Assessment of catheter-associated infection risk with the Hickman right atrial catheter.

One hundred fifty Hickman right atrial catheters were inserted into 143 patients and were followed prospectively until removal. Primary indications for their use were: cancer chemotherapy (45), parenteral nutrition (35), antibiotic therapy (63), and miscellaneous (7). The overall catheter-associated infection rate was 12.0%. Since the mean duration of catheterization was 125 days, the infection/duration rate was 1.0/1,000 days of use. The risk of infection differed significantly according to the primary indication for catheterization: parenteral nutrition greater than antibiotic therapy greater than cancer chemotherapy. The increased risk of catheter-associated infection attributable to duration of catheterization was additive, and the per day risk of such infections remained constant regardless of duration. Nearly two-thirds of patients were discharged home with catheters in place, without adversely affecting infection risk.

Lung biopsy in immunocompromised hosts.

Ninety-five lung biopsy procedures in 78 immunocompromised patients yielded treatable diagnoses in 35 per cent of the needle aspirates, 46 per cent of the cutting needle biopsies and 65 per cent of the open thoracotomies. Complication rates of bleeding or pneumothorax were comparable to those previously described in nonimmunocompromised patients. In patients with coagulation defects, the risk of having bleeding complications was high. Among 44 patients with primary lymphoma, 22 (50 per cent) had a lymphomatous infiltrate in the lung. Twenty-two patients with severe hypoxemia were studied and their mortality, although significantly greater than in the patients without hypoxemia, was marginally improved when a treatable lung lesion was found. The over-all recovery rate was 70 per cent when a treatable diagnosis was made in contrast to 25 per cent when there was no specific diagnosis. Hence lung biopsy, particularly by thoracotomy, appears to be a valuable procedure in immunocompromised patients who have otherwise undiagnosable lung lesions.

Fever in acute myelogenous leukemia.

Implications and course of fever were evaluated during hospitalization of 24 patients with acute myelogenous leukemia. Forty-five febrile episodes were identified. Fever present at admission was usually associated with a diagnosable and treatable infection; fever shortly after induction was self-limited; and fever during granulocytopenia was more likely to be associated with bacteremia. Bacteremia and pneumonia were the most common types of infection. Only Gram-negative bacteria and Candida were identified as causes of infection during life, with Pseudomonas and Klebsiella the most frequently isolated pathogens. Invasive candidiasis was a major postmortem finding. A delay in initiation of empirical treatment beyond the third day of fever was associated with an increase in mortality as was continuation of treatment for longer than 14 days.