Now, we'll apply the principles discussed in the course thus far to the case of a critically ill septic patient. In this segment, we illustrate the thought processes that ought to be used in selecting appropriate empiric antimicrobial therapy in such a patient. The patient was a 56-year-old, previously healthy man who is admitted to the hospital with a severe acute diverticulitis. Empiric antibiotic therapy was initiated at that time with ampicillin plus sulbactam. However, the patient continued to do poorly and required a partial colectomy. Ampicillin-sulbactam was continued, but on the third postoperative day, he developed confusion, fever, leukocytosis, tachycardia, and hypotension. His blood pressure was 90 over 60 with a pulse of 130 and a respiratory rate of 20 per minute. On examination, there was no evidence of infection of the surgical wound and no other obvious source of infection. The central venous catheter site appeared to be non-infected but it was noted that an indwelling bladder catheter was in place. Initial laboratory studies found that the patient had a total white blood count of 14,900 per cubic millimeter and that his serum lactic acid concentration was elevated at 3.9 millimolar. Fluids were rapidly administered and his blood pressure improved. Blood and urine cultures were rapidly obtained. Chest X-ray showed no evident pulmonary disease. Because of concern that the central venous catheter might have been the source of infection, the catheter was removed and replaced at another site. Computerized tomography of the abdomen and pelvis was planned and antibiotic therapy was rapidly initiated. Question at this point is, what antibiotics should be chosen to be initiated in this circumstance and what factors should be considered in making this decision? A critical initial step is the assessment of the potential sources of infection. In this patient, the most likely sites of infection to be considered include the central venous catheter that had been in place for three days and his urinary tract which had an indwelling bladder catheter, as well as an intra-abdominal site related to the recent surgical procedure. Another critical consideration in choosing empiric antibiotic therapy is an assessment of the likely etiologic pathogens, an assessment that is, of course, dependent on knowledge of the likely source of infection. With central venous catheter infection, although gram-positive organisms including ones resistant to most beta-lactam antibiotics are most frequently encountered, gram-negative bacteria and Candida must also be considered. If the catheterized urinary tract is the source, the pathogens are most likely to be gram-negative bacteria. If, however, the source of infection is the abdomen and caused by contamination with intestinal contents related to a leak from the surgical site, the pathogens would, of course, reflect the fecal stream and include a mixed aerobic, anaerobic flora with both gram-positive and gram-negative organisms. The next consideration is an assessment of the likelihood that the pathogen or pathogens causing this infection is antibiotic resistant. This infection was hospital-acquired and it occurred while the patient was receiving ampicillin-sulbactam. At a minimum, one must assume that the infecting pathogen or pathogens are resistant to that beta-lactam, beta-lactamase inhibitor combination. The local antibiogram must be taken into account. In particular, are there multi-drug resistant organisms that are being transmitted within the hospital and/or the relevant patient location? Finally, what is the margin for error? How critically ill is the patient? How steep is the likely downward slope if intervention isn't optimal? What is the likely outcome if the antibiotic regimen you choose is inadequate? Taking all this into account, our empiric choice of antibiotics should include one or more agents active against staphylococci, including especially coagulase-negative staphylococci because of the possibility of central venous catheter infection. Should also provide coverage of enterococci because of the possible intra-abdominal source of infection and given that this infection occurred in an inpatient while receiving ampicillin-sulbactam, we would assume that an enterococcus would be ampicillin resistant. Anaerobes are likely present if the intraabdominal site is the source. Although the infection did occur while the patient was receiving ampicillin-sulbactam, which has excellent activity against anaerobes. Aerobic gram-negative including multi-drug resistant organisms in some circumstances must be considered. We also must consider the possibility that Candida may be playing a role either from an intra-abdominal or a central line infection. The empiric antibiotic regimen chosen for this patient included vancomycin which would cover both methicillin-susceptible and resistant staphylococci as well as ampicillin-resistant enterococci, but would not, of course, cover vancomycin-resistant