Blood culture surveillance: What is useful and cost-effective?

Editorial

Blood Culture Surveillance: Effective?

What is Useful and Cost-

Charles S. Bryan, MD* In this issue, Taylor and colleagues report their experience with nosocomial gram-negative bacteremia at the University of Alberta Hospital, a Canadian tertiary acute care facility.’ The incidence of gram-negative bacteremia remained remarkably constant over 10 years even though the overall incidence of bacteremia rose, primarily because of gram-positive microorganisms. The l-week mortality rate associated with gram-negative bacteremia declined for reasons that were unclear. This report by Taylor et al adds to an enormous, growing, international database on the significance of positive blood cultures, dating back to the pre-antibiotic em2 But, aside from the archival value of monitoring institutional experience, of what use is such surveillance to infection control and patient care? What types of surveillance activities justify their cost? Three key terms need to be examined. First, “bacteremia” -the traditional focus of such surveillancedenotes the presence of bacteria in blood cultures with or without symptoms of disease. Until the late 1970s blood culture surveillance was essentially synonymous with tracking bacteremias, because other types of organisms were infrequently isolated from blood.3 During the 1980s the rising incidence of fungemias made it clear that such surveillance should incorporate, at the very least, bacteremias and fungemias (Figure 1). Mycobacteria, viruses, and parasites can also be detected in blood. There is no generally accepted inclusive term for the presence of microorganisms of any kind in blood. (In preparing a review article,* the author once proposed “microbemia, ))which was roundly rejected by the reviewers, probably for the best!). Another caveat is that the focus for clinical intervention trials has shifted from “septicemia” (that is, sepsis with positive blood cultures) to “sepsis” because many patients with life-threatening infection do not have positive blood cultures.5 The correlation between sepsis and positive blood cultures is now being evaluated in various countries.‘j Monitoring blood cultures provides an easily obtainable, but incomplete

*Department of Medicine, University of South Carolina School of Medicine, and Richland Memorial Hospital, Columbia, South Carolina. Address correspondence Park, Suite 502, Columbia,

182

to Dr. Charles SC 29203.

S. Bryan,

2 Richland

Medical

O.'

IT. 1977

I, 1990

1995

II 1990

>_I 1995

Year Figure 1. Rates of blood culture utilization, bacteremia, and fungemia at Richland Memorial Hospital, a 649-bed regional community teaching facility in Columbia, South Carolina, 1977 through 1995.

overview of serious infections. Developments during the next century, for example, rapid, broad-spectrum tests for microbial DNA or RNA based on the polymerase chain reaction, could conceivably render obsolete the timehonored focus on blood cultures. Second, although “nosocomial” bacteremia obviously refers to acquisition in the hospital rather than in the community, this distinction can be surprisingly difficult. In 1970, nosocomial bacteremia was defined at the First International Conference on Nosocomial Infections as “culture-documented bacteremia in a hospitalized patient admitted with no evidence of bacteremia.“’ In 1975, McGowan and colleagues proposed that nosocomial bacteremia can be defined, for survey purposes, as one on which the first positive blood culture is obtained on or after the third hospital day.* In 1983, the Centers for Disease Control (CDC) definitions were refmed so that “for an infection to be defined as nosocomial, there must be no evidence that the infection was present or incubating at the time of hospital admission!‘9 Still more recently, the adjective “nosohusial” has been introduced for infections related to home health care.‘O The author’s personal preference is to use McGowan’s definition for epidemiologic surveillance purposes, recognizing that such working definitions do not always correlate with what actually happened. The other defmitions require closer scrutiny of individual patient records, which may or not provide important information that justifies labor and cost.

