Intravascular Catheter Use
How to Tell When the Medicine Is Worse Than the Malady

William E. Trick, M.D.

Centers for Disease Control, Atlanta, Georgia

Robert A. Weinstein, M.D.

Cook County Hospital and Rush Medical CollegeChicago, Illinois

	ARTICLE

TOP ARTICLE REFERENCES

Intravascular catheters are essential to the practice of modern medicine. Although technologic advances have improved the safety of these devices, catheter-related infections remain a major problem. There are an estimated 16,000 catheter-related bloodstream infections annually among patients in U.S. intensive-care units (1). To inform clinicians about the risks of intravascular catheters, and to provide information for analyses of cost-effectiveness, precise estimates of the outcomes of these adverse events are needed.

In this issue, Renaud and colleagues (pp. 1584-1590) (2) describe their results from a multicenter study that evaluated outcomes of bacteremia among intensive-care unit patients. They address two primary objectives: (1) to determine the relative proportion of primary, catheter-related, or secondary bacteremia among intensive-care unit patients, and (2) to estimate the influence of each of these sources on mortality and intensive-care-unit length of stay. Because studies addressing outcomes of bacteremia stratified by source are uncommon and because these results are of interest to clinicians, researchers, and administrators, we will focus on the second objective.

Renaud and colleagues report that attributable mortality (the difference between the crude mortality for an affected, e.g., bacteremia, and an unaffected patient population) varied by category: 11.5% for catheter-related bacteremia, 28.6% for primary bacteremia, and 54.8% for secondary bacteremia. Compared with patients who were not bacteremic, the mortality was significantly increased for patients who had a secondary bacteremia or when primary and catheter-related bacteremia were grouped. The investigators also report a median excess length of stay for patients with primary or catheter-related bacteremia of 9.5 d, similar to that reported by other investigators (3, 4). However, when comparing the results of Renaud and colleagues with those of prior studies there are several important considerations: the distribution of pathogens recovered from the study population, and the study methodology, especially case definitions and matching strategies.

Pathogen distribution. Compared with other causes of bloodstream infection, Staphylococcus aureus and Candida spp. appear to result in worse patient outcomes (5, 6); thus, the external validity of a study may be limited if the distribution of pathogens associated with catheter-related bacteremia is inconsistent with other reports. Renaud and colleagues report a distribution of pathogens recovered from patients with primary and catheter-related bacteremia that is consistent with data reported from medical-surgical intensive-care units in the Centers for Disease Control and Prevention's National Nosocomial Infection Surveillance (NNIS) system (7): coagulase-negative Staphylococci, 23% versus NNIS, 39%; S. aureus, 13% versus NNIS, 11.5%; and Candida species, 9.8% versus NNIS, 11.5%.

Case definition. Although Renaud and colleagues conclude that differentiating definite catheter-related bacteremia from bacteremia without an identified source may be useful, their case definitions need to be carefully considered. Their definition of secondary bacteremia (the same organism has been recovered from blood and from an infection at another site) is consistent with the literature; however, their definitions used to distinguish primary and catheter-related bacteremia as separate entities are not entirely consistent with definitions advocated by others (8, 9). By requiring catheter exit site inflammation in addition to recovery of the same organism from blood and catheter tip cultures, Renaud and colleagues' definition of catheter-related bacteremia is highly specific but relatively insensitive. This leads to misclassification of some catheter-related bloodstream infections as primary infections. To support this contention we note that in the study of Renaud and colleagues (2) and in another study (10), the pathogens recovered from patients with primary and catheter-related bacteremias were similar, and differed from the distribution of pathogens recovered from patients with secondary bacteremia. It is possible that a more sensitive case definition would have resulted in a smaller difference in attributable mortality between primary and catheter-related bacteremia.

In Renaud's study, patients who had a secondary bacteremia had a higher mortality than did patients with catheter-related bacteremias, consistent with findings by other investigators that bacteremia secondary to pneumonia results in increased mortality (5). To avoid using overestimates of catheter-related attributable mortality, assessments of the effectiveness of intravascular-catheter modifications should not use results from outcomes studies that include secondary bacteremias.

