Clin North Am 66:3, 1982 10. Gibaldi M: Anatomy of an antibody, and related misadventures in developing an effective treatment for septic shock. Pharmacotherapy 13:302, 1993 11. Kreger BE, Craven DE, McCabe WR: Gram-negative bacterema. IV, Re-evaluation of clinical features and treatment in 612 patients. Am J Med 68:344, 1980 12. Lowry SF: Sepsis and its complications: Clinical definitions and therapeutic prospects. Crit Care Med 22:1, 1994 13. Martin MA: Epidemiology and clinical impact of gram-negative sepsis. Infect Dis Clin North Am 5:739, 1991

558

CUNHA

14. Moellering RC Jr: Principles of anti-infective therapy. In Mandell GL, Douglas RG, Bennct JE (eds): Principles and Practice of Infectious Diseases, ed 3. New York, Churchill Livingstone, 1990, p 206 15. Neu HC: General therapeutic principles. In Gorbach SL, Bartlett JG, Blacklow NR (eds): Infectious Diseases. Philadelphia, WB Saunders, 1992, p 153 16. Quintiliani R: Strategies for the cost-effective use of antibiotics. In Gorbach SL, Bartlett JG, Blacklow NR (eds): Infectious Diseases. Philadelphia, WB Saunders, 1992, p 348 17. Rackow EC, Astiz ME: Pathophysiology and treatment of septic shock. JAMA 266:548, 1991 18. Ristuccia P, Hoeffner RA, Beltran M, et al: The detection of bacteremia by buffy coat smears. Scand J Infect Dis 19:215, 1987 19. Schedel I, Dreikhausen U, Nentwig B, et al: Treatment of gram-negative septic shock with an immunoglobulin preparation. A prospective, randomized clinical trial. Crit Care Med 19:1104, 1991 20. Sheagren IN: Corticosteroids for the treatment of septic shock. Infect Dis Clin North Am 5:875, 1991 21. Siegel JP, Stein KE, Zoon KC: 41 anti-endotoxin monoclonal antibodies. N Engl J Med 327:890,1992 22. Smith LG: Factors in antibiotic selection. In Ristuccia AM, Cunha BA (eds): Antimicrobial Therapy. New York, Raven Press, 1984, pI 23. Suffredini AF: Current prospects for the treatment of clinical sepsis. Crit Care Med 22:12,1994 24. Waage A, Brandtzaeg P, Espevik T, et al: Current understanding of the pathogenesis of gram-negative shock. Infect Dis Clin North Am 5:781, 1991 25. Warren HS, Danner RL, Munford RS: Anti-endotoxin monoclonal antibodies. N Engl J Med 326:1153, 1992 26. Weinstein M, Murphy L Reller L, et al: The clinical significance of positive blood cultures: A comprehensive analysis of 500 episodes of bacteremia and fungemia in adults. II: Clinical observations with special reference to factors influencing prognosis. Rev Infect Dis 5:54, 1983 27. Yoshikawa TT: Empiric antimicrobial therapy. In Yoshikawa TT, Norman DC (eds): Antimicrobial Therapy in the Elderly Patient. New York, Marcel Dekker, 1994, p 469 28. Young LS: Gram-negative sepsis. In Mandell GL, Douglas RG, Bennet JE (eds): Principles and Practice of Infectious Diseases, ed 3. New York, Churchill Livingstone, 1990, p611

Address reprint requests to Burke A. Cunha, MD Infectious Disease Division Winthrop-University Hospital 259 First Street Mineola, NY 11501