Blood Culture Surveillance /Bryan

Finally, the term “mortality rate” must also be qualified. Taylor and colleagues followed each patient “ 1 week from first positive blood culture to determine short-term outcome.” This approach is utilitarian but, as the authors surely recognized, does not necessarily correlate with clinical reality. In analysis of 1186 episodes of gram-negative bacteremia, the authors defined deaths due to infection as being those within 7 days of the last positive blood cultures, with no obvious alternative explanation for death.” Weinstein and his colleagues, in their study of 500 consecutive cases of bacteremia and fungemia, concluded that deaths were attributable to infection only after perusal of the hospital record by two physicianinvestigators, working independently.12s13 Subsequent investigators have distinguished between “crude mortality” (all deaths during hospitalization) and “attributable mortality” (deaths due to the infection). Wenzel defined attributable mortality as the difference between the overall case fatality rate for infected patients and the overall case fatality rate for well-matched control patients.l* Identifying well-matched controls is labor-intensive, at least at most institutions. Nevertheless, the difference between crude mortality and attributable mortality assumes societal importance especially with regard to difficult-to-treat pathogens such as vancomycin-resistant enterococci.15 With these caveats in mind, there are five areas in which blood culture surveillance gives potentially useful information: 1. Ongoing surveillance of nosocomial infection. The endemic or background incidence of positive blood cultures at any institution correlates in a general way with the severity of illness (acuity) among the patient popu1ation.l’ Also, positive blood cultures are invaluable for many diagnoses. One must remember that many microbial species, and especially gram-positive bacteria (notoriously Staphylococcus epidermidis, but also other species), often are contaminants rather than true pathogens. 12,13Of interest, the only blood culture surveillance currently used in the CDC’s National Nosocomial Infections Surveillance (NNIS) system concerns the number of primary infections per 1000 central venous catheter (“line”) days.” Unfortunately, determining that infection is primary, as opposed to secondary from an identified infection, can be arbitrary, and determining the number of line-days can be time-consuming. Gold standards for line-related infections do not exist, and interpretation of such data is difficult unless all clinicians follow a standardized protocol.ls 2. Outbreak investigation. Ongoing blood culture surveillance facilitates early identification of some types of outbreaks. These usually (79% of outbreaks in one recent review) involve gram-negative rods.‘9 Intensive care units are the usual setting for such outbreaks, which are to some extent even predictable. For example, based on reports in the literature a group may discuss ways to

183

manage gentamicin-resistant Klebsiella pneumoniae bacteremia in the neonatal intensive care unit well before being faced with such an outbreak.2o Nationwide (or even international) outbreaks can result from contamination of intravenous fluids or blood products, and potentially from medications.21 Outbreaks due to fungal pathogens (most commonly due to Malassezia furfur from lipid-containing preparations, but also others) are also well recognized. 3. Microbiology laboratory quality control. None of the numerous alternative methods for collecting, processing, and reporting of blood cultures meets all desiderata.22-z4Pseudoepidemics (that is, the epidemic occurrence of false-positive blood cultures) can result from problems with skin contamination, phlebotomy, media preparation, or laboratory processing. 25 Pseudobacteremia due to coagulase-negative staphylococci is hyperendemic at many institutions. At Richland Memorial Hospital, a decision to discontinue routine use of a quantitative blood culture system (the lysis-centrifugation system [ISOLATOR], Wampole Laboratories, Cranbury, New Jersey) decreased the ability to recognize such cases of pseudobacteremia, thereby explaining the increased number of gram-positive bacteremias noted during 1995 (see Figure 1). Monitoring blood cultures also provides a way to give feedback to the medical staff concerning the performance of blood cultures and the significance of the results.26b27It is useful to monitor the ratio of blood cultures obtained to positive results (historically, about 10: 1 at Richland Memorial Hospital). 4. Antibiotic review. Drug-resistant microorganisms are rapidly becoming a major problem, to the extent that many authorities predict a “post-antibiotic era” in which life-threatening infections may be essentially untreatable.z8 Judicious prescribing of antimicrobial agents is widely taught but difficult to ensure. Blood culture surveillance furnishes data that bear on such clinical decisions as whether to prescribe vancomycin empirically for suspected Staphylococcus aureus or enterococcal infection. Concurrent review of antimicrobial therapy for each patient with a positive blood culture helps ensure that therapy is both appropriate and judicious. The method consists of carrying out a prompt review of the clinical scenario and drug treatment for each patient with a newly identified positive culture, offering advice to the physician(s) caring for the patient when indicated.29 An extensive and easily obtainable literature exists for nearly all of the commonly encountered bloodstream pathogens, and clinicians usually find such “automatic consultations” to be useful. However, this activity is labor-intensive and may require additional personnel. 5. Evaluation of overall patterns of clinical care. In the author’s experience, the overall hospital mortality (crude case fatality rate) was 30% for patients with positive blood cultures, compared with 2.5% for the overall

184

International

Journal

of Infectious

Table

1.

Proposed

Level 1, least

Levels

/ Volume 1, Number 4, April 1997

of Intensity

of Routine

Blood

Culture

Surveillance

Procedure intense

2, intermediate

3, most

Diseases

intense

Investigational (useful at selected institutions)

Potential

Monitor number of blood cultures and number of positive cultures

intensity

performed

Benefits

Feedback

to medical

Record number of cultures revealing gram-positive bacteria, number of gram-negative bacteria, number of fungi, and number revealing miscellaneous microorganisms, such as mycobacteria

Feedback antibiotics

to medical staff on empirical for suspected sepsis

Periodically review sensitivity profiles of key microorganisms (for example, Streptococcus pneumoniae, S. aureus, Enterococcus faecium, and Pseudomonas aeruginosa)

Early recognition

Tabulate species of all organisms for institution and for each clinical unit on a regular (e.g., monthly) basis

Feedback infection

Tabulate sensitivity profiles for all blood culture by species, on a regular (e.g., every 6 months)