enterococcus by definition. If the presence of VRE is of concern because of, for example, knowledge of prior colonization or because of a high local prevalence of the organism, the use of linezolid or daptomycin might be preferable for coverage. Gram-negative coverage with piperacillin-tazobactam, which also includes broad anaerobic coverage is reasonable as is use of a carbapenem such as ertapenem. Anti-pseudomonal coverage is not likely to be necessary, so meropenem and imipenem may not be indicated. The combination of a third-generation cephalosporin with metronidazole is also reasonable given the fact that there's already been a decision to administer vancomycin which would cover enterococci. Having chosen an empiric antibiotic regimen, consideration must be given to optimal dosing. For vancomycin, the pharmacodynamic target requires achieving a ratio of the area under the curve to the minimal inhibitory concentration of the pathogen of 400 or greater. It should be noted though that this has generally been applied to MRSA bacteremia and whether it applies as well in this circumstance is not fully known. In practice, of course, the trough vancomycin concentration, Cmin, is generally used as a surrogate measure for area under the curve. Since achieving an adequate trough as soon as possible improves outcomes, it is currently recommended that patients such as this one who was critically ill receive a loading dose of 25-30 milligrams per kilogram of vancomycin. This would then be followed by weight-based dosing at 15 milligrams per kilogram with adjustments designed to achieve a serum trough concentration or Cmin of 15-20 micrograms per ml. Once again, it should be noted that these recommendations are derived from studies focusing on MRSA bacteremia and the target may differ in dealing with infections due to other organisms and other sites of infection. For this patient, the clinician decided piperacillin-tazobactam for broad spectrum coverage. As a beta-lactam, dosing should be aimed at achieving a pharmacodynamic target in which the serum concentration is maintained above the minimal inhibitory concentration of the pathogen for at least approximately 50 percent of the dosing interval. The precise percentage required varies with regard to achievement of bacteriostatic or bactericidal activity and also to some extent depending on the pathogen, but 50 percent is within the general target range. The likelihood of achieving this target especially for susceptible organisms with minimal inhibitory concentrations at/or near the breakpoint is most reliably achieved by use of either a continuous beta-lactam infusion or by prolonging each individual intermittent infusion. For this patient, vancomycin was administered in an initial dose of 2000 milligrams given over two hours followed by 1200 milligrams every 12 hours consistent with his body weight and his normal serum creatinine. Piperacillin-tazobactam was administered in a dose of 4.5 grams of piperacillin every eight hours but with each dose administered over four hours. It was elected in this patient to not initiate empiric coverage for a Candida infection. Two days later, the patient was clinically stable and afebrile with normal vital signs. Computerized tomography of the abdomen and pelvis performed shortly after admission did not reveal intra-abdominal infection. Urine and blood cultures yielded Escherichia coli which as was suspected was resistant to ampicillin-sulbactam, but was susceptible to piperacillin-tazobactam, meropenem, ciprofloxacin, and cefazolin among other antibiotics. At that point, it was appropriate to utilize the available clinical laboratory information to perform an antibiotic time-out and optimize the patient's antibiotic therapy. In general, an antibiotic time-out may result in the administration of more potent agents or broadening of the spectrum of activity based on the microbiological data as well as resulting in resolution of any bug-drug mismatches. Alternatively, the treatment may be refined by de-escalation, which may involve eliminating antibiotic coverage redundancies, narrowing the spectrum of activity, converting from IV to oral administration or in some cases, discontinuation of antibiotic therapy. During the antibiotic time-out, the clinician should also, to the extent possible, provide an estimate of the necessary duration of antibiotic therapy. For further in-depth discussion of the antibiotic time-out and practice applying it to cases, we direct learners to our companion course entitled Optimizing Therapy with Timeouts. For our patient with urinary catheter-associated urosepsis, IV treatment could be de-escalated to cefazolin alone, with consideration, assuming continued clinical improvement, of impending conversion to orally administered ciprofloxacin. It is anticipated that antibiotic administration may continue for 10-14 days although that may be excessive, with most of the antibiotic being administered in the outpatient setting.
