Chapter 26 ICU Infections

573

CHAPTER 26 • ICU Infections

the need for mechanical ventilation are moderate contraindications to forceps biopsy, lavage is virtu- ally always feasible, and gentle, protected specimen brushings can be obtained safely when care is taken to administer platelets and/or deficient clotting fac- tors beforehand. Bronchoscopic yield varies greatly with the specific disease process and with the tim- ing and method of conducting this procedure. For example, when all specimen-gathering techniques (biopsy, brushings, and lavage) are used, a precise diagnosis can be established about 50% of the time. In special instances, such as patients with AIDS, the yield is considerably higher. Open lung biopsy often is delayed because of its perceived morbidity and expense. In fact, open biopsy, a 20- to 40-minute procedure, is usually well tolerated and often helpful if conducted early in the course of the illness. It is the most reliable means of securing tissue for histologic diagnosis while establishing effective hemostasis in patients at high risk for bleeding. VATS is another approach of merit in patients with good hemostasis. The expense of open biopsy should be considered along with the high cost of empiric antibiotic therapy. Not only are multidrug combinations expensive, but also some of the commonly used agents carry substantial risk of toxicity for kidneys and bone marrow. Whatever the value of open lung biopsy may be when undertaken early in the course of dis- ease, it is clearly less valuable after broad-spectrum antibiotics have been given for a prolonged period. In such instances, it is unusual for open biopsy to add sufficient new information to warrant its atten- dant drawbacks. Failure to define a specific etiology should not mandate continuation of empiric antibi- otics indefinitely. Not only can antibiotic manage- ment be streamlined when a specific diagnosis has been made, but also rational reductions in therapy can be made in the patient improving on multiple drugs. Usually, this takes the form of sequentially removing the antibiotic least likely to benefit or most likely to cause toxicity from the combination every 1 to 2 days. The process of trimming anti- biotic coverage often is delayed until patients are no longer granulocytopenic. Even while patients remain on multiple antibiotics, the clinician must remain alert to “superinfection” with a new or resistant organism. Furthermore, if a patient fails to improve with specific therapy directed against a known pathogen, a second organism commonly is present. For example, patients with AIDS and confirmed Pneumocystis pneumonia who fail to

respond to trimethoprim–sulfamethoxazole often have coexistent CMV infection. When no specific diagnosis has been made and the clinician is forced to choose a regimen, it should be remembered that Legionella and Pneumocystis are among the most lethal and common pathogens. For the immuno- compromised patient without a diagnosis, a third- generation cephalosporin and an aminoglycoside or quinolone, plus erythromycin or doxycycline and trimethoprim–sulfamethoxazole, are often chosen as initial therapy. When methicillin-resistant staph- ylococci are prevalent, vancomycin often is added or substituted. In centers in which fungi present a major problem, potent antifungals (e.g., ampho- tericin) are often begun very early in the course. The use of ultra–broad-spectrum antibiotics (such as the carbapenems) may help to greatly simplify initial coverage, but such drugs present their own set of problems in terms of expense and the induc- Neutropenia most frequently is an iatrogenic complication of antineoplastic chemotherapy. These same drugs profoundly impair host ability to maintain the integrity of tissues having rapid cellular turnover (e.g., bowel wall and the muco- sae of gingiva and rectum). Therefore, it is not surprising that diffuse necrotizing colitis, ano- rectal cellulitis, and typhlitis (a severe bacterial infection of the cecum mimicking appendicitis) occur relatively frequently. Violation of the nor- mally intact integument by intravenous catheters, surgical incisions, or decubitus ulcers also opens a portal for bacterial entry. Therefore, for febrile neutropenic patients, the physical examination should routinely include catheter entry sites and the gingival and perirectal regions. Lack of tender- ness with gentle palpation of the anal verge usu- ally suffices to exclude this as a site of infection. All too often, no site is found for bacteremia or the sepsis syndrome. General Treatment Principles Survival of the neutropenic patient depends on early empiric therapy with more than one antibiotic effective against the infecting organism. The most common organisms include Gram-negative rods (particularly Pseudomonas ), Staphylococcus, and fungi (e.g., Aspergillus and Candida ). Patients with impaired cell-mediated immunity are more likely to be infected with Pneumocystis or Candida . When tion of multiply resistant bacteria. Nonpulmonary Sites of Infection