Feedback

isolates, basis

of epidemics

to all health

Feedback to workers care of central lines

Periodically review all cases to one or another pathogen

Feedback pathogens

blood

cultures

due

care workers in intensive

to all health

Assurance that patients with receive appropriate treatment

Complete an epidemiologically oriented fact sheet for each positive blood culture, recording such information as the probable source of infection, antimicrobial therapy, and outcome, as part of an ongoing study of infectious diseases at one’s facility

In-depth selected

Phenotyping

Early recognition of small nosocomial infection

and/or

genotyping

of blood

culture

isolates

Analysis of the cost of medical care of patients with positive blood cultures vs. cost of medical care for matched control patients

Economic measures

Cold storage

Evaluation microbial

of blood

culture

isolates

for future

hospital population. l6 Positive blood cultures thus identify a subset of patients at grave risk of complications. It is well recognized that this picture may be incomplete in certain groups of patients, such as the elderly, in whom manifestations of sepsis may be subtle.30,31Positive blood cultures provide data that are relevant in both human and economic terms, and should therefore be of interest to those who are concerned with the larger socioeconomic dimensions of medicine, including hospital administrators and government policy-makers3* In 1969, Martin proposed that a national bacteremia databank be established in the United States.33 In 1987,

studies

feedback infectious

analysis

use of selected

on efficacy

of all

on antibiotic

use

problem

life-threatening

clusters

of the efficacy

trends

regarding

about

to health care workers disease problems

of long-term ecology

utilization

care units

care workers

Conduct in-depth, concurrent medical record (chart) review of each patient found to have a new positive blood culture, making recommendations to the staff concerning management

culture

and pseudoepidemics

to all health care workers control practice

Tabulate incidence of nosocomial as opposed to communityacquired onset for each case as determined by one or another criterion, with specific emphasis on vascular access (“line”)-associated cases of positive

staff on blood

infections

about

or outbreaks

of infection

of

control

in hospital-associated

Wenzel proposed that attributable mortality due to bacteremia should be introduced as a new vital statistic.‘* No widespread movement to implement these suggestions has been forthcoming. It seems reasonable to suggest that each institution should review its level of commitment to obtaining data concerning positive blood cultures (Table 1). The data reported by Taylor and colleagues in this issue remind us of the value of monitoring positive blood cultures. Each institution will have its unique experience and will find such monitoring to be worthwhile. Blood culture surveillance need not be unusually elaborate or

Blood

time-consuming, and it is gratifying to observe that comprehensive studies based on this activity continue to be reported from various countries around the world.34x3j

ACKNOWLEDGMENT The author thanks Gwen Floyd, RN, James Rampey, MT, Patricia Weathers, RN, and Darlene Whipple, RN, for sharing their ideas concerning the level of intensity of blood culture surveillance.

REFERENCES 1. Taylor GD, Buchanan-Chell M, Kirkland T, McKenzie M, Wiens R. Nosocomial gram-negative bacteremia. Int J Infect Dis 1997; 1:202-205. 2. Keefer CS. The clinical significance of bacteremia. N Y State Med J 1941; 41:976-981. 3. Brenner ER, Bryan CS. Nosocomial bacteremia in perspective: a community-wide study. Infect Control 1981; 2: 219-226. 4. Bryan CS. Clinical implications of positive blood cultures. Clin Microbial Rev 1989; 2:329-353. 5. Wenzel Rp, Pinsky MR, Ulevitch RJ, et al. Current understanding of sepsis. Clin Infect Dis 1996; 22:407-413. 6. Jones GR, Lowes JA. The systemic inflammatory response syndrome as a predictor of bacteraemia and outcome from sepsis. Q J Med 1996; 89:515-522. 7. Garner JS, Bennett jv, Scheckler WE, et al. Surveillance of nosocomial infections. In: Proceedings of the International Conference on Nosocomial Infections, Center for Disease Control, August 3-6, 1970. Chicago: American Hospital Association, 1971:277-281. 8. McGowan JE Jr, Barnes Mvc: Finland M. Bacteremia at Boston City Hospital: occurrence and mortality during 12 selected years (1935-1972) with special reference to hospitalacquired cases. J Infect Dis 1975; 132:315-335. 9. Garner JS, Jarvis WR, Emori TG, et al. CDC definitions of nosocomial infections, 1988. Am J Infect Control 1988; 16:128-140. 10. Graham DR. Nosohusial bacteremia. Proceedings of the 34th Annual Meeting of the Infectious Diseases Society of America, 18-20 September 1996, New Orleans, Louisiana. Il. Bryan CS, Reynolds KL, Brenner ER. Analysis of 1186 episodes of gram-negative bacteremia in non-university hospitals. The effects of antimicrobial therapy. Rev Infect Dis 1983; 5:629-638. 12. Weinstein MP, Reller LB, Murphy JR, Lichtenstein KA. The clinical significance of positive blood cultures: a comprehensive analysis of 500 episodes of bacteremia and fungemia in adults. I. Laboratory and epidemiologic investigations. Rev Infect Dis 1983; 5:35-53. 13. Weinstein MP, Murphy JR, Reller LB, Lichtenstein KA. The clinical significance of positive blood cultures: a comprehensive analysis of 500 episodes of bacteremia and fungemia in adults. II. Clinical observations, with special references to factors influencing prognosis. Rev Infect Dis 1983; 5: 54-70. 14. Wenzel Rl? The mortality of hospital-acquired bloodstream infections: need for a new vital statistic. Trans Am Clin Climatol Assoc 1987; 98:43-48. 15. Edmond MB, Ober JE Dawson JD, Weinbaum DL, Wenzel RF! Vancomycin-resistant enterococcal bacteremia: natural his-