Matching strategies. By multivariate analysis of the entire cohort, Renaud and colleagues identified patient predictors of prolonged length of stay. Then, using these predictors (age, patient location before the ICU admission, admission category, Simplified Acute Physiologic Score II, and a rapidly or ultimately fatal underlying disease), they performed a pair-matched study.

The appropriate day to use when matching clinical characteristics for epidemiologic analyses has varied between studies. One strategy is to match patients with measurements recorded on the day of admission, as in Renaud's study; an alternative strategy is to match on measurements recorded on a prespecified day before the detection of the bacteremia. Matching on the day of admission may not adequately control for severity of illness, e.g., in Renaud's study the case-patients' clinical condition may have deteriorated between the day of admission and the occurrence of bacteremia (median of 10 d). Clinical deterioration, or failure to improve, may be unrelated to bloodstream infection, resulting in an overestimate of attributable mortality, as demonstrated in two recent studies (3, 11). In one of these studies, the severity of illness for the cases and noncases diverged between the day of admission and the day of bacteremia, suggesting that patients who did not improve or who became increasingly ill during their hospitalization had an increased risk of infection (11). In the other study, serial measurements of severity of illness scores demonstrated stable scores until the day of infection, followed by an abrupt clinical deterioration on the day of infection (3). Although the number of participants was small, these two studies that matched on, or adjusted for, severity of illness later in the patients' intensive-care unit stay did not detect a significant increase in mortality for bacteremic patients.

Renaud and colleagues are to be commended for conducting a complex multicenter study that adds additional evidence that bacteremia (regardless of the category) results in a prolonged length of stay and that the attributable mortality for secondary bacteremia may be at least twice as high as catheter-related or primary bacteremia. Additional studies may provide further guidance regarding the best method for differentiating primary and catheter-related bacteremia.

	References

TOP ARTICLE REFERENCES

1. Mermel L. Prevention of intravascular catheter-related infections. Ann Intern Med 2000; 132: 391-402 [Abstract/Free Full Text].

2. Renaud B, Brun-Buisson C. the ICU-Bacteremia Study Group. Outcomes of primary and catheter-related bacteremia: a cohort and case-control study in critically ill patients. Am J Respir Crit Care Med 2001; 163: 1584-1590 [Abstract/Free Full Text].

3. Digiovine B, Chenoweth C, Watts C, Higgins M. The attributable mortality and costs of primary nosocomial bloodstream infections in the intensive care unit. Am J Respir Crit Care Med 1999; 160: 976-981 [Abstract/Free Full Text].

4. Pittet D, Tarara D, Wenzel RP. Nosocomial bloodstream infection in critically ill patients: excess length of stay, extra costs, and attributable mortality. JAMA 1994; 271: 1598-1601 [Abstract].

5. Pittet D, Li N, Woolson RF, Wenzel RP. Microbiological factors influencing the outcome of nosocomial bloodstream infections: a 6-year validated, population-based model. Clin Infect Dis 1997; 24: 1068-1078 [Medline].

6. Arnow PM, Quimosing EM, Beach M. Consequences of intravascular catheter sepsis. Clin Infect Dis 1993; 16: 778-784 [Medline].

7. Richards MJ, Edwards JR, Culver DH, Gaynes RP. the National Nosocomial Infections Surveillance System. Nosocomial infections in combined medical-surgical intensive care units in the United States. Infect Control Hosp Epidemiol 2000; 21: 510-515 [Medline].

8. Maki DG, Mermel LA. Infections due to infusion therapy In: Bennett JV, Brachman PS, editors. Hospital infections. Philadelphia: Lippincott-Raven; 1998. p. 689-724.

9. Sheretz RJ. Surveillance for infections associated with vascular catheters. Infect Control Hosp Epidemiol 1996; 17: 746-752 [Medline].

10. Maki DG, Zilz MA. What is the cause of primary bloodstream infections not linked to an intravascular catheter? Most are IV catheter-related [abstract 661]. Clin Infect Dis 1997; 25: 477 .

11. Soufir L, Timsit JF, Mahe C, Carlet J, Regnier B, Chevret S. Attributable morbidity and mortality of catheter-related septicemia in critically ill patients: a matched, risk-adjusted, cohort study. Infect Control Hosp Epidemiol 1999; 20: 396-401 [Medline].