AGGRESSIVE INITIAL EMPIRIC ANTIMICROBIAL THERAPY
Today’s mantra for antimicrobial prescribing in the ICU reads: “Hit early, hit hard, and then
de-escalate.” Aggressive initial therapy correlates with survival. Limiting the duration of
broad-spectrum therapy reduces the likelihood of drug-resistant pathogens not only for the
patient being treated but also for the ICU, the hospital, and even for society as a whole.
Numerous studies over the past two decades demonstrate that inadequate antimicrobial
therapy leads to increased mortality, prolonged lengths of stay, and poorer outcomes (6–9).
Results of a study involving more than 600 patients indicated that the survival rate decreased by
7.6% for every one-hour delay in treatment (8). Prior to the year 2000, investigations of the effect
of initial “appropriate” antimicrobial therapy [usually defined by the use of agents to which the
eventual pathogen(s) were determined to be susceptible] focused mainly on bloodstream
infections, which allow easy retrospective analysis based on “clean” bacteriologic specimens.
Such studies amply confirmed lower mortality rates for patients who received appropriate
initial antimicrobial therapy (10,11). More recent data extend these observations to patients with
ventilator-associated pneumonia (VAP) and sepsis. The Monoclonal Anti-TNF: A Randomized
Clinical Sepsis (MONARCS) trial was conducted in 157 centers across North America to assess
the safety and efficacy of afelimomab (a TNF-ablocker) in sepsis. Out of a total of 2634 patients
enrolled, 91% got adequate antibiotics. The most common gram-positive organisms were
Staphylococcusspp. andS. pneumoniae, and the most common gram-negative pathogens were
Escherichia coli,Klebsiellaspp., andPseudomonas aeruginosa. Overall mortality rate was 34%; the
breakdown was 33% and 43% for patients who got adequate and inadequate antibiotics,
respectively (12). Another Sepsis trial from Spain found excess in-hospital mortality of 39% with
inadequate initial treatment. There was also an increase in ICU and hospital length of stay (9).
Factors to consider when prescribing initial empiric antimicrobial therapy include the
following (Table 1):
- The duration of hospitalization and recent antimicrobial exposure: Patients who have been
hospitalized for less than 48 hours and who have not had recent exposure to antibiotics
are more likely to have typical “community-acquired” pathogens. Common examples
includeStreptococcus pneumoniaeandHaemophilus influenzaein pneumonia,E. coliin
urinary tract infection (urosepsis), andS. aureus[both methicillin-susceptible (MSSA)
and methicillin-resistant (MRSA)S. aureus] in endocarditis or undifferentiated sepsis
syndrome. Patients who have been hospitalized for longer durations and who have
received multiple prior antibiotics should receive appropriate treatment for drug-
resistant gram-negative bacilli, MRSA, and—if the clinical setting “fits”—anaerobic
pathogens. The guidelines of the American Thoracic Society and the IDSA for the
management of health care–associated pneumonia (HCAP) suggest that risk factors for
multidrug-resistant (MDR) pathogens are antimicrobial therapy within the last three
months, current hospitalization for more than five days, immune suppression, local
epidemiological data suggesting a high frequency of antibiotic resistance in the
community, and risk factors for HCAP (13). The recommended regimens include an
aminoglycosideoran antipseudomonal fluoroquinoloneandan appropriateb-lactam—
if extended-spectrumb-lactamase (ESBL) or MDR pathogens are suspected, then a
carbapenem—andtreatment for MRSA if the latter is suspected. Critically ill patients
are also at risk for yeast infections, with reported rates of 1% to 2% of invasive
candidiasis, although it still remains unclear whether to prescribe empiric antifungal
drugs in the nonneutropenic patient (14). In a recent study of 270 adult ICU patients
with fever despite broad-spectrum antibiotic therapy, empiric use of fluconazole did
not improve the stated outcome compared with placebo, but reduced the incidence of
candidemia in the treated population (15). - The clinical syndrome: Pneumonia in patients who have been hospitalized for more
than 48 hours is most often due to gram-negative bacilli includingP. aeruginosa,
Klebsiella pneumoniae, andAcinetobacter baumanii, but can also involve gram-positive
pathogens including MRSA. Urosepsis in patients with prolonged hospitalization is
commonly due to gram-negative bacilli. Patients who lack an obvious source of
infection are classified as having “primary bacteremia (or fungemia),” which is most
488 Ahuja et al.