Culture

Surveillance

/ Bryan

185

tory and attributable mortality. Clin Infect Dis 1996; 23: 1234-1239. 16. Bryan CS, Hornung CA, Reynolds KL, Brenner ER. Endemic bacteremia in Columbia, South Carolina. Am J Epidemiol 1986; 123:113-127. 17. Hospital Infections Program, National Center for Infectious Diseases, Centers for Disease Control and Prevention, Public Health Service, U.S. Department of Health and Human Services, Atlanta, Georgia. National nosocomial infections surveillance (jNNIS) report, data from October 1986-April 1996, issued May 1996. Am J Infect Control 1996; 24: 380-386. 18. Scheretz RJ. Surveillance for infections associated with vascular catheters. Infect Control Hosp Epidemiol 1996; 17:746-752. 19. Pittet D. Nosocomial bloodstream infections. In: Wenzel Rp, ed. Prevention and control of nosocomial infections. 2nd Ed. Baltimore: Williams &Wilkins, 1993:512-555. 20. Bryan CS, John JF Jr, Pai MS, Austin TL. Gentamicin vs cefotaxime for therapy of neonatal sepsis: relationship to drug resistance. Am J Dis Child 1985; 139:1086-1089. 21, Felts SK, Schaffner W Melly MA, Koenig MG. Sepsis caused by contaminated intravenous fluids: epidemic, clinical, and laboratory investigation of an outbreak in one hospital. Ann Intern Med 1972; 77:881-890. 22. Murray PR, Traynor P, Hopson D. Critical assessment of blood culture techniques: analysis of recovery of obligate and facultative anaerobes, strict aerobic bacteria, and fungi in aerobic and anaerobic blood culture bottles. J Clin Microbiol 1992; 30: 1462-1468. 23. Mermel LA, Maki DG. Detection of bacteremia in adults: consequences of culturing an inadequate volume of blood. Ann Intern Med 1993; 199:270-272. 24. Wilson ML, Weinstein MP General principles in laboratory detection of bacteremia and fungemia. Clin Lab Med 1994; 14:69-82. 25. Heelan JS, Opal SM, Brissette E, Donahue M. The impact of converting to a biphasic blood-culture system on the overall cost and the incidence of pseudobacteremia. Diagn Microbial Infect Dis 1992; 15:5-11. 26. Weinstein MP Clinical importance of blood cultures. Clin Lab Med 1994; 14:9-16. 27. Hirakata Y, Furuya N, Iwata M, Kashitani E Assessment of clinical significance of positive blood cultures of relatively low-virulence isolates. J Med Microbial 1996; 44:195-198. 28. Gold HS, Moellering RC. Antimicrobial-drug resistance. N Engl J Med 1996; 335:1445-1453. 29. Bryan CS. Strategies to improve antibiotic use. Infect Dis Clin North Am 1989; 3:723-734. 30. Barkham TM, Martin FC, Eykyn SJ. Delay in the diagnosis of bacteraemic urinary tract infection in elderly patients. Age Ageing 1996; 25:130-132. 31. Fontanarosa PB, Kaeberlein FJ, Gerson LW, Thomson RB. Difficulty in predicting bacteremia in elderly emergency patients. Ann Emerg Med 1992; 21:842-848. 32. Bates DW, Pruess KE, Lee TH. How bad are bacteremia and sepsis? Outcomes in a cohort with suspected bacteremia. Arch Intern Med 1995; 155:593-598. 33. Martin CM. A national bacteremia registry (Editorial). J Infect Dis 1969; 120:495-496. 34. Mizushima Y Kawasaki A, Hirata H, et al. An analysis of bacteraemia in a university hospital in Japan over a IO-year period. J Hosp Infect 1994; 27:285-298. 35. Salomao R, Castelo Filho A, Pignattari AC, Wey SB. Nosocomial and community acquired bacteremia: variables associated with outcomes. Rev Paul Med (Sao Paulo) 1993; 111: 456-461.